Where it started
A 14-person coffee roaster in Williamsburg reached out because their operations manager was drowning. Retail, wholesale, DTC — six accounts, three spreadsheets, and a morning routine that had quietly grown from 20 minutes to two hours.
The ask when they first emailed was, word for word: "we need a new ERP."
We didn't quote an ERP.
What we found by listening
A real conversation with the ops manager, then a meaningful stretch of time spent watching him work. The picture that emerged was different from the ask.
The issue wasn't that they needed new software. It was that three pieces of information — online sales, wholesale orders, and roasting output — were each accurate, but lived in three different places and nobody was connecting them. Every discrepancy required a scavenger hunt through inboxes. The "ERP" he had in his head was really just "I want these three things to agree with each other."
An ERP would have been a six-figure project, a six-month rollout, and a mountain of change management. Nothing about the business actually called for it.
What we built instead
A focused middleware layer. It watches three sources — the online store, a shared wholesale sheet, and the roasting log the team was already keeping — and reconciles them into their existing accounting tool continuously throughout the day.
Crucially, we replaced nothing. The team kept using every tool they were already using. The fix was a quiet layer of glue, invisible unless it caught something.
On top of that, a morning digest lands in the ops manager's Slack: yesterday's reconciliation summary, any flags (an inventory drop, a pricing anomaly), and a running tally for the week. The two-hour morning routine became a five-minute read.
The outcome
- 8 hours a week returned to the operations manager. Most of them going to the part of the job he actually enjoys — tasting, sourcing, talking to roasters.
- Zero reconciliation errors in the months following launch.
- A pricing discrepancy caught in week two that had quietly been costing them margin on one of their biggest wholesale SKUs for months — a five-figure annual recovery on its own.
- A wholesale expansion they'd been postponing for two years. With the ops manager's mornings free, they took on three new wholesale accounts that pushed annual revenue ~35% higher in the year following the build.
- An ERP project avoided.
"I didn't realize how much mental overhead the morning was costing me until it was gone. I'm actually using that time to taste new lots now — which is why I started this company in the first place." — Operations Manager, the Williamsburg coffee roaster
Why it worked
Because we listened before we prescribed. The ask was an ERP. The actual problem was a reconciliation gap that no ERP would have solved without years of overhead and most of an annual budget. Naming the real problem — and then refusing to scope the bigger version — was the work the engagement was paid for.
The code lives on their infrastructure, owned by them, with documentation a competent engineer could pick up. The work pays its own way every quarter, in time returned and margin recovered.