Year 5 — the destination: $100M on $2.0B GMV
Gigabite is the system of record for how 1,000+ foodservice buyers source private-label goods. Every spec, quote, order, shipment, invoice, and payment is an evidence-backed ledger event. AI does the work; the ledger keeps the truth — 90% of facts auto-approve, humans touch only material exceptions.
Transaction revenue
~3.5% blended take on $2.0B of sourced goods flowing through the ledger.
Embedded finance
Net terms underwritten from our own payment ledger — ~$1.2B financed volume at a blended ~1.5% fee per cycle.
Software & data
Ledger seats, supplier intelligence, category benchmarks.
The ARR ladder, read backwards
EXIT ARR BY FISCAL YEAR · $MFull operating model, year by year → Business roadmap
Year 4 — scale to $1.0B GMV
For $100M in Y5 to happen, all of this must already be true by the end of Y4:
National coverage without service decay
Freight + 3PL network and a supplier bench deep enough to absorb $2.0B GMV — fill rate and OTIF hold as volume doubles.
Credit engine underwriting from our own ledger
Net terms priced from proprietary DSO, exposure, and payment history. Fintech revenue grows faster than risk headcount.
A category machine, not category projects
New product family live in under 30 days: attribute schema → ontology → auto-matching. 25 families spanning disposables, jan-san, packaging, and PPE.
Year 3 — the machine becomes repeatable
For Y4 to scale, growth must stop depending on heroics:
Repeatable go-to-market
Sales playbook with under-6-month CAC payback; quote-to-order conversion proven cohort over cohort, not deal by deal.
Embedded finance launches
Net terms offered in-quote, underwritten from ledger history; credit exposure tracked live in the financial sub-ledger.
The moat becomes measurable
Match approval rates rise with volume; supplier scores start predicting repeat-order performance. Expansion revenue passes new-logo revenue for the first time.
EXITING YEAR 3
- $450M
- GMV through the ledger
- 200
- active buyers
- 16
- product families
- 75%
- facts auto-approved
Year 2 — the pilot becomes a product
For Y3 to repeat, the platform has to run without us in the loop:
Self-serve document intake
Buyers upload item lists, invoices, catalogs. AI extraction plus review queues run by their ops teams, not ours.
Ops leverage proven
Reconciliation agents on quotes, POs, invoices, BOLs — GMV per ops FTE nearly doubles year over year.
10 families, auto-matching on
Required-attribute schemas defined per family. Auto-approval is the default path; review is the exception.
First expansion motion
Existing buyers add categories. Repeat-order cohorts become the sales reference, not the demo.
Year 1 — prove the ledger, one quarter at a time
For Y2 to become a platform, Y1 has to earn trust in the ledger itself:
Ledger live
Append-only events, evidence-linked candidate facts, review queues. Postgres system of record + object-store evidence, per the final architecture.
First anchor buyers
5 buyers live. 8 disposable families with required-attribute schemas. Evidence-backed landed-cost quotes v1.
Close the loop
Quote → order → invoice → payment on-ledger. Reconciliation agents on invoices, packing lists, BOLs.
First cohort proof
20 buyers, $45M GMV run-rate, first repeat-order cohorts. Auto-approval reaches 40% of facts.
How this gets built, phase by phase → Technical roadmap
Every transaction makes the next one cheaper
The data moat is the plan. Normalized demand, SKU equivalency, supplier capability, cost, quality, and payment data — none of it exists anywhere else in this category, and all of it accrues to the ledger.
SKU equivalency
Every approved match trains attribute-level equivalency — match quality rises with volume, review cost falls.
Supplier intelligence
Certification scope, quality history, and landed-cost history compound into scores no entrant can replicate.
Payment history
Proprietary receivables and payment data underwrites net terms at margins banks can't price.
Ops per FTE
Reconciliation agents absorb volume: GMV per ops FTE climbs from $15M in Y1 to $50M in Y5.
THE SECOND MOAT · THE PRICING SPINE
One system — not a dozen apps
Every surface reads and writes the same ledger. One product record, one price that firms up, one shared credit line — so no two surfaces ever disagree. Only Analysis ships today; the rest are specced against the final architecture and built in sequence.
- Analysis LIVE
- Samples SPEC
- Packaging Review SPEC
- Container Builder SPEC
- Repeat Orders SPEC
- Tariff Watch SPEC
- Statements SPEC
- True-up SPEC
- Rebate SPEC
- Traceability & Recall SPEC
- Backhaul SPEC
- ERP Integration SPEC
- EDI / SOC 2 SPEC
UNDERNEATH EVERY SURFACE
One product record · one price that firms up · one shared credit line · one append-only evidence log. Build any surface on its own — none of them can ever disagree about a product or a price. See the architecture →
The ship order, surface by surface → Product roadmap