RefRelay

1. One-liner

Gets the EUDR reference number to your customs broker before the container lands, not after it’s stuck.

2. Trend signal — why now?

The EU Deforestation Regulation (Regulation EU 2023/1115) is enacted and the clock has actually started. Large and medium operators must comply from 30 December 2026; micro and small operators from 30 June 2027. This is not a proposal, not a roadmap, not a vetoed bill — the Commission adopted implementing regulations on 13 July 2026.

Three things changed in the last twelve months that make this the right moment:

The customs gate became real. A Due Diligence Statement must be submitted per shipment or batch, and the resulting reference number goes into the customs declaration. The operative sentence, from a German customs practice: “Without a valid reference number or valid declaration identifier, a product will not be released at customs.” Gerlach Customs is blunter about the consequence: “Incorrect or missing EUDR documentation therefore becomes a direct logistics risk, with possible delays, storage costs and supply chain disruption.”

The plumbing got rebuilt underneath the incumbents. The EUDR Information System no longer depends on TRACES. Every vendor who sold “TRACES NT integration” as their headline feature is re-tooling right now. That’s a rare window where the incumbent’s main asset is temporarily a liability.

The market is spectacularly unprepared. One survey of 47 roasters across Italy, Spain, Germany and France found 73% still don’t understand what “deforestation-free” means for their business and 89% have done zero traceability work. Their stated plan: “We’ll deal with it when the deadline gets closer.” The deadline is now four months away.

Meanwhile pricing is stuck at enterprise altitude. Coolset starts around €12,000/year. osapiens and TraceX are sales-led with no public pricing. One EUDR consultancy cites a Nigerian client quoted €40,000/year just for deforestation analysis — “a cost that would eat a large part of their profit.” The Commission’s own estimate of aggregate annual compliance cost runs €175M–€2.6B.

Provenance:

• Signal 1 (Demand): 73% of EU coffee roasters don’t understand “deforestation-free”; 89% have done zero traceability work; ~60% of DDS rejections trace to data gaps — https://www.fiegenbaum.solutions/en/blog/eu-deforestation-regulation-eudr-coffee-industry-challenges-strategies and https://getbindu.com/blog/eudr-coffee-importers — observed 2026-08-26
• Signal 2 (Feasibility): EUDR Information System decoupled from TRACES; implementing regulations adopted 13 July 2026; polygon validation commoditized to free/open-source (GeoRoots) — https://www.coolset.com/academy/the-eu-deforestation-regulation-eudr-what-businesses-need-to-know-and-do — observed 2026-08-26
• Signal 3 (Economic): Entry EUDR software ~€12,000/yr; €40,000/yr quoted for a single mid-size analysis; fines up to 4% of EU turnover; demurrage $100–150/container/day — https://eudrready.eu/eudr-software-small-business/ and https://www.eprod-solutions.com/articles/is-eudr-cocoa-compliance-priced-for-the-african-market — observed 2026-08-26 Category: Regulatory arbitrage

3. The opportunity

Everybody is selling the wrong half of this problem.

Walk the vendor landscape and you find two clusters. Upstream, there’s a crowded field selling polygon validation and satellite risk analysis — Silvatrace, geocledian’s EUDR Check, TraceX, Koltiva, plus GeoRoots which is free and open source. That job is commoditized. Nobody should build there.

Downstream, EUDRReady sells a €29–79/month document vault for SME traders who only need to collect and store DDS reference numbers from suppliers. Its own comparison table marks satellite monitoring, TRACES submission, full risk assessment, and GPS/polygon collection as ”❌ Phase 2.” It’s a filing cabinet, and it’s honest about it.

Between those two sits the actual operational failure, and nothing is sold into it: the per-shipment race between a DDS reference number and a container arriving at Rotterdam.

Here’s the shape of the failure. The reference number is generated when the operator submits the DDS. It must be in the customs declaration — as document code C716 in Box 44 of the SAD — at the moment of declaration. The importer’s broker files that declaration. So a reference number that exists but hasn’t reached the broker in the right form, matched to the right consignment, with the right net weight and TARIC code, is worth exactly nothing. The container sits.

And the mismatch is structural, not careless. A single reference number can cover multiple shipments across a year. Practitioners report companies “frequently fail to align inbound, storage and customs declarations properly, creating mismatches at border control.” The regulation also splits duties in a counterintuitive way — downstream collection is passive, meaning many importers don’t even know whose job it is to chase the number.

Incumbents treat the reference number as an output — the receipt you get when the compliance work finishes. For a small importer it is an input to a logistics deadline they don’t control. That inversion is the gap. It’s the same pattern as capture-vs-defense: the industry solved generating the artifact and left the consequences of the artifact arriving late completely unbuilt.

4. Target market

• Primary customer: Owner-operators and single-person compliance/ops staff at EU importers of coffee, cocoa, rubber, palm, soy and timber doing €2M–40M annual turnover and 20–400 import consignments a year. Concentrated in the Netherlands, Germany, Belgium, Italy and Spain. Job titles: Managing Director, Head of Operations, Import Manager, Green Coffee Buyer. Crucially these are firms with a customs broker on retainer but no sustainability manager — the exact profile the €12k tools are mispriced for.

• Why they buy: Not because they love compliance. Because a held container has a meter running. Demurrage and detention run $100–150 per container per day, with only 3–7 free days, and reefer containers run 1.5–2× that. One held coffee container over a long weekend costs more than a year of this software. The regulatory fine ceiling — 4% of EU turnover — is the tail risk that gets it approved, but demurrage is what gets it bought this week.

• Rough TAM reasoning: The EU imports roughly 40% of the world’s coffee, and cocoa alone accounts for about a third of the EU’s deforestation exposure. Across the seven regulated commodities, EU importers in the €2–40M band plausibly number in the low tens of thousands. I need ~700 of them at €250/mo to hit $2M ARR. That’s a small single-digit share of a specific, listable, addressable set — and unlike most compliance TAMs, this one has a public register forming around it.

• Why now for them: Their first consignment under the regime lands in the weeks after 30 December 2026. Small operators get until 30 June 2027, but their large customers hit the earlier date and will start demanding numbers upstream immediately. The buying window opens roughly October 2026 and stays hot through mid-2027.

5. Product sketch (MVP)

• Shipment countdown board — every inbound consignment with ETA, commodity, supplier, and a single status: reference number in broker’s hands / requested / missing. Sorted by days-to-arrival, so the thing about to cost money is at the top.
• Reference chase agent — automatically emails the upstream operator or exporter for the DDS reference at a lead time you set (default: ETA minus 10 days), escalates on silence, and files the reply. Handles the passive-collection ambiguity by chasing anyway.
• Inbox extraction — forward or connect the mailbox; it pulls reference numbers, declaration identifiers and verification numbers out of supplier PDFs and email bodies, and tells you which of the three you actually received. Practitioners note these three terms are “regularly confused in practice” — so the product names the confusion out loud.
• Broker handoff packet — one-click export giving the customs broker exactly what the declaration needs: reference number, TARIC document code, net weight in kg, tariff number, and the consignment it maps to. Email or CSV, in the broker’s format.
• Coverage-vs-consignment matcher — when one reference covers multiple shipments across a year, it tracks remaining coverage against actual inbound volume and warns before you over-draw a reference.
• Pre-declaration checks — flags the known rejection triggers before submission: HS/TARIC mismatch, missing supplier identity, weight discrepancy, batch traceability gap.
• Five-year evidence file — every number, every chase email, every broker handoff, timestamped and exportable, because retention is five years and competent authorities can request evidence at any point.
• Substantiated-concern responder — if a concern lands, it assembles what you sent, when, and to whom, in the notification format the competent authority expects.

6. AI angle — what’s load-bearing

Remove the AI and this collapses into a spreadsheet with a calendar reminder — which is precisely what these importers already have and precisely why containers get held.

Two places AI does real work:

Extraction from unstructured supplier comms. The reference number does not arrive in a clean API payload. It arrives in a PDF attachment from an Ethiopian exporter, in the body of a WhatsApp-forwarded email, in a scanned bill of lading, in three languages, sometimes mislabeled as a verification number when it’s actually a declaration identifier. Distinguishing those three is a semantic judgment on messy input, and getting it wrong means filing a declaration with a number that doesn’t validate. Cheap multimodal extraction is what makes this tractable for a two-person team in 2026 — it was not in 2023.

Chase correspondence that actually lands. Chasing a reference number from a smallholder cooperative’s export desk is not a templated dunning email. It requires explaining, in the recipient’s language, what a reference number is, why you need it, and which of the three things they might send is the right one. Generated, context-aware, multilingual chase mail is materially better than a fixed template — and the chase is the entire product.

What is emphatically not AI here: deforestation risk scoring. That’s commoditized, partly free, and I’d rather integrate GeoRoots or a satellite vendor than pretend to compete there.

7. Localization angle

This is an EU-market play by construction — the regulation defines the geography. But there’s real localization inside it:

• Broker formats differ by member state. A Dutch declaration workflow and a German one want the handoff packet differently. Shipping NL + DE first, then IT/ES/BE, is the sequencing.
• The chase is outbound to origin countries. Chase mail needs Portuguese (Brazil), Spanish (Colombia, Peru), Vietnamese, Bahasa, French (Côte d’Ivoire) and Amharic-adjacent English. The product’s outbound surface is emerging-market-facing even though the customer is European.
• Pricing. €199–349/mo works for a Dutch importer where €12,000/yr does not. That’s the whole arbitrage.

8. Business model — path to $1M–$5M ARR

• Pricing: €199/mo Starter (up to 50 consignments/yr, 1 commodity), €349/mo Standard (up to 200 consignments, unlimited commodities, multi-broker), €699/mo Multi-entity (several importing entities, API to broker software). Annual billing at 2 months free.
• ACV: ~€3,400 blended (mix skewed to Standard).
• Rough math to $1M ARR: 275 customers × €349/mo × 12 ≈ €1.15M. Realistic in 18 months if the Dec-2026 window converts.
• Rough math to $5M ARR: ~1,300 customers, which needs (a) all seven commodities not just coffee/cocoa, (b) broker-channel distribution rather than direct-only, and (c) the €699 tier carrying 25% of the base. Achievable but it’s a 3-year shape, not an 18-month one.
• Expansion path: Consignment volume tiers are the natural meter — customers grow into higher bands without a sales conversation. Then multi-entity. Then sell a seat to the broker, who has 40 importer clients with the same problem — that flips the model from direct SMB to channel and is where the second €2M lives.

Gross margin is healthy: the expensive call is extraction, and a consignment involves a handful of documents. Even at 1,000 customers averaging 150 consignments/yr, inference is a low-single-digit percentage of revenue.

9. Go-to-market wedge — first 100 customers

• Customs brokers are the channel, and there are few enough to name. In NL, DE and BE the mid-market customs brokerage field is a few hundred firms, publicly listed via national customs authorities and forwarder associations (FENEX in NL, AEB/DSLV networks in DE). A broker with 40 importer clients is about to get 40 panicked calls in January 2027. Approach 200 brokers with a co-branded offer — they look competent, we get distribution, they take 20%. Ten brokers closing five clients each is 50 customers.
• Commodity trade associations with member lists. The European Coffee Federation, the European Cocoa Association, national green-coffee associations, and the Cocoa Association of Asia all publish or share member directories and are actively circulating EUDR guidance to members. Sponsor the EUDR webinar rather than buying a booth — the webinar attendee list is the qualified lead list.
• Trade-show timing, not trade-show presence. World of Coffee and the Amsterdam/Hamburg green coffee circuit run in the exact pre-deadline window. Don’t take a stand — take meetings. Fifty pre-booked 20-minute conversations with import managers beats a €15k booth.
• The held-container postmortem, as content. Publish a plain-language teardown of exactly what goes in Box 44, what C716 means, and the three-numbers confusion — the single most searched, worst-explained thing in this regulation. This is the one place SEO genuinely works, because the search intent is “my container is held” and that searcher buys today.
• Direct to the unprepared. The roaster survey says 89% have done nothing. National business registries plus commodity import data identify importers by commodity code. Scrape, enrich with EORI-linked filings, send a two-minute Loom showing their likely January exposure. At 2,000 contacted and a 3% conversion, that’s 60 customers.

10. Build complexity — justification

Medium. Off the shelf: document extraction, email infrastructure, standard web stack, and — importantly — deforestation risk analysis, which I integrate rather than build. The custom work is the consignment-to-reference state machine, the coverage matcher, and the broker handoff formats, which need to be right per member state. The one genuine unknown is integration with the new EUDR Information System, which decoupled from TRACES and whose interface is still stabilizing — that’s a real schedule risk, and the reason v1 should treat number capture and relay as the core and treat submission as a fast-follow. Two people, 12–16 weeks to a v1 that a real importer can run a season on.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Helps importers meet an EU obligation; no regulated-activity licence needed to assist with filings
Ethical — no harm / dark patterns	✅	The product’s incentive is aligned with the regulation’s intent — it makes traceability arrive on time, not disappear
Market exists (evidence above)	✅	Enacted regulation, dated deadline, €12k incumbent floor, documented rejection rates
1–5 person team can build this	✅	Two people, 12–16 weeks
Launchable with <$50K / ₹40L	✅	~€25–35K covering two people’s runway, satellite/risk API fees, and member-list sponsorships

All five pass.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	17/20	Held container bleeds $100–150/day with 3–7 free days; fine ceiling 4% of turnover. Pain is dated, metered, and lands on a specific person. Not 19 because it’s acute-episodic rather than daily.
Demand evidence	15	12/15	Strong indirect: 89% unprepared, ~60% of rejections from data gaps, €12k incumbent floor, €40k quote. Docked because I could not source direct verbatim importer complaints about this specific reference-relay failure — the inference is mine.
Build feasibility	15	11/15	Standard stack plus extraction, but the new Information System interface is still stabilizing and broker formats vary by member state. 12–16 weeks, not 6.
Distribution clarity	15	12/15	Brokers are a named, countable, incentive-aligned channel with existing client books. Association member lists are real. Docked because broker partnerships take longer to close than the math implies.
Revenue mechanics	15	12/15	€199–699 is well below the €12k incumbent and well above hobby pricing; 275 customers to $1M is credible. The €5M path needs the broker channel to work, which is unproven.
Time to first revenue	10	7/10	Deadline-driven urgency compresses the sales cycle, but the buying window genuinely opens around Oct 2026 — you can pre-sell before then, not invoice much. 6–10 weeks post-launch.
Defensibility	10	5/10	Honest answer: execution and channel, not moat. Broker relationships and accumulated per-member-state handoff formats compound by month 12. But a funded incumbent could add this to an existing suite.
Total	100	76/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · domain-expertise-required

You need someone who can hold the customs-declaration mechanics in their head — TARIC codes, Box 44, the three-numbers distinction — and someone who can ship extraction that works on bad scans. A pure marketer will drown in the domain; a pure engineer will build the polygon validator nobody needs.

Key assumptions to validate (3–5)

1. Assumption: Small importers experience reference-number-to-broker timing as a distinct, named problem — not as an undifferentiated part of “EUDR compliance.” How to test: 25 structured interviews with import managers at €2–40M coffee/cocoa importers in NL/DE. Ask them to walk through their January 2027 first consignment step by step. Do they spontaneously identify the broker handoff? If fewer than 8 do, the wedge is wrong even if the market is right.

2. Assumption: Customs brokers will co-sell rather than absorb this into their own service billing. How to test: Pitch 15 mid-market brokers directly. A broker who says “we’ll just do this for our clients as a paid service” is a competitor, not a channel — and if more than half say that, the distribution plan collapses to direct-only.

3. Assumption: €199–349/mo clears the willingness-to-pay bar for a firm that currently pays €0. How to test: Pre-sell annual contracts at 50% discount to the first 20 interviewees before writing code. Fewer than 5 conversions means the pain isn’t priced yet.

4. Assumption: The new EUDR Information System exposes a workable interface for submission and reference retrieval by Q4 2026. How to test: Register as an operator now, attempt test submissions under the July 2026 implementing rules, and document what’s actually available.

Risk flags

1. Regulatory risk — the deadline slips again. EUDR has already been delayed once and “simplified” repeatedly; a review clause sits at April 2026 with an explicit brief to cut SME administrative burden by a target share. Another postponement doesn’t kill the product but pushes the revenue window by a year, which for a bootstrapper is the difference between viable and not. This is the single biggest risk on the page.

2. Platform dependency — the Information System is the product’s spine. It just decoupled from TRACES and is still stabilizing, with contingency arrangements for outages only required to be published by 30 December 2026. Interface churn hits directly.

3. Channel-becomes-competitor. Customs brokers and freight forwarders (DHL, Customs Support Group) already market EUDR as a service line. The channel I want to sell through may prefer to sell this themselves. Mitigation is being infrastructure they resell, not a brand competing for their client.

4. Simplification erodes the wedge. If the Commission’s burden-reduction push exempts a large slice of the €2–40M band, or lets one annual declaration cover a year of shipments cleanly, the per-shipment urgency softens considerably.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical pair, one with customs/freight domain exposure
Time to revenue:        6–10 weeks post-launch; pre-sales possible from Oct 2026
Capital to launch:      €25–35K ($28–40K)
Top 3 assumptions to validate first:
  1. Importers name the broker-handoff timing as a distinct problem — 25 interviews, need 8+ spontaneous mentions
  2. Brokers co-sell rather than compete — pitch 15, need fewer than half saying "we'd do this ourselves"
  3. €199–349/mo clears WTP — pre-sell 20 discounted annuals, need 5+ conversions before writing code
Kill criteria:
  - Abandon if EUDR is postponed beyond December 2027 (revenue window gone for a bootstrapper)
  - Abandon if fewer than 5 of 20 pre-sale attempts convert at half price
  - Abandon if >8 of 15 brokers say they'll build or absorb this in-house
  - Abandon if simplification exempts the €2–40M turnover band or removes per-shipment DDS

15. Next step — 1-week validation sprint

• Day 1–2: Register as an operator in the EUDR Information System and attempt a test DDS submission under the July 2026 implementing rules. Document exactly what the interface gives you and what it doesn’t. This is falsifiable in 48 hours and it de-risks the biggest build unknown.
• Day 3–4: Book and run 12 calls — 8 import managers at €2–40M coffee/cocoa importers in NL/DE, 4 mid-market customs brokers. Single question to the importers: “Walk me through your first consignment after 30 December. Where does the reference number come from and who gives it to your broker?” Say nothing about the product until they’ve answered.
• Day 5: Pre-sell. Offer the first 20 a €999 annual founding price. Decide on hard numbers.

Go / no-go: Proceed only if ≥5 of 8 importers cannot confidently answer who hands the reference to their broker, AND ≥3 of 20 pre-sale offers convert to a paid deposit, AND ≤2 of 4 brokers say they’d build it themselves. Anything less and the pain is real but not yet priced — revisit in October when the deadline is close enough to hurt.