The situation
Most of the regulated Salesforce programmes I’ve recovered came to me looking similar from the outside.
A multi-million-pound budget, half-spent. A steering committee meeting that everyone dreads. A partner
relationship that’s gone from collaborative to defensive. A team that knows it’s not working but
hasn’t been allowed to say so. And, usually, an integration with the customer’s underlying systems that
nobody fully understands.
The pattern is rarely that the technology is broken. The pattern is that the operating model around
the programme is broken. The data model decisions got punted to the next sprint. The renewals motion
that was supposed to follow the implementation isn’t designed yet. The success criteria changed three
times and nobody updated the plan.
What I actually do
The first six weeks is diagnosis, not recovery. I sit in the stand-ups, read the architecture decisions,
talk to the ten people who actually know what’s going on (they’re never the ten on the steering committee),
and write a one-page assessment that tells the board what we’re going to do and what we’re going to stop
doing. The hardest part of that page is usually the “stop doing” list.
Then we restructure the programme into something that can actually ship. That nearly always means
fewer parallel workstreams, a smaller team for the first ninety days, and a public commitment to
re-baseline. We rebuild the data model conversations from first principles — what are the entities, who
owns them, what’s the source of truth, what happens at renewal. We re-earn the steering committee’s
confidence through small, visible wins on a weekly drumbeat, not big quarterly promises.
I’ll be in the room. I’m not the kind of recovery lead who writes the strategy paper and leaves the
team to deliver it. The team needs to see that the new direction is being held to from the inside, including
on the hard calls — the difficult vendor conversation, the scope cut the customer doesn’t want, the
person on the team who’s lovely but in the wrong role.
The result
Across CloudShift and VASS the recovery playbook drove a 40% reduction in delivery overruns at the
practice level — which is the way you actually measure whether the change held, not how a single
programme finished. Individual programmes shipped to production after recovery, with operational
ownership transferred cleanly to the customer’s BAU teams.
The thing I won’t promise is the heroic save. Most recoveries finish six months later than the original
plan, at a margin you wouldn’t have chosen, and with a customer who’s still cross with you for a quarter
or two. They finish. That’s the point.
When this kind of brief is right
If you’re a CIO, COO or CTO with a programme that needs an honest read and someone who’ll do something
about it: a six-week diagnostic is usually the right starting point. Fixed fee, written report, no
commitment to a longer engagement at the end. I’ll tell you whether the programme is recoverable, what
it’ll cost, and who you actually need on it.