Then stop measuring the consultancy by the reports.
A report is useful when it causes a better decision or a useful change. If every review produces another deck, another action tracker and another workshop while the same problems remain, the consultancy is documenting failure rather than fixing it. A report can be useful. A report that exists because the previous report recommended another report is paperwork breeding in captivity.
If you want someone to look at the recommendations and tell you what is actually worth doing, get onto us. We are much more interested in the answer than how many pages it takes to present it.
Ask what changed because of the last report
Pick the previous review. Which recommendations were implemented? Which outcomes improved? Did release lead time fall? Did defect escapes change? Did environment availability improve? Did management get a clearer risk view? If the old report offered forty actions and no movement, begin by asking which review recommendations really matter.
If nobody knows, commissioning another assessment is unlikely to create enlightenment.
Find out whether the recommendations were actionable
‘Improve governance’, ‘enhance quality culture’ and ‘increase automation maturity’ sound respectable and can mean almost anything. Useful recommendations identify the problem, consequence, proposed change and who needs to decide or act.
If a recommendation cannot be translated into a concrete decision, challenge it before adding it to a tracker.
Stop outsourcing ownership of improvement
Consultants can identify problems, propose changes and help implement them. The organisation still needs owners who care whether the change sticks. A programme cannot subcontract its appetite for making difficult decisions.
Sometimes everyone already knows the answer but prefers a consultancy to keep studying it because the actual fix is politically awkward: reduce a supplier team, stop a pointless test phase, invest in environments, delay a release or make one leader accountable.
If the action from ‘too much governance’ is to establish a governance improvement workstream, the system has developed antibodies.
Commission the next piece of work around a decision
Instead of ‘review our testing’, ask ‘why does regression take three weeks and what should we stop or change?’ Instead of ‘assess governance’, ask ‘why does nobody own the release risk?’
A narrow question can produce far more value than a broad diagnostic programme. If the consultancy cannot work without expanding every problem into a transformation roadmap, you may be buying consultancy-shaped activity rather than improvement.
Put a stop condition on diagnostic work
Every review should have a point at which enough is known to recommend action. Otherwise discovery can continue indefinitely because there is always another stakeholder, data set or maturity dimension to inspect.
Agree the questions up front and define what evidence would be sufficient to answer them. If new information changes the scope materially, discuss it. Do not quietly expand the engagement because broader analysis produces more billable certainty.
There is a useful professional skill in knowing when to stop analysing. At some point the organisation has to do something with the answer.
Pay attention when every problem becomes a new workstream
A consultancy can make money by decomposing one difficult problem into many respectable-looking streams: governance, capability, tooling, operating model, culture and maturity. Sometimes that decomposition is necessary. Sometimes it is how a clear problem becomes a programme. The best consultancy deliverable is sometimes an email saying ‘stop doing these three things’. This will upset PowerPoint but it will recover.
Keep asking which pieces are essential to the original decision. You can always widen the work later. It is much harder to shrink a consultancy programme after thirty people have been given named roles in it.