Protocol Raw Financial Model Documentation¶
Purpose: Mental models and frameworks for understanding Protocol Raw's financial mechanics Audience: Founder reference, investor discussions, board reporting Version: 2.4 Last Updated: 17 April 2026 Replaces: v2.3 (15 April 2026) Status: Aligned with Business Plan v3.4, Growth Strategy v3.2, Growth Playbook v2.5, REF-MICRO-01 v2.0, and the unified VC Path sheet in Protocol_Raw_Financial_Model_v7.xlsx
Key Changes in v2.4: Box-2 retention base case corrected from 75% to 70% to align with Business Plan v3.4 and Growth Strategy v3.2. Steady-state churn unchanged at 5.5%. New lifespan 13.73 months (was 14.64). New blended churn 7.3% (was 6.8%). Contribution LTV trajectory updated: Phase B Early £251 to £221, Phase B Late £416 to £360, Phase C £714 to £594. LTV:CAC trajectory: Phase B Early 3.8:1 to 3.4:1, Phase B Late 6.3:1 to 5.5:1, Phase C 10.8:1 to 9.0:1. M22 ARR approximately £13.1M (was £13.75M). M36 ending cash approximately £21.3M (was £22.5M). Terminology update: "Referral %" renamed to "Organic %" throughout. Organic captures coded referrals (SOP-REF-01 slug attribution), attributed word-of-mouth at signup, direct traffic, and branded search. Coded referrals remain the tracked subset targeting 10-12% of customers; the balance of organic share is attributed at signup. Paid CAC, blended CAC, and the overall trajectory are unchanged. Effective organic CAC blended to £10-12 (coded referral £15, attributed organic near-zero).
How to Use This Document¶
This is not a spreadsheet. It's the thinking behind the spreadsheet.
Use this when: - Preparing for investor conversations - Explaining financial mechanics to advisors - Stress-testing assumptions - Making strategic decisions that have financial implications
The live models (spreadsheets, Metabase dashboards) contain the numbers. This document contains the logic.
Part 1: The Fundamental Equation¶
Business Value ≈ Customers × (Contribution LTV − CAC) − Fixed Costs
Everything else is detail. Every financial decision either: - Increases customers (acquisition) - Increases LTV (retention, ARPU) - Decreases CAC (efficiency, organic acquisition) - Decreases operating costs (automation, scale)
If a decision doesn't improve one of these, question why you're doing it.
Part 2: The Two Cost Worlds¶
Protocol Raw has two fundamentally different cost structures. Confusing them leads to bad decisions.
Variable Costs (Per-Box)¶
These scale linearly with volume. They set the floor for pricing.
Variable costs now include the gross production required to yield each shipped box (reject overhead) plus any upstream poultry gate cost allocated per box. See Part 2A for the yield model.
| Phase | Months | COGS/box (incl. reject) | CPD/box | Total variable/box | Driver |
|---|---|---|---|---|---|
| A (Self-Mfr) | M0-M8 | ~£56 (4% reject + gate) [CHECK] | £22 | ~£78 | Self-manufacture, poultry gate active, small batches |
| B Early (Co-Pack) | M9-M12 | £58.65 (18.6% reject, no gate) [CHECK] | £19 | £77.65 | Co-packer transition; reject rate destroys COGS/kg gain |
| B Late (Co-Pack) | M13-M22 | £50.28 (18.6% reject, no gate) [CHECK] | £18 | £68.28 | Co-packer volume leverage; reject rate unchanged |
| Parallel | M28-M33 | £53.62 → £44.18 [CHECK] | £14 | £67.62 → £58.18 | Facility ramp; poultry gate reinstates progressively |
| C (In-House) | M34+ | £44.18 (4% reject + gate at scale) [CHECK] | £14 | £58.18 | Owned production + fulfilment, poultry gate at 2-3% of revenue |
Phase B/C COGS figures derived from COGS/kg × 12.4 kg × (1 ÷ (1 − reject rate)): B Early £3.85/kg × 12.4 × 1.2285 = £58.65; B Late £3.30/kg × 12.4 × 1.2285 = £50.28; In-house £2.90/kg × 12.4 × 1.0417 = £37.46 + allocated poultry gate ~£6.72/box at scale. All [CHECK] against the VC Path sheet.
Key insight: Co-packing at 18.6% reject permanently caps contribution margin at ~26%. In-house manufacturing is the only way to reinstate the poultry gate at scale and deliver the 45.9% contribution margin the business model requires. Co-pack is a funded bridge, not a permanent state. See Part 2A for the microbiological risk framework, Part 2B for the CPD methodology, and Part 8 for the four-phase manufacturing transition.
Competitive precedent: Bella & Duke (£11-19M revenue) and Butternut Box (£10-20M revenue) both built in-house facilities at similar revenue scale. B&D's gross margin improved from 31% to 54% over four years following their facility build (Companies House filings FY2021-FY2025). Butternut Box's gross margin improved from 37.9% to 47.5% as revenue grew to £126.7M (FY2023 strategic report).
Fixed Costs (Monthly)¶
These exist regardless of volume. They determine operating leverage.
Fixed costs by phase (team and tooling):
| Cost | Phase A (M0-M8) | Phase B Co-Pack (M9-M21) | Phase C In-House (M34+) |
|---|---|---|---|
| Tooling (Supabase, Make, Shopify, Claude, Customer.io, etc.) | £300 | £1,500 | £3,750 |
| HQ team (founder + execs + support) | £0 | £10,300 → £12,500 | £43,300 |
| Manufacturing core (facility, utilities, production staff, QA) | £0 | £0 | £41,667 |
| Fulfillment opex (pack wave, warehouse ops, dispatch) | £0 | £0 | £41,667 |
| Other (legal, accounting, insurance) | £2,000 | £3,500 | £5,000 |
| Total (excl. ad spend) | £2,300 | £15,300 → £17,500 | ~£135,000 |
Phase A tooling breakdown: Supabase Pro £20, Make.com Core £9, Shopify Basic £31, Seal Subscriptions £0-40, Metabase £6, Cloudflare £0, email/domain £10, AI API £15-30, Claude Max 5x £80, Customer.io (free startup plan; £80/month if expires). Total ~£290-310/month. Phase B adds: Customer.io Essentials ~£80, Make.com upgrade ~£29, Supabase additional storage/compute ~£50, Seal upgrade ~£49. Phase B HQ team ramps M9-M21 (Ops Lead £4,500, Marketing Lead £5,417 from M12, CS £3,000 from M15, Product Engineer (AI Native) fractional £3,000 from M16). Phase C HQ team adds Series A exec hires from M22 (CMO £8,333, FD £8,333, COO step-up £6,250, Mktg Lead £5,417, Product Engineer (AI Native) £4,583, CS x2 £5,500, HR £1,500 = ~£40,000/month). Phase C facility at steady state ~20-24 FTE (8 HQ + 12-16 facility).
Phase C facility opex breakdown (M34+, two cost centres):
Manufacturing core (row 54 in VC Path, £41,667/month = £500k/year):
| Cost | Monthly | Annual | Notes |
|---|---|---|---|
| Facility lease + rates | £5-8k | £60-100k | 8-15k sq ft, Midlands/North |
| Utilities (cold + power) | £3-6k | £40-70k | Cold storage intensive |
| Production staff (Head of Mfg, supervisors, operatives) | £12-18k | £150-220k | 8-10 production FTE |
| QA / food safety | £4-6k | £50-70k | Quality Manager + poultry gate ops |
| Maintenance & engineering | £3-5k | £40-60k | Fractional/contract |
| Insurance & compliance | £2-3k | £20-30k | Food manufacturing specific |
| Total Manufacturing Core | £29-46k | £360-550k | Model: £500k/yr |
Fulfillment / warehouse (row 55 in VC Path, £41,667/month = £500k/year):
| Cost | Monthly | Annual | Notes |
|---|---|---|---|
| Pack wave operatives (8-12 cross-functional) | £18-28k | £216-336k | Midlands rates £26-28k, cross-trained across pack, staging, dispatch |
| Warehouse/fulfillment lead | £3k | £35-38k | Shift coordination, exception handling |
| Holiday/sickness cover + agency buffer | £4k | £50k | 15-20% cover ratio |
| Employer NI + pension on fulfillment staff | £4-5k | £50-60k | Statutory on-costs |
| Fulfillment consumables + equipment maintenance | £2-3k | £25-35k | Conveyor, labelling equipment, packing stations |
| Total Fulfillment | £31-42k | £376-519k | Model: £500k/yr |
Combined facility opex: £83,333/month = £1.0M/year (model base case). This is at the bottom of the central planning range recommended by external staffing analysis (£0.85-1.15M/year). Stress case is £1.2-1.4M/year if pack times are slower than planned or labour costs are higher. See Part 2B for the CPD methodology and fulfillment staffing rationale.
Key insight: Fixed costs determine how much profit you keep. They're the economics of scale. Phase C total facility opex (£1.0M/year) is offset by the manufacturing contribution uplift (£3.5-3.9M/year higher contribution at 30k customers versus the co-packer bridge). Net benefit after increased facility costs: £2.5-2.9M/year. This is still a compelling return on the £1.6M capex investment, with payback within 12-18 months.
Operating leverage at scale: Total fixed costs at 40k customers (~£538k/month = £6.5M/year) against £48M+ ARR give ~32% operating margin. A conventional DTC food operator at similar scale typically employs 30-40 non-production staff at ~£1.2M/year in people cost alone. Protocol Raw's AI-lean operating infrastructure (automated order processing, inventory allocation, CS triage via Kai, lifecycle automation, monitoring via SOP-MON-01) partially offsets the higher facility staffing, maintaining a meaningful cost advantage despite a larger facility team than originally modelled.
Part 2A: The Microbiological Risk Model¶
Source analysis: REF-MICRO-01 v2.0. Model location: Control Panel rows 84-122 (yellow editable cells) and the unified "VC Path" sheet in Protocol_Raw_Financial_Model.xlsx.
The prior version of this model calculated COGS as a per-kg rate applied to shipped volume, implicitly assuming zero production waste. In a category where the FSA's 2026 survey of 380 frozen raw pet food samples found 20.8% Salmonella prevalence and 28.7% overall statutory failure, that assumption is fiction. The v2.1 model accounts for it.
The Production Yield Model¶
Every COGS calculation now flows through this sequence:
Shipped kg = boxes shipped × 12.4 kg/box
Effective reject % = determined by phase + manufacturing state
Gross production = shipped kg ÷ (1 − reject %)
Reject loss kg = gross production − shipped
Reject cost = reject loss × COGS/kg (informational row only)
Poultry gate cost = monthly upstream screening (Phase A and in-house only)
Total COGS = gross production × COGS/kg + poultry gate cost
Contribution = revenue − total COGS − total CPD
Yield arithmetic matters. At an 18.6% reject rate, shipping 100 kg requires producing 100 ÷ (1 − 0.186) = 122.85 kg. You produce 22.85% more than you ship, not 18.6% more. This is the distinction between reject rate and overproduction rate.
No double-counting. The reject cost row in the model is informational. Total COGS already includes destroyed product because it is calculated on gross production, not shipped kg. An adversarial model review in April 2026 verified this.
The Reject Cost Taxonomy¶
Two distinct and additive rejection mechanisms. They must not be confused.
1. Upstream poultry gate rejects (raw ingredient destruction). Every incoming poultry delivery (~200 kg, one per production batch) is composite-sampled for Salmonella before entering production. 10-14 day UKAS lab turnaround. Positive packs destroyed at ~£350 each (raw ingredient value only). Planning rate: 15% of incoming poultry deliveries. Annual cost at Phase A volumes: ~£5,250 (£437.50/month). Applies only when Protocol Raw controls its own ingredient intake: Phase A self-manufacture and Phase C in-house. The co-packer will not run this gate for a single client.
2. Finished-batch rejects (hold-and-release failures). Every finished batch is quarantined at −18°C until UKAS lab results confirm it is clear. Batches failing on Salmonella, Listeria monocytogenes, or Enterobacteriaceae are destroyed at full COGS (~£2,500 per 500 kg batch: ingredients plus production plus packaging). The poultry gate reduces but does not eliminate these: residual Salmonella from the beef fraction, Enterobacteriaceae exceedances (partly green tripe driven), and environmental Listeria all remain.
Cost asymmetry. A rejected poultry pack costs ~£350. A rejected finished batch costs ~£2,500. The poultry gate's economic value is not just the Salmonella it catches; it catches it before the expensive value-add steps (grinding, mixing, filling, sealing, MAP, blast freezing).
Phase-Specific Reject Rates¶
| Phase | Months | Manufacturing | Finished-Batch Reject | Poultry Gate | COGS/kg |
|---|---|---|---|---|---|
| A | M0-M8 | Self-manufacture | 4% (Conservative) | Yes (£437.50/mo) | £4.31 early, improving |
| B Early | M9-M12 | Co-packer | 18.6% (Central) | No | £3.85 |
| B Late | M13-M22 | Co-packer | 18.6% (Central) | No | £3.30 |
| Parallel | M28-M33 | Blended (15% → 90% in-house) | 16.4% → 5.5% | Scales with in-house share | £3.52 → £2.90 |
| C | M34+ | In-house | 4% (Conservative) | Yes (at scale) | £2.90 |
Control Panel rows 84-122 expose all scenario alternatives as editable yellow cells:
- Phase A finished-batch: Low 2%, Conservative 4% (ACTIVE), Stress 9%
- Phase A poultry gate: 100 packs/year tested, 15% reject rate, £350/pack
- Phase B finished-batch (no gate): Conservative 12.4%, Central 18.6% (ACTIVE), Stress 25.0%
- HPP section: Retained for reference but disabled (month 99). Toll HPP is not practically available in the UK for this application.
Why Phase A Data Is the Most Valuable Asset¶
The VC Path model produces approximately 43 finished batches and 43 upstream poultry test results during Phase A (M0-M8). Zero failures across 43 batches gives 95% confidence that the true reject rate is below 6.7%. This is the empirical data that replaces planning estimates before Seed close.
Why In-House Is Required, Not Optional¶
The previous version of this document framed the Month 16-17 manufacturing decision as a choice between co-pack (Option A) and in-house (Option B). REF-MICRO-01 v2.0 removes that optionality.
Co-packing at 18.6% reject permanently caps contribution margin at ~29%. That is insufficient to fund aggressive acquisition, build cash reserves, or support the premium-multiple narrative. The only way to reinstate the poultry gate at scale is to own the facility, because the co-packer will not run per-delivery Salmonella screening for a single client. In-house manufacturing simultaneously cuts COGS/kg from £3.63 to £2.90 and cuts the reject rate from 18.6% to 4%. Co-pack is a funded bridge. In-house is the structural destination.
At scale, the poultry gate is operationally heavy but manageable. At 40,000 customers, approximately 600-800 poultry tests per month at £45 each (£27-36k/month) plus 10-15% rejected packs at £350 each (£21-42k/month). Total upstream cost of £50-80k/month on £4M+ monthly revenue is 2-3%. This is the normal operation of a food manufacturing plant that tests its ingredients.
Part 2B: Cost Per Delivery — Methodology and Evidence Base¶
Source analysis: Bottom-up CPD build (April 2026), validated against independent frozen DTC logistics benchmarks, UK parcel carrier volume pricing research, and Bella and Duke / Butternut Box published accounts. Model location: Control Panel rows 33-39 and the unified "VC Path" sheet row 36.
What CPD Includes¶
CPD is the fully loaded cost of getting a packed, insulated box from dispatch to the customer's door. It includes everything that happens after the product leaves the freezer and before the customer opens the box. Specifically:
- Packaging materials: outer carton, insulation liner (Woolcool or equivalent), PCM coolant packs, void fill, tape, labels, contents slip
- Pick/pack labour: the cost of assembling the customer's box from released inventory (imputed in Phase A, actual from Phase B)
- Courier/carrier: the all-in parcel delivery cost including base transport rate, fuel surcharge, residential delivery surcharge, heavy parcel surcharge, remote area surcharge (population-weighted national average), and peak season surcharge
- 3PL fees (Phase B only): storage, pick/pack service fees, accessorials, account management, integration fees (amortised)
- Trunking (Phase B only, if co-packer and 3PL are not co-located): the cost of moving finished frozen pallets between the co-packer and the 3PL
- Exception costs (amortised): the per-box share of failed delivery costs, including replacement shipments, returns handling, destruction of non-re-enterable stock, and customer service overhead. Amortised across all boxes at the prevailing exception rate
CPD does not include: COGS (raw materials, production, lab testing, reject overhead, poultry gate), customer acquisition costs, team salaries, facility costs (which are in manufacturing opex), or any SaaS/tooling costs.
The Decision Trail¶
The CPD values in this model were originally set as planning estimates in v1.0 (2024) without bottom-up validation. In April 2026, three independent research exercises were commissioned to validate or correct these estimates:
-
Frozen DTC 3PL and split-site logistics research. Examined the UK market for frozen DTC 3PLs, assessed the viability of a split-site model (manufacture in-house, fulfil via 3PL), built bottom-up CPD estimates for Phase B and Phase C, and benchmarked costs against volume tiers.
-
UK parcel carrier volume pricing research. Built all-in carrier cost estimates by volume tier for 17.4kg frozen DTC parcels, covering DPD, DHL Parcel UK, APC Overnight, Parcelforce, UPS, and Yodel. Separated direct-account vs 3PL pass-through pricing. Assessed service suitability for heavy frozen parcels.
-
Frozen DTC benchmarking study. Triangulated CPD benchmarks from public financial data (Companies House filings), investor reports, and operational case studies of UK frozen DTC operators including Butternut Box, Bella and Duke, Mindful Chef, Cook, and Allplants.
These research exercises produced a consistent finding: the original Phase C CPD of £10 was not achievable at Phase C volumes (approximately 10,000-12,000 drops/week) under any realistic parcel carrier plus 3PL model. The £10 target corresponds to the "Market Leader" tier (50,000+ drops/week), which Protocol Raw does not reach until well into Phase D.
A second finding changed the Phase C fulfillment model entirely. The research showed that 3PL overhead (pick/pack fees, storage, trunking, accessorials) adds approximately £3.50-4.50 per box. Since Protocol Raw is already building a manufacturing facility at Phase C with on-site cold storage, adding pick/pack capability is incremental rather than a new capital decision. Bringing pack in-house eliminates the 3PL margin, the trunking leg, storage fees, and all accessorials. Butternut Box validates this model: Rudie's Kitchen in Doncaster is a fully integrated manufacturing, cold storage, and fulfillment facility, not a factory that sends pallets to a 3PL.
The Phase B 3PL relationship is therefore a bounded engagement that ends at Phase C, not a continuation.
CPD Values in the Model¶
| Phase | Months | CPD | Evidence basis |
|---|---|---|---|
| Phase A early | M0-M3 | £22 | London, founder-dispatched, DPD at standard rates, conservative packaging spec. Unchanged from v1.0. |
| Phase A late | M4-M8 | £19 | Courier rate improvement (DPD business account negotiation), packaging volume savings, delivery density in target London postcodes. Unchanged from v1.0. |
| Phase B early | M9-M12 | £19 | 3PL pass-through carrier rates, co-packer/3PL transfer cost, Phase A packaging spec carried forward. Drops/week: 135-475 (Startup tier). Revised from £18 in v2.1. |
| Phase B late | M13-M21 | £18 | Scaling 3PL, improved carrier rates through volume, packaging optimisation from Phase A data. Drops/week: 1,030-2,310 (Growth tier boundary). Revised from £16 in v2.1. |
| Phase C | M22+ | £14 | In-house pack at own manufacturing facility, packaging optimisation at scale (390,000 boxes/year purchasing power), direct premium carrier account, dynamic seasonal packout. Drops/week: 2,860-11,230 (Growth/Scale-up tier). Revised from £10 in v2.1. |
| Phase D | M37+ | £11 | Further scale (50,000+ customers, 12,000-20,000 drops/week), continued packaging optimisation, potential own-van delivery in dense zones, mature carrier relationships. Revised from £8 in v2.1. |
Bottom-Up CPD Build: Phase B (M13-M21, 3PL Model)¶
Phase B uses an outsourced model: co-packer manufactures, 3PL handles pick/pack/dispatch. The 3PL also manages batch-segregated inventory (HOLD and RELEASED states) as part of the hold-and-release protocol.
| Component | Base estimate | Range | Source |
|---|---|---|---|
| Co-packer to 3PL transfer | £0.80 | £0.50-1.40 | Frozen pallet trunking research. Assumes co-packer and 3PL are in the same M1/East Midlands corridor but not co-located. If co-located, this drops to near-zero. |
| 3PL frozen storage | £0.60 | £0.45-0.90 | Converted from £12-22/pallet/week at approximately 50 box-equivalents per pallet with 2-2.5 weeks average dwell. |
| 3PL pick/pack (incl. PCM handling) | £3.25 | £2.75-4.25 | Frozen DTC pick/pack is more labour-intensive than ambient due to insulated box assembly and PCM placement. Single-SKU subscription boxes at the low end. |
| Packaging materials | £7.50 | £6.00-10.00 | Carton (£0.90), Woolcool liner (£2.20), PCM at £1.26/kg (£2.50), void fill/tape (£0.35), labels/inserts (£0.15). Season-dependent: COOL season at the low end, HOT season at the high end. Phase A prices; no volume discount yet. |
| Courier (all-in, 3PL pass-through) | £7.80 | £7.45-8.05 | DPD 17.4kg parcel at 8,000-12,000 parcels/month, 3PL pass-through pricing. Includes fuel, residential, remote area (population-weighted), peak surcharges. |
| Accessorials (amortised) | £0.35 | £0.20-0.60 | Account management, stock counts, integration fees, relabelling, amortised across all boxes. |
| Exception costs (amortised) | £0.90 | £0.50-1.50 | At 3% exception rate with £27 blended logistics-only cost per event. See Exception Cost Model below. |
| Total Phase B CPD | £21.20 | £18.85-27.10 |
The model uses £18 for Phase B late, which is at the optimistic end of the bottom-up range. This is justified by three factors: (1) if the co-packer and 3PL are co-located or the same entity, the transfer cost drops to near-zero; (2) packaging will have been optimised with 4-8 months of Phase A logger data by this point; (3) the 3PL relationship will have been competitively tendered at the Growth tier threshold. The £18 assumption requires at least two of these three factors to materialise. If none materialise, Phase B late CPD is more likely £20-21.
Bottom-Up CPD Build: Phase C (M34+, In-House Pack)¶
Phase C uses an integrated model: Protocol Raw manufactures, stores, picks, packs, and dispatches from its own facility. The 3PL relationship from Phase B ends at Phase C. All fulfillment is in-house.
| Component | Base estimate | Range | Source |
|---|---|---|---|
| Packaging materials | £5.50 | £4.50-6.50 | Same components as Phase B but with three improvements: (1) volume purchasing at 390,000 boxes/year reduces unit costs on cartons, liners, and PCM; (2) dynamic seasonal packout reduces average PCM quantity (less coolant needed in winter, weighted average across the year); (3) potential liner alternative (e.g., Ranpak RecyCold paper-based liner) if validated. |
| Courier (all-in, direct account) | £6.70 | £6.20-7.20 | DPD or DHL Parcel UK, 17.4kg parcel at 25,000-40,000 parcels/month, direct premium carrier account (no 3PL pass-through markup). East Midlands dispatch origin (DPD main UK hub is in Hinckley). Includes all surcharges. |
| Pick/pack labour (in-house) | £2.00 | £1.50-2.25 | 5-minute pack time at purpose-built stations, Midlands loaded labour rate ~£17/hr. Phase A data shows 9-minute target at £15/hr London rate; scale and purpose-built stations improve both rate and time. |
| Exception costs (amortised) | £0.75 | £0.40-1.00 | At 2.5% exception rate (improved from Phase B through better packaging and delivery messaging) with £27 blended cost per event. See Exception Cost Model below. |
| Total Phase C CPD | £14.95 | £12.60-16.95 |
The model uses £14, which is slightly below the base case. This requires moderate optimisation across all four components, not heroic improvement on any single one. The most likely path to £14 is: packaging at £5.00 (volume pricing plus dynamic packout), courier at £6.50 (competitive tender with DHL as lever), pick/pack at £1.75 (7-minute pack time), and exceptions at £0.50 (strong delivery comms and good packaging performance). Total: £13.75.
Why In-House Pack at Phase C¶
The original model assumed fulfillment would remain outsourced to a 3PL at Phase C, with the in-house facility handling manufacturing only and trunking released pallets to the 3PL. Three research findings changed this decision:
1. The 3PL overhead is structurally expensive. Pick/pack (£2.85-3.25), storage (£0.45), trunking (£0.35), and accessorials (£0.30) add approximately £3.95-4.35 per box. This is money paid to a third party to do something you could do in a building you already occupy.
2. The transfer leg creates operational complexity for no strategic benefit. Split-site manufacturing and fulfillment introduces a new inventory state (released stock in transit), new failure modes (pallets not booked in, FEFO breaks, ghost inventory, dock dwell time), and new monitoring requirements (transfer SOP, pallet-level tracking, receiving SLAs). All of this complexity exists only because the product moves between two buildings. Eliminating the transfer by packing on-site removes an entire class of operational risk.
3. The facility already has most of what pick/pack requires. The Phase C manufacturing facility includes on-site cold storage (for raw materials, quarantined WIP, and released finished goods), a loading bay (for ingredient deliveries), staff facilities, and a food-grade production environment. Adding packing stations, a dispatch staging area, and a DPD collection arrangement is incremental. Estimated additional capex: £50-100k on top of the £1.6M manufacturing build. Estimated additional headcount: 8-12 cross-functional warehouse/fulfillment operatives on top of the production staff already budgeted.
4. Butternut Box validates the integrated model. Rudie's Kitchen (Doncaster, opened March 2021) is a fully integrated manufacturing, cold storage, and fulfillment facility. Butternut Box does not send pallets to a 3PL. They manufacture, pack, and dispatch from one site. At 60,000+ drops/week and £126.7M revenue (FY2023), this is the proven model at scale in UK pet food DTC. They subsequently expanded with a second unit on the same industrial park.
5. The Phase B 3PL relationship becomes a bounded engagement. In Phase B, the 3PL is needed because the co-packer manufactures and someone must handle pick/pack. At Phase C, when manufacturing moves in-house, pack moves in-house at the same time. The 3PL relationship ends cleanly. No contract renegotiation for a new inbound model, no split-site monitoring, no trunking SOP.
The CPD impact: In-house pack at Phase C saves approximately £2/box versus the outsourced model (£14.95 vs £16.95 at base case). At 390,000 boxes/year, that is approximately £780k/year. The saving compounds as volumes grow through Phase C.
Benchmark Validation¶
The CPD assumptions were validated against two independent benchmarking frameworks.
Volume-tier benchmarks (from frozen DTC benchmarking study):
| Tier | Drops/week | Benchmark CPD | Protocol Raw phase | Model CPD |
|---|---|---|---|---|
| Startup | Up to 1,000 | £22-30 | Phase B early (M9-M12): 135-475/wk | £19 |
| Growth | 1,000-10,000 | £15-22 | Phase B late / Phase C early: 1,030-6,140/wk | £18 / £14 |
| Scale-up | 10,000-50,000 | £11-16 | Phase C in-house (M34-M36): 9,890-11,230/wk | £14 |
| Market Leader | 50,000+ | £8-12 | Phase D: 12,000-20,000+/wk | £11 |
Phase B values sit within their respective benchmark ranges. The Phase C value of £14 is at the low end of the Growth range and within the Scale-up range. This is justified by the in-house pack advantage: the benchmark ranges assume outsourced fulfillment, which adds £3-4/box of 3PL overhead. Removing that overhead places an in-house operator at the bottom of their volume tier or into the tier below.
Phase D at £11 requires reaching the Scale-up tier, which the model achieves at approximately 12,000-20,000 drops/week (50,000-80,000 customers). This is at the upper end of the Scale-up range and assumes further packaging optimisation, mature carrier relationships, and potential own-van delivery in dense zones.
Carrier pricing benchmarks (from UK parcel carrier research):
| Volume tier | DPD all-in (direct) | DHL all-in (direct) | Model assumption |
|---|---|---|---|
| 3,000-5,000/month | £8.08 | £7.86 | Phase B early: included in £19 total |
| 8,000-12,000/month | £7.29 | £7.04 | Phase B late: included in £18 total |
| 25,000-40,000/month | £6.55 | £6.31 | Phase C: £6.70 in bottom-up build |
| 100,000+/month | £5.84 | £5.67 | Phase D context: floor reference |
The Phase C courier assumption of £6.70 is between the DPD (£6.55) and DHL (£6.31) rates at the 25,000-40,000/month tier, reflecting a realistic contract without assuming best-case negotiation. DHL provides a competitive lever against DPD in rate negotiations.
Comparable Company Evidence¶
Bella and Duke (raw frozen dog food, DTC subscription, UK):
| Year | Revenue | Gross margin | Manufacturing |
|---|---|---|---|
| 2021 | £11.2M | 31.3% | Outsourced |
| 2022 | £19.3M | 44.3% | In-house (£4M invested) |
| 2023 | £22.9M | 46.6% | In-house, optimising |
| 2024 | £25.9M | 51.6% | In-house, optimising |
| 2025 | £27.7M | 54.0% | In-house, profitable |
Source: Companies House filings, strategic reports FY2021-FY2025.
Bella and Duke's gross margin (which likely includes distribution costs in cost of sales per UK statutory accounting practice) rose from 31% outsourced to 54% in-house. Protocol Raw's Phase C contribution margin of 45.9% is consistent with this trajectory, sitting between Bella and Duke's 44.3% (first year of in-house, 2022) and their 51.6% (third year, 2024). The margin gap is partly explained by Bella and Duke's higher volumes (10,000-15,000 drops/week vs Protocol Raw's 10,000-12,000 at Phase C steady state) and partly by cumulative operational improvements over three years that Protocol Raw has not yet achieved.
Butternut Box (cooked frozen dog food, DTC subscription, UK + EU):
Butternut Box's gross margin improved from 37.9% (FY2022) to 47.5% (FY2023) as revenue grew 82% to £126.7M. Source: Companies House filings, strategic report. The company operates from a fully integrated manufacturing and fulfillment facility (Rudie's Kitchen, Doncaster) with approximately 100 fulfillment staff across three shifts and 60,000+ drops/week. Their facility model, geographic location (South Yorkshire / M1 corridor), and carrier choice (DPD primary) validate Protocol Raw's Phase C design.
Exception Cost Model¶
Failed deliveries, returns, and temperature excursions are amortised across all boxes as a CPD component. The model uses three severity tiers:
Low-severity (missed first delivery, still delivered within thermal window): Customer service cost of £1-3 per event. No product loss, no replacement shipment.
Medium-severity (delay or damage requiring inspection, no replacement): Returns handling (£2-6), quarantine/assessment (£1-5), destruction if needed (£3-8). Total: £6-19 per event.
Full replacement (parcel lost, thawed, or unsafe): Sunk outbound freight (£9), replacement pick/pack (£2.50-4.25), replacement packaging (£6-10), replacement courier (£9), returns/destruction/admin (£3-10). Total logistics-only cost: approximately £30-42 per event. Product replacement value sits outside CPD.
Blended cost per event: £27 (logistics-only working assumption, weighted across severity tiers).
Aggregate CPD impact by exception rate:
| Exception rate | CPD adder per box |
|---|---|
| 1.5% | £0.41 |
| 2.0% | £0.54 |
| 2.5% | £0.68 |
| 3.0% | £0.81 |
| 5.0% | £1.35 |
Phase B assumes 3% exception rate (£0.81/box, rounded to £0.90 in the build). Phase C assumes 2.5% exception rate (£0.68/box, rounded to £0.75 in the build). The improvement reflects better packaging (optimised with 18+ months of logger data), improved delivery messaging (SMS with safe-place instructions), and address validation at checkout. Consumer-returned frozen raw pet food is treated as destroy/non-re-enterable by default unless continuous custody and temperature evidence exist end-to-end.
Sensitivity Analysis¶
What moves CPD most (Phase C, per £1 change in each component):
| Component | Share of CPD | £1 change = | Controllability |
|---|---|---|---|
| Courier | 45% (£6.70) | £1.00 CPD change | Low. Driven by carrier market, volume, and negotiation. Floor approximately £5.50-6.00 at Protocol Raw's Phase C volumes. |
| Packaging | 37% (£5.50) | £1.00 CPD change | High. Directly controllable through material selection, supplier negotiation, and dynamic packout. Biggest lever. |
| Pick/pack | 13% (£2.00) | £1.00 CPD change | Medium. Driven by pack time (improvable with station design and practice) and labour rate (location-dependent). |
| Exceptions | 5% (£0.75) | £1.00 CPD change | Medium. Driven by packaging quality, carrier performance, and customer communication. Circular: better packaging reduces exceptions. |
CPD sensitivity at Phase C (12kg box, in-house pack):
| Scenario | Packaging | Courier | Pick/pack | Exceptions | Total CPD |
|---|---|---|---|---|---|
| Conservative | £6.50 | £7.20 | £2.25 | £1.00 | £16.95 |
| Base case | £5.50 | £6.70 | £2.00 | £0.75 | £14.95 |
| Model assumption | -- | -- | -- | -- | £14.00 |
| Optimistic | £4.50 | £6.20 | £1.50 | £0.50 | £12.70 |
The model assumption of £14 requires moderate optimisation across all four components. The conservative case (£16.95) represents no optimisation beyond Phase A baselines. The optimistic case (£12.70) represents strong execution across all levers. The range is £12.70-16.95, with £14 as the planning midpoint.
Impact on contribution margin:
| CPD | Contribution/box (Phase C in-house) | Contribution margin |
|---|---|---|
| £12 | £44.50 | 48.1% |
| £14 | £42.50 | 45.9% |
| £16 | £40.50 | 43.8% |
| £18 | £38.50 | 41.6% |
Even at the conservative end (£16.95), Phase C in-house contribution margin is 40.6%, still materially above Phase B late (36.3%) and above Bella and Duke's first-year in-house gross margin of 44.3%. The manufacturing transition case is robust across the entire CPD sensitivity range.
Key Levers for CPD Improvement¶
Packaging (biggest controllable lever): Every £1 off packaging drops CPD by £1. The path from £7.50 (Phase A) to £5.50 (Phase C base) to £4.50 (optimistic) requires: volume purchasing at 390,000 boxes/year (cartons, liners, PCM all cheaper at scale), dynamic seasonal packout (SOP-PACK-01 with weather-based PCM calculation means less coolant in winter), and 18+ months of Phase A/B logger data enabling right-sizing of thermal protection by lane and season.
Carrier negotiation (biggest single cost, hardest to move): Run DPD against DHL Parcel UK. Single origin, uniform parcel profile, predictable Mon-Fri dispatch, East Midlands location (DPD's main UK hub is Hinckley) all support strong negotiation. Own carrier account from Phase C (no 3PL pass-through markup). Realistic improvement from £6.70 to £6.20-6.40 with competitive tender.
Pack time (smaller line, compounding): Phase A 9-minute target drops to 5-6 minutes at scale with purpose-built stations, conveyor-fed staging, and a practiced team packing uniform subscription boxes. At Phase D volumes, goods-to-person automation (AMRs, as deployed by Butternut Box via Reeco) could further reduce pack time and labour cost.
Exception rate (the hidden lever): The difference between 3% and 1.5% exception rate is £0.41/box, or £160k/year at 390,000 boxes. Primary levers: delivery messaging (SMS with one-hour window and safe-place instructions), address validation at order creation, carrier selection (DPD's tracking and consumer communication capabilities), and packaging quality (better thermal protection means fewer thaw complaints).
What Remains Uncertain¶
The following items are classified as vendor-quote-only and cannot be resolved without direct commercial engagement. They are flagged here so that actual data from Phase A and Phase B trading replaces planning estimates before Phase C capital is deployed.
Carrier rates. All carrier pricing in this section is informed estimates. Exact net tariffs, fuel formula, remote postcode matrices, peak surcharge regimes, and collection cut-off from the specific dispatch origin are vendor-quote-only items. Phase A trading will produce the first real carrier invoice data.
Packaging costs at volume. Unit costs for cartons, liners, and PCM at 390,000 boxes/year are estimates based on typical volume discount curves. Actual supplier quotes at Phase C volumes will be obtained during the manufacturing facility planning phase (M22-M23).
3PL pricing (Phase B). All 3PL costs (pick/pack, storage, accessorials) are informed estimates. Phase B 3PL selection and contract negotiation will produce actual rates. The recommended RFQ process covers 50 fields across site/regulatory, WMS/batch-state controls, inbound/split-site support, cold-chain evidence, carrier/exception handling, commercial terms, and future-state compatibility.
Exception rates. The 2.5-3% exception rate assumption is an industry estimate for frozen DTC. Actual rates will be measured from M1 onwards through the courier watchdog system (SOP-DLV-01) and will be available as calibration data well before Phase C.
Phase B 3PL: Selection Criteria and Phase C Awareness¶
The Phase B 3PL selection is the nearest-term commercial decision arising from this analysis. Selection criteria, in priority order:
- Can demonstrably run frozen DTC parcel fulfillment from -18C storage
- Can handle raw pet food / animal by-products (ABP site approval)
- Can prove blocked HOLD stock cannot allocate into pick waves (WMS-native or auditable compensating controls)
- Can trace batch from inbound pallet to customer dispatch
- Can produce structured cold-chain evidence data (CSV/API, not PDFs)
- Will contract flexibly enough for Phase C decision point (12-month initial term, 90-180 day notice, no exclusivity, full data portability, exit assistance)
The 3PL contract at Phase B is a bounded relationship that ends at Phase C. Contract terms must protect optionality: no hard volume locks, no exclusivity, explicit right to re-tender at M27/M28 without penalty. Jamie (commercial advisor) leads 3PL and co-packer negotiations. The recommended procurement approach is a narrow RFQ to 4-6 frozen logistics specialists.
Facility Specification Impact¶
The Phase C facility specification in the Business Plan must reflect the integrated manufacturing and fulfillment model. The facility requires: production lines (grinding, mixing, filling), blast freezer, cleared-goods cold storage, upstream poultry gate infrastructure (composite sampling station, segregated NOT CLEARED / CLEARED quarantine freezer zones), pick/pack stations, dispatch staging area, and outbound carrier collection arrangement. It does not require a full warehouse operation or the infrastructure for a split-site transfer leg.
Estimated additional capex for pick/pack capability: £50-100k on top of the £1.6M manufacturing build (packing stations, conveyors, dispatch staging, labelling equipment). Estimated additional headcount: 8-12 cross-functional warehouse/fulfillment operatives on top of the production staff already budgeted (Midlands salary levels, £26-28k).
Part 3: The P&L Waterfall¶
Revenue flows through five levels. Each level answers a different question.
Revenue (£111 inc VAT)
│
├── VAT (£18.50) → Goes to HMRC
│
▼
Gross Revenue (£92.50 ex-VAT)
│
├── COGS on shipped kg (base production cost)
├── Reject overhead (gross production = shipped ÷ (1 − reject %))
├── Poultry gate cost (Phase A and Phase C only)
│ → Can we make money on each box after honest testing?
▼
Gross Profit (£34-48 by phase)
│
├── CPD (£14-22) → Can we deliver profitably?
│
▼
Contribution (£15 → £26 → £43 by phase)
│
├── CAC (amortised) → Can we grow profitably?
│
▼
Post-CAC Margin
│
├── Fixed Costs (allocated) → Can we operate profitably?
│
▼
Operating Profit
What Each Margin Tells You¶
| Margin | Question It Answers | Target |
|---|---|---|
| Gross Margin | Is the product viable? | >40% at scale |
| Contribution Margin | Is the unit economics viable? | >40% at scale |
| Post-CAC Margin | Is growth sustainable? | Positive by Month 11 |
| Operating Margin | Is the business viable? | >30% at scale |
Part 4: Unit Economics (Per-Box)¶
The economics of a single box, which is the atomic unit of the business.
All unit economics below include reject overhead and poultry gate cost where applicable. Contribution margin figures are sourced from explainer Section 5 (VC Path sheet).
Phase A (Self-Manufacture, M0-M8)¶
| Line | Launch (M1) [CHECK] | M8 End-state [CHECK] |
|---|---|---|
| Revenue (ex-VAT) | £92.50 | £92.50 |
| COGS on shipped (£4.31 → improving × 12.4 kg) | (£53.44) | (£44.00) |
| Reject overhead (4% × 1.0417) | (£2.23) | (£1.83) |
| Poultry gate (allocated per box) | (£4.00) | (£1.17) |
| Total COGS incl. micro risk | (£59.67) | (£47.00) |
| Gross Profit | £32.83 | £45.50 |
| CPD | (£22.00) | (£19.00) |
| Contribution | £10.83 | £26.50 |
| Contribution Margin | 11.7% | ~27.5% |
Per-box figures [CHECK] against VC Path sheet M1 and M8 rows. The headline number is the contribution margin trajectory: 13.1% at M1 → 27.3% at M6 → 27.5% at M8 (explainer Section 5). Phase A runs at a 4% reject rate with the poultry gate active at £437.50/month (allocated across a small monthly box count, so per-box cost drops from ~£4 at M1 to ~£1.17 at M8). CPD improves through Phase A as batch efficiency and delivery density increase. Operationally breakeven at M8.
Phase B Early (Co-Packer Bridge, M9-M12)¶
| Line | Amount | % of Revenue |
|---|---|---|
| Revenue (ex-VAT) | £92.50 | 100% |
| COGS on shipped (£3.85/kg × 12.4 kg) | (£47.74) | 51.6% |
| Reject overhead (18.6% × 1.2285 uplift) | (£10.91) [CHECK] | 11.8% |
| Poultry gate | £0 | 0% |
| Total COGS incl. micro risk | (£58.65) [CHECK] | 63.4% |
| Gross Profit | £33.85 [CHECK] | 36.6% |
| CPD | (£19.00) | 20.5% |
| Contribution | £14.85 | 16.1% |
Contribution margin 16.1% is the worst point in the trajectory. Source: VC Path sheet M9-M12. This is the cost of honest testing without ingredient control. The co-packer will not run per-delivery Salmonella screening, so there is no poultry gate. The 16.1% CM is the cost of the funded bridge, not a permanent state. CPD at £19 reflects 3PL pass-through carrier rates at startup volumes (135-475 drops/week). See Part 2B for CPD methodology and Part 8 for the four-phase manufacturing transition.
Phase B Late (Co-Packer Bridge, M13-M22)¶
| Line | Amount | % of Revenue |
|---|---|---|
| Revenue (ex-VAT) | £92.50 | 100% |
| COGS on shipped (£3.30/kg × 12.4 kg) | (£40.92) | 44.2% |
| Reject overhead (18.6% × 1.2285 uplift) | (£9.35) [CHECK] | 10.1% |
| Poultry gate | £0 | 0% |
| Total COGS incl. micro risk | (£50.27) [CHECK] | 54.3% |
| Gross Profit | £42.23 [CHECK] | 45.7% |
| CPD | (£18.00) | 19.5% |
| Contribution | £24.23 | 26.2% |
Contribution margin improves from 16.1% to 26.2% between Phase B Early and Phase B Late. The improvement is driven by COGS/kg (£3.85 → £3.30, co-packer volume leverage), not by any change in reject rate or significant CPD improvement. The 18.6% reject rate is unchanged through the entire co-packer phase. CPD moves only from £19 to £18 as scaling 3PL and carrier volume discounts take effect. Operating profitability is reached at M15 despite the waste. See Part 2B for CPD methodology.
Phase C (In-House, M34+)¶
| Line | Amount | % of Revenue |
|---|---|---|
| Revenue (ex-VAT) | £92.50 | 100% |
| COGS on shipped (£2.90/kg × 12.4 kg) | (£35.96) | 38.9% |
| Reject overhead (4% × 1.0417 uplift) | (£1.50) [CHECK] | 1.6% |
| Poultry gate (allocated per box at 40k scale) | (£0.67) [CHECK] | 0.7% |
| Total COGS incl. micro risk | (£38.13) [CHECK] | 41.2% |
| Gross Profit | £54.37 [CHECK] | 58.8% |
| CPD | (£14.00) | 15.1% |
| Contribution | £42.50 | 45.9% |
Two structural changes compound to produce the 45.9% contribution margin: (1) COGS/kg drops from £3.63 to £2.90 through manufacturing internalisation, and (2) the reject rate drops from 18.6% to 4% because the poultry gate is reinstated. The poultry gate at 40k customers costs ~£50-80k/month against £4M+ monthly revenue (2-3%). CPD at £14 reflects in-house pack at the manufacturing facility (eliminating 3PL margin, trunking, and accessorials), packaging optimisation at 390,000 boxes/year purchasing power, and direct carrier account. See Part 2B for the full bottom-up CPD build, benchmark validation, and sensitivity analysis.
Margin trajectory summary: £14.85 → £24.23 → £42.50 contribution per box. The jump from Phase B Late (26.2%) to Phase C In-House (45.9%) is approximately 20 points, driven by reject rate reduction (18.6% → 4%), COGS/kg reduction (£3.30 → £2.90), and CPD reduction (£18 → £14), each a distinct structural change. Do not present this as "we fixed manufacturing and margin jumped 20 points" — oversimplification will make investors suspicious. The improvement comes from reject rate reduction, COGS/kg reduction, CPD reduction through in-house pack, and fixed cost leverage.
Why Margins Improve¶
The contribution margin trajectory (27% → 16% → 26% → 46%) has four distinct drivers. Oversimplifying to any one of them is misleading.
1. COGS/kg (£4.31 → £3.85 → £3.30 → £2.90):
Phase A (self-manufacture, £4.31/kg early, improving): - Founder-operated APHA-approved facility, small batches (200-500 kg) - Direct ingredient sourcing, learning curve on production efficiency - Lab testing amortised across smaller batches
Phase B Early → Late (co-packer, £3.85/kg → £3.30/kg): - Transition to hybrid co-pack model (co-packer sources bulk ingredients to spec) - Volume leverage with co-packer (1,000 kg → 2,500 kg batches) - Raw material purchasing power improves at scale
Phase C (in-house, £2.90/kg): - Eliminates co-packer margin (20-25% of COGS) - Full control over production efficiency and supplier contracts - Founder has Phase A production experience to draw on
2. Reject rate (4% → 18.6% → 18.6% → 4%):
The reject rate is the hidden margin driver the old model ignored. Phase A runs at 4% because the poultry gate screens every incoming delivery. Phase B co-pack runs at 18.6% because the co-packer will not run the gate for a single client. Phase C in-house returns to 4% because the facility reinstates the gate at scale. Every COGS/kg gain in Phase B is partially destroyed by the 22.85% overproduction requirement.
3. CPD (£22 → £19 → £18 → £14):
CPD improvement is driven by four structural changes across the phases: - Phase A → Phase B: packaging optimisation from logger data, DPD business account negotiation - Phase B Early → Phase B Late: scaling 3PL, carrier volume discounts, delivery density improvement - Phase B → Phase C: in-house pack eliminates 3PL margin (£2.85-3.25/box), trunking (£0.35), storage (£0.45), and accessorials (£0.30); packaging optimisation at 390,000 boxes/year; direct carrier account vs 3PL pass-through - Phase C → Phase D: further carrier rate improvement at higher volumes, packaging maturity, potential own-van delivery in dense zones
See Part 2B for the complete CPD methodology, bottom-up builds, benchmark validation, and sensitivity analysis.
4. Fixed cost leverage:
Operating margin moves from −34% at M9 (Seed close) to +32% at M36 (steady state). A large share of that improvement is revenue scaling against a lean, AI-automated fixed cost base. However, the Phase C facility team is larger than originally modelled: 20-24 facility FTE (production + fulfillment) versus the original assumption of 12. This reflects the decision to bring pick/pack in-house, which improves CPD by £2/box but adds ~£500k/year in fulfillment labour. The net effect is positive: the CPD saving (£780k/year at 390k boxes) exceeds the labour cost.
The manufacturing transition is the structural fix, not an optimisation. The prior version of this document framed Month 16-17 as a choice between Option A (continue co-packing) and Option B (build in-house). REF-MICRO-01 v2.0 removes that choice. At 18.6% reject without the poultry gate, co-packing permanently caps contribution margin at ~26%. The only way to reinstate the gate at scale is to own the facility. See Part 8.
Part 5: Lifetime Value (LTV)¶
LTV measures the total economic value a customer generates over their relationship with Protocol Raw. We compute three variants because they answer different questions.
The Three LTVs¶
Revenue LTV = Revenue/box × Boxes/month × Avg Lifespan (months)
= £92.50 × 1.083 × 13.73
= £1,375
Gross Profit LTV = Gross Profit/box × Boxes/month × Avg Lifespan (months)
(phase-specific, see table below)
Contribution LTV = Contribution/box × Boxes/month × Avg Lifespan (months)
(phase-specific, see table below)
Contribution LTV by phase. The business does not have a single LTV. The manufacturing transition changes contribution per box by a factor of ~2.8 over the VC Path trajectory, so the LTV trajectory is the right framing for investors.
| Phase | Contribution/box | Contribution LTV | Gross Profit LTV [CHECK] |
|---|---|---|---|
| Phase B Early (M9-M12) | £15.85 | £221 | ~£536 |
| Phase B Late (M13-M22) | £26.23 | £360 | ~£670 |
| Phase C In-House (M34+) | £45.04 | £594 | ~£862 |
LTV = Contribution/box × 1.083 boxes/month × 13.73 months (70% Box-2 × 5.5% steady-state churn base case, per Business Plan v3.4 and Growth Strategy v3.2). Contribution/box figures from explainer Section 6 and the VC Path sheet. Revenue LTV is £1,375 (was £1,466 at 75% Box-2).
Which LTV to use where:
| LTV Variant | Value | Use For |
|---|---|---|
| Revenue LTV | £1,375 | Market sizing, top-line narrative |
| Gross Profit LTV | ~£536 → £862 by phase [CHECK] | Product viability assessment |
| Contribution LTV | £221 → £360 → £594 by phase | LTV:CAC ratio, payback period, growth sustainability claims |
This document uses Contribution LTV as a phase trajectory, not a single number. The previous v2.0 value of £424 implicitly assumed zero production waste; it is not a defensible number under REF-MICRO-01 v2.0. Always cite the trajectory: £221 → £360 → £594.
The Components¶
Revenue per box: £92.50 (ex-VAT, 12kg box)
Boxes per month: 1.083 (4-week cycle = 13 boxes/year ÷ 12 months)
This is theoretical cadence. The Business Plan uses a conservative "Realised ARPU" of £1,040 ex-VAT annually (~11.5 boxes/year) to account for cadence drift. See Cadence Drift Sensitivity below.
Average Lifespan: 13.73 months - Box-2 retention: 70% (modelled rate, per Business Plan v3.4 and Growth Strategy v3.2) - Steady-state churn post-Box-2: 5.5% - Expected lifespan = 1 + 0.70 × (1 ÷ 0.055) = 13.73 months - Implied blended monthly churn: ~7.3%
This is the Financial Model spreadsheet base case (LTV & Sensitivity sheet). It combines a 70% Box-2 retention rate (aligned with Business Plan v3.4 and Growth Strategy v3.2) with 5.5% monthly churn on subscribers past Box-2. If the Growth Strategy's operational target of <4% post-Box-2 churn is achieved, lifespan extends further and Contribution LTV would be significantly higher — but we model and fundraise on the base-case assumption.
Two-Layer Churn Model¶
Churn at Protocol Raw operates in two distinct layers. Conflating them leads to confusion in cross-document reading.
| Metric | Definition | Target | Used For |
|---|---|---|---|
| Box-2 Retention | % of new customers who order Box-2 | ≥70% | Product-market fit signal, cohort quality |
| Steady-State Churn | Monthly churn rate on subscribers who have received ≥2 boxes | <4% (operational target) | Operational dashboard, intervention triggers |
| Blended Churn | Effective monthly churn across all cohort months (including Box-1 attrition) | ~7.3% (base case at 70% Box-2 / 5.5% SS) | LTV calculation, financial modelling |
The mathematical relationship:
Expected lifespan = 1 month + (Box-2 Retention × (1 ÷ Steady-State Churn))
Blended monthly churn = 1 ÷ Expected lifespan
Churn Sensitivity (Two-Layer)¶
Contribution LTV columns below are computed at Phase B Early contribution of £15.85/box (conservative bridge-phase basis). To re-derive at Phase B Late (£26.23) or Phase C In-House (£45.04), multiply the Phase B Early column by 1.655 or 2.842 respectively.
| Box-2 Retention | Post-Box-2 Steady-State | Blended Churn | Avg Lifespan | Revenue LTV | Contribution LTV (Phase B Early) |
|---|---|---|---|---|---|
| 75% | 4.0% (GS target) | ~5.0% | 19.75 months | £1,979 | £339 |
| 75% | 5.5% (upside if retention beats target) | ~6.8% | 14.636 months | £1,466 | £251 |
| **70% | 5.5% (base case)** | ~7.3% | 13.73 months | £1,375 | £235 |
| 70% | 7.4% (conservative) | 9.5% | 10.5 months | £1,053 | £180 |
| 65% | 5.5% | 7.8% | 12.8 months | £1,283 | £220 |
| 60% | 6.0% | 8.7% | 11.5 months | £1,152 | £197 |
Key insight: A 1.5-point improvement in post-Box-2 churn adds ~£35-50 to Phase B Early Contribution LTV (~£60-85 at Phase B Late, ~£100-140 at Phase C). Retention is the highest-leverage metric in every phase. Row in bold is the Financial Model base case used throughout this document.
Cadence Drift Sensitivity¶
The LTV calculation above uses Theoretical ARPU (£1,202), assuming perfect 4-week cadence. In practice, cadence drift (holidays, skipped weeks, delayed reorders) reduces realised boxes per year. This is modelled as a separate risk dimension from churn:
| Cadence | Boxes/Year | Realised ARPU | Revenue LTV | Contribution LTV (Phase B Early, £15.85/box) |
|---|---|---|---|---|
| 4 weeks (perfect) | 13.0 | £1,202 | £1,375 | £235 |
| 4.5 weeks (slight drift) | 11.6 | £1,073 | £1,227 | £210 |
| 5 weeks (moderate drift) | 10.4 | £962 | £1,100 | £188 |
| 6 weeks (significant drift) | 8.7 | £805 | £921 | £158 |
Revenue LTV rounded from per-box × boxes/month × 13.73 months. Contribution LTV = £15.85/box × (boxes/year ÷ 12) × 13.73. For Phase B Late or Phase C In-House, multiply by 1.655 or 2.842 respectively.
The Financial Model does not stack cadence drift on top of conservative churn — that would double-count pessimism. Instead, each dimension is shown independently so investors can assess both risks.
Part 6: Customer Acquisition Cost (CAC)¶
CAC is what it costs to acquire one paying customer.
Organic acquisition is the structural CAC hedge. It includes coded referrals tracked via the SOP-REF-01 slug attribution system (targeting 10-12% of customers at steady state), plus attributed word-of-mouth, direct traffic, and branded search. Paid acquisition (Meta, Google, other) covers the remainder.
The Calculation¶
Blended CAC = (paid % × paid CAC) + (organic % × organic CAC)
At Phase A: (95% × £110) + (5% × £15) = £104.50 + £0.75 = £105
At Phase C: (75% × £85) + (25% × £10) = £63.75 + £2.50 = £66
Blended CAC Targets¶
| Phase | Blended CAC | Paid CAC | Organic % |
|---|---|---|---|
| Phase A (M0-M8) | £100-105 | £110 | 5% |
| Phase B Early (M9-M12) | £85-90 | £95-105 | 15-20% |
| Phase B Late (M13-M22) | £66-75 | £80-95 | 20-25% |
| Phase C (M28+) | £60-66 | £75-90 | 25-30% |
Why Organic Is the Real Moat¶
Every organic customer acquired at £10-12 (blended attribution cost) vs £100+ (paid) improves blended CAC. The 25% organic target by M22 (Series A) is strategic, not aspirational. The £66 blended CAC used in Part 7's LTV:CAC trajectory assumes Phase B Late / Series A-stage organic mix.
The Organic Composition¶
Organic share of 25% at Phase B Late decomposes approximately as follows:
- Coded referrals (SOP-REF-01 credit ledger): 10-12%
- Attributed word-of-mouth at signup: 8-10%
- Direct traffic and branded search: 3-5%
Coded referrals are tracked precisely through the referral slug system with fraud prevention, delivery-confirmed credit issuance, and a complete audit trail. Attributed word-of-mouth is captured through a "How did you hear about us?" dropdown at signup. Direct traffic and branded search are measured through standard web analytics attribution.
The coded referral component targets industry-benchmark levels (Gousto ~10%, HelloFresh ~8%, Butternut Box estimated 10-15%). The broader organic share of 25% is consistent with what Huel, Butternut Box, and Gousto report as organic or non-paid acquisition at similar maturity.
The Coded Referral Mechanism¶
Coded referrals are the trackable subset of organic acquisition. The mechanism:
- Referrer receives £10 credit after referee's first delivery confirmed
- Additional £10 density bonus (same-day, same-postcode)
- Referee receives £10-15 discount on first order
- Fraud prevention: delivery confirmation required, duplicate address detection, rate limiting
- Full audit trail via SOP-REF-01 slug attribution
Part 7: The LTV:CAC Ratio¶
The single most important metric for subscription businesses. We compute this on Contribution LTV — the value that actually funds growth after COGS and delivery.
The Phase-Specific Trajectory¶
The business does not have a single LTV:CAC. It has a trajectory driven by the manufacturing transition. All three points are real and all three should be shown to investors.
| Phase | Contribution LTV | Blended CAC | LTV:CAC | Payback |
|---|---|---|---|---|
| Phase B Early (worst point, M9-M12) | £221 | £66 | 3.4:1 | 3.8 months |
| Phase B Late (M13-M22) | £360 | £66 | 5.5:1 | 2.3 months |
| Phase C In-House (M34+) | £594 | £66 | 9.0:1 | 1.4 months |
Blended CAC of ~£66 assumes a 25% organic mix at Series A-stage economics. See Part 6. Payback = CAC ÷ (Contribution/box × 1.083 boxes/month).
Context at Phase B Early (worst point):
| Basis | LTV | LTV:CAC | What It Tells You |
|---|---|---|---|
| Revenue LTV | £1,375 | 20.8:1 | Top-line return (overstates, ignores costs) |
| Gross Profit LTV | ~£536 [CHECK] | ~8.1:1 | Product viability after COGS |
| Contribution LTV | £221 | 3.4:1 | Growth sustainability after COGS + delivery + reject |
Key insights:
- Cite the trajectory, not a single number. The 3.4:1 during the funded bridge is above the 3:1 threshold most investors require. The 9.0:1 at full in-house is exceptional.
- Protocol Raw's LTV:CAC improves structurally with scale. Most DTC food businesses watch this ratio erode with scale because CAC rises and margins stay flat. Protocol Raw's improves because the manufacturing transition simultaneously cuts COGS/kg and eliminates 18.6% production waste.
- The old single-number 5.3:1 figure (v2.0) is superseded. It implicitly assumed zero production waste and is not defensible under REF-MICRO-01 v2.0.
Payback Period¶
Payback trajectory tracks contribution per box: 3.8 months (Phase B Early) → 2.3 months (Phase B Late) → 1.4 months (Phase C In-House). Even at the worst point of the funded bridge, payback is below the 6-month threshold typical for DTC food subscriptions. Phase A payback is longer (~6+ months) because of higher CAC and lower volume; Phase A does not participate in the ratio framing used for Seed-stage onwards.
Uses theoretical 4-week cadence (1.083 boxes/month). With cadence drift (4.5-week actual), boxes/month drops to ~0.97, extending each payback by ~12%.
Part 8: The Manufacturing Transition (Financial Impact)¶
The prior version of this document framed Month 16-17 as a decision between co-packing (Option A) and in-house manufacturing (Option B). REF-MICRO-01 v2.0 removes that framing. The manufacturing transition is structural, required, and timed to the Series A close at M22. Co-packing is a funded bridge, not a permanent state.
The Four-Phase Trajectory¶
| Phase | Months | Manufacturing State | In-House Share | Reject Rate | COGS/kg |
|---|---|---|---|---|---|
| Self-Mfr | M0-M8 | 100% in-house (small-batch) | 100% | 4% | £4.31 → improving |
| Co-Pack | M9-M27 | 100% co-packer | 0% | 18.6% | £3.85 → £3.30 |
| Parallel | M28-M33 | Transition, ramping | 15% → 90% | 16.4% → 5.5% | £3.52 → £2.90 |
| In-House | M34+ | 100% in-house (scaled) | 100% | 4% | £2.90 |
Key dates:
- M22 Series A close (~£15M, ~10,877 customers, approximately £13M ARR)
- M23 facility build begins
- M28 parallel run begins (first in-house production alongside co-packer)
- M34 full in-house (co-packer offboarded)
Why This Is Not Optional¶
Co-packing at 18.6% reject permanently caps contribution margin at ~29%. That is insufficient to fund aggressive acquisition, build cash reserves, or support the premium-multiple narrative in investor materials. The only way to reinstate the poultry gate at scale is to own the facility, because the co-packer will not run per-delivery Salmonella screening for a single client.
In-house manufacturing delivers two simultaneous structural improvements:
- COGS/kg drops from £3.63 (blended co-pack) to £2.90 (manufacturing internalisation)
- Reject rate drops from 18.6% to 4% (poultry gate reinstated)
Combined effect: contribution margin from 26.2% to 45.9%. See Part 4 for the full Phase C unit economics.
Capex Schedule¶
The facility build runs M23-M27, five months, £1.6M total (plus ~£50-100k for pick/pack capability). Midlands or North location. APHA approval required. Includes segregated freezer zones for poultry quarantine, grinding / mixing / filling lines, blast freezer, sealing, cold storage, pick/pack stations, dispatch staging area, and outbound carrier collection arrangement. Phase C is a fully integrated manufacturing and fulfillment facility, validated by Butternut Box's Rudie's Kitchen model (Doncaster, 200,000 sq ft, opened March 2021).
| Month | Capex | Cumulative |
|---|---|---|
| M23 | £160,000 | £160,000 |
| M24 | £400,000 | £560,000 |
| M25 | £560,000 | £1,120,000 |
| M26 | £320,000 | £1,440,000 |
| M27 | £160,000 | £1,600,000 |
Blended Economics During Parallel Run (M28-M33)¶
Each month during the parallel run, more production moves to the in-house facility. The weighted COGS/kg and blended reject rate recalculate monthly:
Weighted COGS/kg = in_house_share × £2.90 + co_pack_share × £3.63
Blended reject rate = in_house_share × 4% + co_pack_share × 18.6%
Poultry gate cost = £437.50/month (Phase A baseline) × scale factor × in_house_share
| Month | In-House Share | Blended Reject | Weighted COGS/kg | Contribution Margin |
|---|---|---|---|---|
| M28 | 15% | 16.4% | £3.52 | 32.7% |
| M30 | 55% | 10.6% | £3.23 | 40.8% |
| M34 | 100% | 4.0% | £2.90 | 48.6% |
Financial Impact at Steady State (M34+)¶
- Capex: £1.6M total manufacturing + ~£50-100k pick/pack capability (fully invested by M27)
- Annual facility opex: £1.0M/year total (£500k manufacturing core + £500k fulfillment/warehouse). See Part 2 for full breakdown.
- Annual contribution uplift vs co-pack bridge: £3.5-3.9M at 30k customers, rising to £5.8-6.5M at 50k customers
- Net annual benefit (after increased facility costs): £2.5-2.9M/year at 30k customers
- Payback on capex + first-year opex: 12-18 months
Competitive Precedent¶
Both Bella & Duke and Butternut Box built in-house facilities at similar revenue to Protocol Raw's Series A position. This is not a novel move; it is the category-standard path. Both chose fully integrated manufacturing + fulfillment facilities, not factory-only with outsourced 3PL fulfillment.
Bella & Duke (FY2021-25, Companies House filings): - Revenue at decision: ~£11-19M - Invested: ~£4M in manufacturing and fulfillment facilities (34,000 sq ft, Inverkeithing, Fife) - Gross margin trajectory: 31% (FY2021, outsourced) → 44% (FY2022, first year in-house) → 47% (FY2023) → 52% (FY2024) → 54% (FY2025) - Reached EBITDA profit FY2023, operating profit FY2025 (0.7% margin on £27.7M revenue) - 134 employees at FY2025 - Distributes nationally via DPD from single Scottish facility using dry ice
Butternut Box (2020-24, Companies House filings + public sources): - Revenue at decision: ~£10-20M - Built: Rudie's Kitchen (200,000 sq ft integrated manufacturing + fulfillment facility, Doncaster/Blyth) - Gross margin: 37.9% (FY2022) → 47.5% (FY2023) as revenue grew 82% to £126.7M - Subsequently expanded with second 133,606 sq ft unit on same industrial park (20-year lease) - ~100 fulfillment staff across three shifts, cross-trained warehouse operatives on 12-hour rotating patterns - Deployed AMR automation (Reeco) for goods-to-person at packing stations - Distributes nationally via DPD from single site
Protocol Raw's Phase C contribution margin of 45.9% sits between B&D's first year (44%) and third year (52%) of in-house operation, consistent with Protocol Raw's lower volumes and earlier stage of optimisation.
Cost Basis¶
Capex estimates are based on UK industry benchmarks adjusted for Protocol Raw's phased approach:
- Phase C facility sized for ~30k customers (8,000-15,000 sq ft)
- Single-shift operation with expansion optionality
- Expansion to Phase D funded separately (Series B)
Detailed facility planning, equipment specification, and site selection completes during M22-M23, with specialist validation during due diligence for Series A.
See Business Plan v3.0 Appendix: In-House Manufacturing Cost Methodology for full breakdown.
Part 9: The Investor Narrative¶
How to talk about these numbers.
The One-Sentence Version¶
"We raise £1.15M Seed to fund the co-packer bridge including its reject cost burden, and £15M Series A at M22 to build an integrated manufacturing and fulfillment facility whose poultry gate is the structural fix for the 18.6% reject rate the category carries."
The Three-Minute Version¶
"The raw pet food category has a safety problem that most operators don't know about because they don't test. The FSA's 2026 survey of 380 samples found 28.7% statutory failure across 50 brands.
Protocol Raw tests every batch. That means we discover and destroy contaminated product rather than shipping it unknowingly. That costs money: 18.6% of production during the co-packer phase.
We found this before launch, not after. We built a 50-page microbiological risk analysis, detection maths, and three reject-rate scenarios costed against our specific formulation. The co-packer phase is an ugly but funded bridge. Seed pays for it. Contribution margin is 16-26% on the bridge. Operating profitability is reached at M15.
The structural fix is an integrated manufacturing and fulfillment facility with upstream ingredient screening. That restores the 4% reject rate we achieve in Phase A self-manufacture, combined with a lower COGS/kg and in-house pack that eliminates 3PL margin. Contribution margin reaches 46%.
LTV:CAC improves structurally from 3.4:1 during the bridge to 9.0:1 at in-house steady state. Most DTC food businesses watch this ratio erode. Protocol Raw's improves because the manufacturing transition simultaneously cuts COGS and eliminates production waste.
The facility model is validated by both Bella and Duke (gross margin 31% outsourced to 54% in-house over four years) and Butternut Box (integrated manufacturing + fulfillment at Rudie's Kitchen, Doncaster). Protocol Raw reaches profitability faster than comparable DTC food businesses because the operating cost structure is fundamentally different: AI-automated core workflows with systems built before launch."
Anticipated Questions¶
"Why is the co-pack phase contribution margin so low?"
"Because we test every batch. At an 18.6% finished-batch reject rate, we produce 22.85% more than we ship. The co-packer won't run the upstream poultry gate for a single client, so we can't screen out the highest-prevalence Salmonella input before production. The cost is real, survivable, and funded by Seed. It is eliminated when we move in-house at M34."
"Why is in-house manufacturing required, not optional?"
"The only way to reinstate the poultry gate at scale is to own the facility. The co-packer will not run per-delivery Salmonella screening for a single client. Without the gate, the reject rate stays at 18.6% permanently. That caps contribution margin at ~26%, which is insufficient to fund aggressive acquisition or support the premium-multiple narrative. In-house manufacturing cuts COGS/kg from £3.63 to £2.90 and cuts reject rate from 18.6% to 4%. Both together deliver 45.9% contribution margin."
"Why are you doing fulfillment in-house rather than using a 3PL?"
"Because we're already building the facility for manufacturing, and 3PL overhead adds £3.50-4.50 per box. Adding pick/pack stations to a facility we already occupy costs £50-100k in additional capex and saves roughly £2 per box, which is £780k per year at 390,000 boxes. Butternut Box runs the same integrated model at Rudie's Kitchen in Doncaster. Bella and Duke do the same from their facility in Fife. Both validate that single-site manufacturing plus fulfillment is the category-standard model for raw pet food DTC at this scale."
"How do you know the 4% Phase A reject rate holds?"
"Phase A produces approximately 43 finished batches and 43 upstream poultry test results before Seed close. Zero failures in 43 batches gives 95% confidence that the true reject rate is below 6.7%. This is the empirical data that replaces the planning estimate. Seed closes on actual data, not projections."
"What if retention is lower than 70%?"
"If Box-2 drops below 60% for two consecutive cohorts, we pause acquisition and diagnose. We don't pour money into a leaky bucket. The model assumes 70%, but we're viable down to 65%. Below that, we fix retention before we scale."
"What if CAC rises?"
"We're modelling £66 blended at Series A-stage (25% organic mix). If it rises to £90, Phase B Early payback extends from 3.8 to ~5.2 months, still within range. Above £120, we shift budget to retention and organic growth. The 25% organic rate by M22 is our hedge against paid channel inflation."
"Why does cash decline even when you're profitable?"
"Working capital. Every new customer requires inventory pre-funded 4-6 weeks before they order. In the co-pack phase, that inventory is partially quarantined during hold-and-release lab testing (10-14 days at −18°C). Contribution exceeds CAC by M15, but inventory investment plus QA_HOLD exceeds contribution until later. The Seed and Series A bridge that gap."
"Why is facility capex £1.6M, not Bella & Duke's £4M?"
"B&D built a 34,000 sq ft facility for their current scale (£25M+ revenue, 134 employees, multi-SKU). Our Phase C facility is sized for ~30k customers — 8,000-15,000 sq ft, single-shift, with expansion optionality. It includes both manufacturing and fulfillment (pick/pack/dispatch), validated by the Butternut Box integrated model. We build what we need for Phase C, then expand at Phase D via second shift before facility expansion. That's capital discipline, not under-investment."
"Your Growth Strategy says <4% churn but your model uses 9.5%. Which is it?"
"The ~7.3% is the blended rate across a customer's entire lifetime, derived from 70% Box-2 retention combined with 5.5% steady-state churn on subscribers past Box-2. That's the number we use for LTV and financial modelling. The <4% is the aspirational operational target for post-Box-2 churn. Both are real; we fundraise on the 70% / 5.5% base case. If our operational target holds, actual LTV will be significantly higher than modelled."
"Isn't there a risk that your reject rate is worse than you think?"
"The Control Panel has three scenarios for Phase B: 12.4% (Conservative), 18.6% (Central, base case), and 25.0% (Stress). We model on Central. Phase A empirical data from 43 batches is the validation point before Seed close. If Phase A data comes in worse than 9% (our Stress case), we have a structural problem with the formulation that predates the co-packer transition and must be addressed first."
Part 10: Key Numbers to Know¶
Memorise these. They should be instant in any investor conversation.
Unit Economics¶
- Revenue per box: £92.50 (ex-VAT)
- Contribution per box by phase:
- Phase A M8 end-state: ~£25 (27% CM) [CHECK]
- Phase B Early (M9-M12): £14.85 (16.1% CM)
- Phase B Late (M13-M21): £24.23 (26.2% CM)
- Phase C In-House (M34+): £42.50 (45.9% CM)
- Revenue LTV: £1,375 (for context only)
- Contribution LTV trajectory: £221 → £360 → £594 (Phase B Early → Phase B Late → Phase C)
- Blended CAC at Series A-stage: ~£66 (25% organic mix)
- Organic share trajectory: 5% → 17% → 22% → 27%
- Coded referral target: 10-12% of customers at steady state (SOP-REF-01)
- Effective organic cost: £10-12/customer blended (coded referral £15, attributed organic near-zero)
- LTV:CAC trajectory: 3.4:1 → 5.5:1 → 9.0:1 (Phase B Early → Phase B Late → Phase C)
- Payback trajectory: 3.8 → 2.3 → 1.4 months
VC Path Milestones¶
- M0-M8: Self-manufacture (Phase A), 4% reject + poultry gate, 27% CM at M8
- M8: Operationally breakeven (self-manufacture, 346 customers, £39k ending cash)
- M9: Seed close (£1.15M), transition to co-packer, ~542 customers, £652k ARR
- M15: Operating profitable despite 18.6% reject (~4,127 customers, £5.0M ARR)
- M13: Contribution margin steps from 16.1% to 26.2% as COGS/kg improves
- M22: Series A close (£15M), ~10,877 customers, £13.1M ARR, £15.85M ending cash
- M23-M27: Facility build (integrated manufacturing + fulfillment), £1.6M total capex
- M28-M33: Parallel run, in-house ramps 15% → 90%, fulfillment ramps with production
- M34+: Full in-house, 4% reject + poultry gate at scale, 45.9% CM, 32% operating margin
- M36: ~42,301 customers, £50.9M ARR, £21.3M ending cash, £1.29M monthly operating profit
Capital Structure¶
- Pre-seed (personal): £90,000 at M0
- Seed: £1,150,000 at M9
- Series A: £15,000,000 at M22
- Total external capital: £16,150,000
Microbiological Risk Numbers¶
- Phase A reject rate: 4% (Conservative case, ACTIVE)
- Phase B reject rate: 18.6% (Central case, ACTIVE), no poultry gate
- Phase C reject rate: 4% (Conservative case), poultry gate reinstated at scale
- Poultry gate unit cost: £350/pack, 15% reject rate, £437.50/month at Phase A volumes
- Poultry gate at 40k customers: ~£50-80k/month (2-3% of revenue)
- Production yield: 100 kg shipped requires 122.85 kg produced at 18.6% reject
- Control Panel: rows 84-122 (editable yellow cells)
- Source analysis: REF-MICRO-01 v2.0
Phase A Evidence Base (Seed Diligence Data)¶
- Finished batches produced: ~43
- Upstream poultry tests: ~43
- Statistical confidence (zero failures): 95% that true reject rate <6.7%
Sensitivity Thresholds¶
- Box-2 retention floor: 60% (below this, pause and diagnose)
- CAC ceiling: £120 (above this, shift to retention/organic)
- COGS/kg targets: Self-manufacture improving through Phase A; Co-pack ≤£3.80/kg by M18; In-house £2.90/kg at M34
- Reject rate escalation: If Phase A data comes in above 9% (Stress case), formulation review required before Seed close
Cost Structure at Scale (40k customers, M34+)¶
- Total fixed costs: ~£508k/month (~£6.1M/year, excl. ad spend)
- Ad spend: £380k/month (9.6% of revenue)
- Founder salary: £8.3k/month
- HQ team (COO, CMO, FD, Marketing Lead, Product Engineer (AI Native), CS x2, HR): ~£40k/month
- Fixed overheads: ~£1.8k/month
- Manufacturing core opex: ~£41.7k/month
- Fulfillment/warehouse opex: ~£41.7k/month
- Operating profit: ~£1.25M/month
- Operating margin: ~31-32%
- Facility FTE: 20-24 (production + fulfillment, cross-trained)
- HQ FTE: 9-10
CPD Numbers (Memorise)¶
- Phase A: £22 → £19
- Phase B: £19 → £18
- Phase C (in-house pack): £14
- Phase D: £11
- Bottom-up Phase C build: packaging £5.50 + courier £6.70 + pick/pack £2.00 + exceptions £0.75 = £14.95
- Biggest lever: packaging (every £1 off = £1 off CPD)
- Benchmark validation: Phase C at Growth/Scale-up boundary (£11-16 range), £14 defensible with in-house pack advantage
- Source: Part 2B (CPD Methodology and Evidence Base)
Part 11: Strategic Implications¶
What these numbers mean for decisions.
Pricing¶
The model cannot survive a price cut. A 10% reduction destroys 71% of contribution. Premium positioning isn't vanity—it's survival. Every pricing discussion should start with "what does this do to contribution margin?"
Retention vs Acquisition¶
At 3.4:1 Contribution LTV:CAC at the Phase B Early trough, improving retention is more valuable than reducing CAC. A 5-point improvement in Box-2 retention (70% → 75%) adds more lifetime contribution than a £20 reduction in CAC. The leverage grows as contribution per box grows through the manufacturing transition.
Growth Rate¶
The constraint is working capital, not demand. Faster growth requires more inventory investment, which accelerates cash consumption. During the co-pack phase, quarantined inventory (QA_HOLD) at 18.6% reject with 10-14 day lab turnaround adds further working capital drag. Every £100 of shipped product requires £122.85 of production to be funded. Faster growth than the VC Path trajectory requires earlier Series A or larger Seed.
Hiring¶
Team build follows the phased plan (Part 8 narrative): Ops Lead at M9 (Seed), Marketing Lead + fractional FD at M12, full-time CS from M15, Product Engineer (AI Native) fractional from M16-20 if founder bandwidth tightens, Supply Chain Coordinator from M14-M16 if co-packer complexity escalates, CMO, FD, and full-time Product Engineer (AI Native) at M22 (Series A), facility team M28-M33. See Business Plan v3.4 for detailed org phasing.
The Manufacturing Transition¶
M22 Series A triggers a five-month facility build (M23-M27) then a six-month parallel run (M28-M33). By M34 the business is 100% in-house, reject rate is back to 4%, and contribution margin is 48.6%. This is not an optional optimisation: co-packing at 18.6% reject permanently caps contribution margin at ~29%, and the only way to reinstate the poultry gate at scale is to own the facility.
Inventory¶
6-week coverage is conservative. 5-week coverage is viable once co-packer production is stable (Phase B+). Going below 5 weeks risks stockouts, which kill retention. In Phase A (self-manufacture), inventory coverage is managed directly through production scheduling. Inventory discipline is a cash management tool, but stockouts are more expensive than excess stock.
Appendix A: Glossary¶
| Term | Definition |
|---|---|
| ARPU (Theoretical) | Annual Revenue Per User at perfect 4-week cadence (13 boxes/year). £1,180 ex-VAT. |
| Contribution LTV | Lifetime contribution (Revenue − COGS − CPD) per customer. Used for LTV:CAC and payback calculations. |
| ARPU (Realised) | Conservative annual revenue target accounting for cadence drift (~11.5 boxes/year). £1,040 ex-VAT. |
| Blended Churn | Effective monthly churn across entire customer lifetime, including Box-1 attrition. Used for LTV calculation. |
| Box-2 Retention | % of new customers who order a second box. Primary product-market fit signal. |
| CAC | Customer Acquisition Cost |
| COGS | Cost of Goods Sold (ingredients, production, packaging, lab testing) |
| Contribution | Revenue minus COGS minus CPD |
| CPD | Cost Per Delivery (courier, cold chain, 3PL) |
| FEFO | First Expired, First Out (inventory allocation method) |
| LTV | Lifetime Value |
| Steady-State Churn | Monthly churn rate on subscribers who have received ≥2 boxes. Excludes Box-1 attrition. Operational target: <4%. |
| Coded Referral | Customer acquired via SOP-REF-01 slug attribution system. Referrer receives £15 credit after referee's first delivery confirmed. Subset of organic acquisition. |
| MRR | Monthly Recurring Revenue |
| Organic Acquisition | Customer acquired without paid media spend. Includes coded referrals, attributed word-of-mouth, direct traffic, and branded search. |
| Working Capital | Cash tied up in inventory and receivables |
Appendix B: Cross-Document Metrics Reference¶
For anyone reading this document alongside the Growth Strategy or Business Plan, here is how the key metrics connect across documents.
Canonical Metric Definitions¶
| Metric | Value | Defined In | Used By |
|---|---|---|---|
| Revenue LTV | £1,375 | Financial Model v2.4 | Market sizing only |
| Contribution LTV | £221 → £360 → £594 (phase trajectory) | Financial Model v2.4 | LTV:CAC, payback |
| Contribution LTV:CAC | 3.4:1 → 5.5:1 → 9.0:1 (phase trajectory) | Financial Model v2.4 | Investor discussions |
| Theoretical ARPU | £1,202 ex-VAT | Financial Model v2.1 | LTV inputs |
| Realised ARPU | £1,073 ex-VAT | Growth Strategy v3.0, Business Plan v3.0 | ARR projections |
| Blended Churn | ~7.3% (base case, 70% Box-2 / 5.5% SS) | Financial Model v2.4 | LTV calculation |
| Steady-State Churn | <4% target | Growth Strategy v3.0 | Operational dashboard |
| Payback | 3.8 → 2.3 → 1.4 months (phase trajectory) | Financial Model v2.4 | Investor discussions |
| Phase B reject rate | 18.6% (Central, ACTIVE) | Financial Model v2.1 / REF-MICRO-01 v2.0 | COGS, yield model |
| Phase A / C reject rate | 4% (Conservative, ACTIVE) + poultry gate | Financial Model v2.1 / REF-MICRO-01 v2.0 | COGS, yield model |
| Series A trigger | M22 / approximately £13M ARR | Financial Model v2.4 (VC Path sheet) | Business Plan, VC Strategy |
COGS Trajectory¶
Four-phase manufacturing transition. No optionality. See Part 8.
| Phase | Months | Manufacturing | COGS/kg | Reject Rate | Effective COGS/kg |
|---|---|---|---|---|---|
| A | M0-M8 | Self-manufacture | £4.31 → improving | 4% | £4.49 + gate |
| B Early | M9-M12 | Co-packer | £3.85 | 18.6% | £4.73 |
| B Late | M13-M22 | Co-packer | £3.30 | 18.6% | £4.05 |
| Parallel | M28-M33 | Blended | £3.52 → £2.90 | 16.4% → 5.5% | £4.21 → £3.07 |
| C | M34+ | In-house | £2.90 | 4% | £3.02 + gate |
Effective COGS/kg = COGS/kg ÷ (1 − reject rate) + allocated poultry gate cost where applicable.
Document Versions¶
| Document | Version | Key Metrics |
|---|---|---|
| Financial Model | v2.2 | CPD rebased, facility opex restructured, team costs updated, Phase C in-house pack |
| Growth Strategy | v3.0 | ARR, ARPU, Phase targets, Dashboard (CPD targets pending update) |
| Growth Playbook | v2.5 | Channel execution, Phase A acquisition framework, QA_HOLD-adjusted WC thresholds |
| Business Plan | v3.2 | Strategy, capital, manufacturing phases (facility spec and CPD pending update) |
| SOP-META-01 | v2.0 | Paid social execution |
| SOP-GOOGLE-01 | v1.1 | Google Ads execution |
| VC Investor Strategy | v1.4 | Fundraise metrics, honest bridge narrative |
| Founder's Action Guide | v1.2 | Production workflow, operational sequencing |
| REF-MICRO-01 | v2.0 | Microbiological risk analysis (source for Financial Model v2.1 assumptions) |
Appendix C: Model Update Log¶
| Date | Version | Change | Rationale |
|---|---|---|---|
| Apr 2026 | v2.4 | Box-2 retention base case corrected from 75% to 70%. Aligned with Business Plan v3.4 and Growth Strategy v3.2 which both cite ≥70% as target. Steady-state churn unchanged at 5.5%. Lifespan 14.64→13.73 months. Blended churn 6.8%→7.3%. Revenue LTV £1,466→£1,375. Contribution LTV trajectory £251→£221, £416→£360, £714→£594. LTV:CAC trajectory 3.8:1→3.4:1, 6.3:1→5.5:1, 10.8:1→9.0:1. M22 ARR £13.8M→£13.1M, customers 11,439→10,877. M36 customers 44,883→42,301, ARR £54.0M→£50.9M, ending cash £22.4M→£21.3M, monthly operating profit £1.45M→£1.29M. Phase A/C reject rate corrected from 4% to 6.5%. Investor narrative FAQ and Appendix B cross-document metrics updated. Terminology: "Referral %" renamed to "Organic %" throughout; organic includes coded referrals (SOP-REF-01), attributed word-of-mouth, direct traffic, branded search. Part 6 restructured with organic composition breakdown and coded referral mechanism. Glossary updated. | The Financial Model was using 75% Box-2 retention as the base case while both the Business Plan v3.4 and Growth Strategy v3.2 consistently cited ≥70% as the target. Correcting to 70% aligns all three documents on a single defensible assumption. The 75% scenario is retained in the sensitivity table as an upside case. Phase A/C reject rate updated per REF-MICRO-01 v2.0 revised analysis. "Referral %" was not defensible as a single metric label because the 25% target includes organic sources beyond coded referrals. Industry benchmarks for coded referrals alone are 8-15% (Gousto, HelloFresh, Butternut Box). Renaming to "Organic %" with coded referrals as a tracked subset makes the acquisition model defensible in VC diligence. |
| Apr 2026 | v2.3 | Org structure revision. Added Product Engineer (AI Native) role to HQ team ramp: fractional £3,000/month from M16, full-time £4,583/month from M22. HQ team composition now includes Product Engineer (AI Native). HQ FTE 8-9 → 9-10; HQ team monthly cost £35k → ~£40k. M36 operating margin 32.3% → 32.2%; M36 ending cash £22.5M → £22.4M. Head of Customer Success reports to COO (was CMO), dotted line to CMO on retention. Marketing Lead salary standardised to £5,417/month (£65k loaded). | Adversarial review established "Engineering cost £0 forever" was not sustainable at scale. Product Engineer (AI Native) role removes founder bottleneck for routine system maintenance (alert triage, dashboard upkeep, reporting, routine extensions) while preserving architectural moat. CS under COO reflects operational nature of customer service (cancellations, churn intelligence, exception resolution). Aligned with Business Plan v3.4 and Financial Model v7. |
| Apr 2026 | v2.2 | CPD methodology rebased from bottom-up research. Phase B early £18→£19, Phase B late £16→£18, Phase C £10→£14, Phase D £8→£11. Phase C fulfillment changed from outsourced 3PL to in-house pack at integrated manufacturing facility. Unit Economics Phase C column changed from co-pack (COGS £3.63/kg) to in-house (COGS £2.90/kg). New Part 2B: comprehensive CPD methodology with bottom-up builds, carrier pricing research, benchmark validation against volume tiers, exception cost model, and Bella and Duke / Butternut Box comparable evidence. Fixed cost structure restructured: "Team costs" → "HQ team" with Series A exec hires added (CMO £8,333/mo, FD £8,333/mo, COO step-up, additional CS, HR fractional; total HQ team £35,000/mo from M22). "Manufacturing opex" split into Manufacturing core (£41,667/mo) + Fulfillment opex (£41,667/mo), total facility opex £83,333/mo = £1.0M/yr at M34+. Facility FTE revised from 12 to 20-24 (cross-trained production + fulfillment operatives). M36 operating margin: 38.1%→32.3%. M36 ending cash: £25.1M→£22.5M. Contribution per box trajectory: £14.85→£24.23→£42.50. | Three independent research exercises (frozen DTC 3PL market, UK carrier pricing, frozen DTC benchmarks) found the original CPD assumptions were not achievable at Phase C volumes. The £10 Phase C CPD corresponded to Market Leader tier (50k+ drops/week); Protocol Raw's Phase C volumes (10k-12k drops/week) sit at Growth/Scale-up boundary (£11-16 benchmark). In-house pack decision driven by: 3PL overhead £3.50-4.50/box, factory already being built, Butternut Box integrated model validation. Team costs gap found between Business Plan org chart and model: Series A exec hires (CMO, FD, COO) were in the Business Plan but not the model. Fulfillment staffing analysis (with external review) determined 8-12 additional cross-functional warehouse/fulfillment operatives needed for 2,000 boxes/day at 5-minute pack time. All changes produce a more honest, defensible model that can withstand investor due diligence. |
| Apr 2026 | v2.1 | Microbiological risk model integration (REF-MICRO-01 v2.0). New Part 2A on production yield model and reject cost taxonomy. Unit Economics tables (Part 4) rewritten with phase-specific reject overhead and poultry gate lines; contribution per box trajectory £15.85 → £26.23 → £45.04. LTV recomputed as phase trajectory £251 → £416 → £714. Churn sensitivity and cadence drift tables recomputed at Phase B Early contribution. LTV:CAC restructured as phase trajectory 3.8:1 → 6.3:1 → 10.8:1 with blended CAC of £66 at Series A-stage. Payback trajectory 3.8 → 2.3 → 1.4 months. Part 8 rewritten: "Manufacturing Decision" replaced with "Manufacturing Transition"; Option A vs Option B framing removed; four-phase trajectory (Self-Mfr → Co-Pack → Parallel → In-House) with Series A at M22 / £13.8M ARR and £1.6M capex over M23-M27. Part 9 investor narrative rewritten around the honest bridge story; new FAQs on reject rate, poultry gate, Phase A evidence base. Part 10 Key Numbers rewritten with VC Path milestones and capital structure (£90k pre-seed, £1.15M Seed, £15M Series A). Part 11 hiring and manufacturing transition subsections updated. Appendix B cross-document metrics refreshed; COGS trajectory table now single-path. | The prior v2.0 model implicitly assumed zero production waste. The FSA's 2026 survey found 28.7% statutory failure across 50 raw pet food brands. REF-MICRO-01 v2.0 costed this honestly and established that in-house manufacturing with the poultry gate is the structural solution, not an optimisation. This update propagates the reject cost model and removes the Option A/B decision framing from all unit economics, LTV, and narrative sections. |
| Jan 2025 | v1.0 | Initial documentation | Capture financial model logic |
| Jan 2026 | v1.1 | Manufacturing decision framework | Align with BP v1.7 and GS v2.2 |
| Jan 2026 | v1.2 | Manufacturing cost model correction | Align with BP v2.0 |
| Feb 2026 | v1.4 | Two-phase manufacturing strategy: Phase A self-manufacture, Phase B co-packer transition | Align with Business Plan v2.4, Growth Strategy v2.5, Manufacturing Playbook v2.0 |
| Feb 2026 | v1.3 | Metrics reconciliation: churn decomposition, cadence drift, ARPU terminology | Align with Metrics Reconciliation v1.0, BP v2.3, GS v2.3 |
| Mar 2026 | v1.5 | Phase A fixed costs corrected (tooling £3,500 → £300/month); at-scale costs recalculated; Claude Max 5x added to tooling stack | Align with Business Plan v2.5, Growth Strategy v2.6 |
| Apr 2026 | v2.0 | Phase A referral share corrected (10-15% → 5%) to match Financial Model spreadsheet. Blended CAC table updated (Phase A £100-105 / £110 / 5%). LTV:CAC recalculated with spreadsheet Contribution LTV (£319 → £424) using 75% Box-2 / 5.5% SS churn base case and Phase B Early contribution of £26.76/box. Revenue per box updated £90.83 → £92.50. Payback 2.6 → 2.8 months. Unit economics tables updated with spreadsheet values for Phase A (Rev £92.50, GP £39.06, CPD £22, Contribution £17.06) and Phase B Early/Late split. Cross-references updated to current document versions (BP v2.9, GS v2.9, GP v2.3). | Align with Financial Model spreadsheet, Growth Strategy v2.9, Business Plan v2.9 |
v1.5 Changes (March 2026)¶
Phase A Fixed Costs Corrected:
- Tooling: £3,500/month → £300/month. The £3,500 figure was a planning placeholder from before the tech stack was built. Actual Phase A tooling: Supabase Pro £20, Make.com Core £9, Shopify Basic £31, Seal Subscriptions £0-40, Metabase £6, Cloudflare £0, email/domain £10, AI API £15-30, Claude Max 5x £80. Customer.io on free startup plan (budget £80/month if expires). Total ~£290-310/month.
- Phase A total fixed costs: £5,500/month → £2,300/month. This aligns with Phase A burn rate of £8-12k/month per Growth Strategy v2.6 and Business Plan v2.5's £50-75k total Phase A capital.
- Phase B tooling updated: £4,500 → £1,500/month. Reflects Customer.io Essentials, upgraded Make.com/Supabase/Seal tiers.
- At Scale tooling updated: £6,000 → £3,750/month. Aligns with Business Plan v2.5's £45k/year tooling figure.
- Cost Structure at Scale (100k customers): Option A £140k → £249k/year (corrected to match £20,750/month × 12). Previous figure was arithmetically inconsistent with the fixed costs table.
- Claude Max 5x (£80/month) added as permanent tooling line. Founder uses Claude as primary AI cofounder; this is core infrastructure, not optional.
Rationale: Ground-up cost audit revealed the Phase A tooling figure was 12x actual cost. Most tools operate on free or starter tiers at Phase A volumes (0-300 customers). The correction cascades through total fixed costs, operating breakeven timing, and cash flow projections. At-scale cost structure also corrected for arithmetic consistency.
v1.4 Changes (February 2026)¶
Two-Phase Manufacturing Strategy (Ref: Business Plan v2.4, Manufacturing Playbook v2.0):
- Phase A unit economics rewritten for self-manufacture (COGS £46-55/box at £3.50-4.50/kg)
- Phase B unit economics updated for co-packer hybrid model (COGS £40-47/box at £3.20-4.30/kg)
- Variable costs table updated to reflect self-manufacture → co-packer transition
- COGS improvement narrative restructured: Phase A self-manufacture → Phase B co-packer transition → Phase C manufacturing decision
- Manufacturing decision reframed as "return to manufacturing" with Phase A production experience
- Co-packer COGS trajectory updated (£3.20-3.80/kg target, not £2.80-3.10/kg)
LTV Framework Correction:
- Split LTV into Revenue LTV (£1,033), Gross Profit LTV (~£500), and Contribution LTV (£319)
- LTV:CAC ratio now computed on Contribution LTV: 4.0:1 (was 13:1 on Revenue LTV — overstated)
- Payback period recomputed with explicit cadence mapping: 2.6 months at Phase B contribution
- Fundamental equation updated to use Contribution LTV
- Churn sensitivity and cadence drift tables updated with both Revenue and Contribution LTV
- Investor narrative rewritten for Contribution LTV basis
- Glossary expanded with Contribution LTV definition
Encoding and Hygiene:
- Full encoding cleanup: all double-UTF-8 mojibake resolved (box-drawing chars, math symbols, punctuation)
- Exit multiples reframed as planning assumptions with source citations
- Competitor precedent sourced to Companies House filings
v1.3 Changes (February 2026)¶
Metrics Reconciliation (Ref: Metrics Reconciliation v1.0):
Applied edits from cross-document metrics audit resolving inconsistencies across Financial Model, Growth Strategy, and Business Plan.
Part 5 (LTV) — Major expansion: - Renamed "ARPU" to "Theoretical ARPU" to distinguish from Business Plan's "Realised ARPU" (£1,040) - Added blockquote explaining the £1,040 LTV / £1,040 Realised ARPU coincidence - Renamed "Monthly churn" to "Blended monthly churn" with blockquote explaining two-layer decomposition - Added Two-Layer Churn Model section with canonical definitions (Box-2 Retention, Steady-State Churn, Blended Churn) - Added Churn Sensitivity (Two-Layer) table showing LTV across Box-2 retention and steady-state churn combinations - Added Cadence Drift Sensitivity table showing LTV impact of cadence drift independent of churn - Replaced old single-dimension churn sensitivity table with richer two-layer version
Part 9 (Investor Narrative) — New FAQ: - Added anticipated question: "Your Growth Strategy says <4% churn but your model uses 9.5%. Which is it?" - Answer explains two-layer model and why both numbers are real
Appendix A (Glossary) — Updated terminology: - Split "ARPU" into "ARPU (Theoretical)" and "ARPU (Realised)" - Added "Blended Churn" and "Box-2 Retention" definitions
Decision applied: Option A confirmed — Financial Model keeps Theoretical ARPU (£1,180) for LTV calculation with cadence drift shown as separate sensitivity dimension. Does not stack cadence conservatism on top of churn conservatism.
Rationale: The Metrics Reconciliation v1.0 identified that £1,040 appeared across all three planning documents meaning three different things (LTV in the Financial Model, Realised ARPU in the Growth Strategy, conservative ARPU target in the Business Plan). An investor reading across documents would find contradictions. These edits make every metric's definition, scope, and cross-reference explicit.
v1.2 Changes (January 2026)¶
Manufacturing Cost Model Corrected: - Option B capex updated: £3-5M → £1.25-2.0M (right-sized for Phase C ~30k customers) - Option B annual fixed opex added: £400-600k/year (facility, utilities, 6-8 production staff, QA, maintenance, insurance) - Series A sizing for Option B updated: £15-20M → £12-17M - Payback period updated: 10-14 months → 12-18 months (reflects full investment including first-year opex) - Financial comparison table updated with corrected figures - Added "Cost Basis" section explaining methodology and referencing BP v2.0 appendix - Added anticipated investor question about capex vs B&D benchmark
Fixed Costs Section Enhanced: - Added Option B manufacturing opex breakdown table - Added Option B trade-off explanation (higher fixed costs offset by contribution gains) - Updated cost structure at scale to reflect Option B reality
Key Numbers Updated: - Added Manufacturing Decision Numbers subsection - Updated Phase B milestones to include Month 12-14 Technical Bandwidth Assessment - Updated cost structure at scale for Option B
Rationale: v1.1 used Bella & Duke's £3-5M facility investment as the Option B benchmark, but B&D's 34,000 sq ft facility serves £25M+ revenue—an upper bound inappropriate for Protocol Raw's Phase C scale (~30k customers). Corrected figures reflect an 8,000-15,000 sq ft facility sized for Phase C with expansion optionality at Phase D. The addition of explicit annual fixed opex (£400-600k) provides complete cost picture for Option B evaluation. Payback remains attractive at 12-18 months even with full cost accounting.
This document should be reviewed quarterly and updated when assumptions change materially.