Three very different products are sold under one label. Buying the wrong one is expensive and slow to notice.
"Impact consulting" covers at least three distinct products. They have different deliverables, different durations, different success conditions and different failure modes. They are frequently sold under the same name, occasionally by the same firm, and the buyer often does not learn which one they bought until the engagement ends.
Here is how to tell them apart before you sign, rather than after.
Product one: evidence production
What it is. Someone comes in and produces evidence you do not currently have. An evaluation, a baseline, a theory of change, an impact assessment, a report for a funder.
What you get. A document, and the credibility that comes with it having been produced independently.
When it is the right purchase. When you genuinely lack the evidence, when independence is the point, or when a funder requires an external hand.
The failure mode. You already had the evidence and what you lacked was the ability to act on it. This is the most common misdiagnosis in the sector. The engagement completes, the report is good, and eighteen months later the same conversation is happening with a more recent document. If your last three evaluations reached broadly similar conclusions, buying a fourth is not an evidence problem.
Product two: capability building
What it is. Someone builds your team's ability to do the work themselves. Training, coaching, accompaniment, tools your staff will still be using after they leave.
What you get. People who can do a thing they could not do before.
When it is the right purchase. When the gap is skill, when you will need this capability repeatedly, and when you can protect the time for people to actually learn.
The failure mode. Training people to operate a system that does not exist. If nobody is obliged to use the skill, and no forum requires its output, capability decays within a year and the organisation concludes the training was poor. It usually was not.
Product three: operating design
What it is. Someone rebuilds how impact runs through your organisation. Governance, decision rights, role ownership, operating rhythms, the infrastructure underneath.
What you get. A changed set of arrangements — who decides what, on what evidence, at what point.
When it is the right purchase. When you have evidence and it changes nothing. When impact sits beside the organisation rather than running through it. When the same recommendation has appeared in three consecutive reports.
The failure mode. It is commissioned without the authority to actually change anything. Operating design is redistributive by nature — it moves decision rights — and an engagement scoped as advice will produce advice.
The diagnostic
One question sorts most cases. If a perfect report landed on your desk tomorrow, would anything change?
If yes, you have an evidence problem. Buy product one.
If no, and the reason is that nobody would know what to do with it, buy product two.
If no, and the reason is that nobody is obliged to do anything with it, buy product three. And be aware that this is the one most likely to be quietly substituted for product one, because product one is easier to scope, easier to price and much easier to approve.
Questions worth asking a prospective firm
- What is the deliverable, and what is different in the organisation after it lands? A firm that can only answer the first half is selling product one whatever the proposal says.
- Whose authority changes? If the answer is nobody's, this is not operating design.
- What happens twelve months after you leave? Listen for whether the answer depends on them coming back.
- Whose framework is this? A firm applying a borrowed framework and a firm that wrote one are different propositions. Neither is automatically better — but you should know which you are buying.
Most disappointing impact engagements are not badly executed. They are correctly executed against the wrong purchase, and the mismatch is set at commissioning, months before anyone could notice.