Killed the rebuild. Kept the ERP they owned.
A funded plan to rebuild the core system from scratch.
A mid-sized manufacturer, family-run, around forty people. Their operations system was a decade old and “couldn't do reporting.” A vendor had shown them something modern; the board had blessed a rebuild at roughly six figures over about eight months. The decision was, in their words, “basically made”, they brought us in to sanity-check the build plan, not to question whether there should be one. The operations manager, the person who'd actually live through the migration, was the only one in the room who looked nervous.
Don't rebuild. What you already own does about ninety percent of this.
We opened the hour with the conclusion, not a discovery deck: the rebuild solved a problem they didn't have and left the one they did. The pain was two missing reports and one piece of double-entry between the floor and accounting, none of which needed a new platform, eight months, or six figures. Keep the system. Spend a fraction. Keep the year.
So they could have reached it without us.
We didn't take “it can't do reporting” on faith. We sat with the operations manager and listed only what actually hurt that quarter: late shipping reports compiled by hand, and numbers re-keyed between the production system and the accounting tool. Then we opened what they already paid for. The existing system had a reporting module nobody had ever switched on. The accounting package had an import path they'd never wired up. The “rebuild” was, underneath, a configuration-and-training problem wearing a platform migration's clothes. We put both paths on one page: rebuild, eight months, six figures, real risk to a running factory; or configure what exists, wire one integration, train two people, a few weeks, low five figures, almost no risk. Same gap closed. We showed the arithmetic so the call was theirs, not ours.
We argued ourselves out of the build, the migration, and the bigger invoice.
We told them to kill the rebuild and decline the new vendor. We named the one small piece we'd help with, the single integration, and said the reporting and training could be done in-house or with their current vendor's standard support, which is revenue we pointed away from ourselves on the record. The full recommendation and the number went to them in writing, free, so they could act on it with us, with someone cheaper, or alone. The honest part: the engagement we talked them out of was worth far more to us than the one we kept.
A named person sat across the table and said “don't rebuild”, and put it in writing knowing it cost us the larger piece of work. Not a faceless “we”, not a proposal hedged to keep the door open. That accountability is the whole reason the format works: the no only means something if someone's name is on it.