The test that matters
Most consulting engagements are judged on results during the engagement. Pipeline created, CAC reduced, conversion improved — all measured in the window where the consultant is present and the client's attention is highest.
That's the wrong test, and I say that as someone who spent three years optimising for it.
The right one is twelve months later. Is the thing still running? Has anyone touched the model? Did the weekly meeting survive? Or did the whole apparatus quietly revert within two quarters of the last invoice?
I've been on both sides of that outcome enough times to have a view about what separates them, and it has very little to do with the quality of the strategy.
Why good work reverts
It depended on a person who left. The RevOps manager who owned the scoring model moves on, and nobody else understands it. Six months later it's ignored, and nobody can say when that started.
It required effort nobody had allocated. The weekly pipeline review works brilliantly while a consultant chairs it. Once that ends, it competes with everything else in the calendar and loses within a month.
It was never actually adopted, only accepted. The recommendations were approved, the deck was well received, and behaviour never changed. This is the most common failure and the hardest to detect, because everything looks fine right up until you check the numbers a year later.
It was too sophisticated for the team that inherited it. A well-tuned attribution model or a complex scoring system that only its author can maintain will be abandoned the moment its author is gone. This is the failure I'm most personally guilty of.
What actually persists
Four things, in rough order of durability.
1. Changed definitions.
The most durable thing you can leave. A joint MQL definition that sales and marketing negotiated together outlives every dashboard, because it lives in people's heads and in how they talk about the work.
Definitions are cheap to create and almost impossible to un-learn. If I could only leave one artefact, it would be this.
2. Recurring meetings with an owner and a format.
Not the meeting itself — the format. Four questions, a named chair, one decision required before it ends. Formats survive; agendas don't.
The critical detail is naming an internal owner from day one and having them run it while you're still there. A meeting you chair is a meeting that ends when you leave.
3. Simple models people can explain.
A six-criterion scoring model that any rep can reconstruct in their head will outlive a well-tuned model nobody understands, even though the second one is more accurate.
This is the trade-off I got wrong for years. Accuracy that depends on the author's presence isn't accuracy — it's a temporary loan.
4. Instrumentation.
An activation event defined and tracked persists because it's in the system, not in a person. Once "activated users" appears on a dashboard next to signups, it keeps informing decisions without anyone maintaining it.
What reliably doesn't persist
- Anything requiring manual weekly effort that isn't in someone's job description
- Documentation nobody has read since it was written
- Tools bought during the engagement and configured by the consultant
- Strategy documents, in almost all cases
The last one is worth sitting with. The deliverable clients most often ask for is the one least likely to change anything twelve months on.
How I structure engagements now
Name the internal owner in week one. Not the sponsor — the person who will run this afterwards. If there isn't one, that's the first thing to fix, and if it can't be fixed I'd rather decline the work.
They do it, I watch. Backwards from how most engagements run, and much more uncomfortable for everyone. But the first pipeline review I chair is a demonstration; the first one they chair is a transfer.
Build the simplest version that works. Then resist improving it. The sophisticated version is for me; the simple version is for them.
Do a handover review, not a final report. Sit with the owner, run the process together, and find out what they can't yet do unaided. That's a more useful last day than presenting results everyone already knows.
Check in at six months, unpaid. Partly because I want to know, and mostly because it's the only honest feedback on whether the work was any good.
What this costs
It makes engagements less impressive. A simple model and a meeting format is a less compelling deliverable than a comprehensive growth programme, and it prices lower.
It also means the clients who get the most durable value are the ones least likely to need me again — which is an awkward incentive to sit with, and the main reason I think this test is worth applying deliberately rather than hoping it takes care of itself.
The measure I'd want to be judged on: a year after I leave, could someone tell that I'd been there? Not because my name is on a document, but because the way the team works is still different.
Hilal Tasdan
B2B SaaS Growth Marketing Consultant & Fractional CMO. Partner in Growth.