The four questions a Sales Cloud rescue should answer in week one.
Inheriting a Salesforce programme in trouble is, before anything else, a triage problem. The temptation is to open the org, recoil at the technical debt, and start rewriting. Resist it. The first week is for diagnosis, not surgery.
Over a decade of rescues, the same four questions have done more to set direction than any code review. They tell you whether you're looking at a two-week stabilisation or a two-quarter rebuild — and, just as importantly, which of the previous team's decisions are load-bearing and which are simply load.
01What is the org actually for?
Not what the statement of work says. What does the business do on a Tuesday that this org has to support? A surprising number of failing programmes failed because the build chased a process nobody runs. Sit with three actual users before you read a single requirements document.
A failing org is almost never failing at the thing the project plan is worried about.
02Which decisions are reversible?
Some architecture choices are cheap to undo and some are concrete. Data model and integration boundaries are concrete; automation and UI are usually not. Map the irreversible decisions first, because those are the ones worth slowing down to argue about. Everything else you can change later, so don't spend week-one capital on it.
03Where is the trust broken?
By the time a programme reaches a third failed go-live, the real problem is rarely technical — it’s that the business has stopped believing a date. Find the smallest credible thing you can ship in two weeks and ship it. You are rebuilding the relationship as much as the system, and nothing rebuilds it like a delivered promise.
04What does "done" mean to the person who called you?
Get the single metric that, if it moved, would make this engagement a success in the eyes of the person who signed the cheque. Forecast variance. Handle time. Time-to-listing. Write it on the wall. Every architecture decision for the next quarter gets weighed against that number — and the decisions get dramatically easier once there’s only one of them to optimise.
The point
None of these four questions are technical, and that's deliberate. The Salesforce part of a Salesforce rescue is the part we're least worried about. The hard part is pointing a capable team at the right problem — and the first week is the only cheap time to get that right.