The Review to Run Before You Call a Consultant

A diagnostic starts from a symptom. Margin is soft, a launch missed its number, a team keeps rebuilding the same plan — someone names the problem, and the diagnostic works backward from that sentence to find what actually produced it. The diagnostic you run before you hire anyone is a description of that process, and it assumes the symptom already exists.

A business review has no symptom to start from. It looks at the whole business on a schedule, the way a physical checkup happens whether or not anything currently hurts. Nobody has to be worried for a review to be worth running, which is exactly why most companies do not run one until they are.

The two are not competitors. A review that is already current is what makes a later diagnostic fast, because the diagnostic no longer has to build its own baseline before it can use it.

What the earlier piece assumed

The argument for running a diagnostic before a shortlist rests on a claim: the evidence is already in the building, in documents nobody wrote for this purpose. That is true in a company that has been looking at itself regularly. It is not true in a company that only opens its own numbers when a quarter goes wrong, because in that company the baseline was never written down in the first place.

A standing review is what keeps that claim true. At minimum, it holds a small, consistent set of facts current enough to be trusted the day someone needs them:

• Revenue and margin cut by segment or cohort, not blended into one line that hides which part is actually paying for the others.
• The two or three initiatives that consumed the most senior time last quarter, regardless of what the plan said would.
• Which of last year's stated priorities were actually resourced, versus named and then quietly deprioritized.
• Where headcount grew, and whether the reason it grew still applies.

None of this requires new instrumentation. Most companies already have the underlying data somewhere. What is missing is the habit of assembling it on a schedule, before anyone needs it for a decision that is already under pressure.

A review built from these four holds up because none of them require anyone's cooperation to stay honest. They come from systems the company already runs for other reasons — billing, the calendar, the org chart — so there is no incentive sitting inside the number to make it look better than it is by the time someone finally reads it.

A review that only happens after something has gone wrong is not a review — it is the first week of a diagnostic wearing a different name. The value of a standing review comes specifically from its indifference to how the quarter feels. It runs when the numbers are good and nobody wants to look at them as carefully as when they are bad, which is precisely when a slow separation between the headline metric and the thing underneath it is easiest to catch and cheapest to fix.

This is also why a review should never be asked to produce a finding. A diagnostic is allowed to conclude something and recommend a change; that is its job. A review's job is narrower — keep the description of the business current — and asking it to also justify itself with a discovery is how reviews quietly turn into occasional, dramatic events instead of a standing practice. The unglamorous version is the one that survives a good quarter.

The alternative is familiar enough to have a shape: a leadership team that only opens its own numbers in earnest once a board member asks a hard question, at which point the review and the crisis arrive in the same meeting, and nobody can tell whether the finding caused the alarm or the alarm produced the finding.

The company that skips straight from a bad quarter to a consultant search is not skipping a step so much as outsourcing it, at a markup, under time pressure. The early phase of any competent engagement — interviews, document review, reconstructing what actually happened last year — is the review the company never ran, now being assembled for the first time by people who do not yet know where anything is kept.

That is not wasted work exactly. It has to happen somewhere. But it is expensive to buy on demand what could have been maintained continuously and for a fraction of the cost, and there is a second cost that does not show up on the invoice: a company with no current review has no independent record to check a consultant's framing of the problem against. The diagnostic's own account becomes the only account, because it is the only one that exists.

A company that keeps its own review current enters that same conversation differently. It can say which parts of the picture it already trusts and which it does not, which narrows what an engagement needs to cover and sometimes removes the need for one — the review having already answered the question a consultant would otherwise have been paid to ask first.

There is a third cost that is easy to miss because it never appears as a line item: time. A diagnostic run against a review that already exists can move directly to the constraint, because the description of the business is not itself in question. A diagnostic run with no review has to spend its early weeks establishing facts that should never have been in doubt, and every week spent proving what the review would have already shown is a week the actual constraint sits unexamined.

The pattern underneath all of this is the same one that shows up across the pillar this note belongs to: the quality of a decision is set well before the decision is made, by whatever description of reality was available going into it. A diagnostic can only be as fast and as honest as the review that came before it, and a company with no review is always diagnosing from zero, on the clock, at consulting rates.

This is also where the review changes what kind of help is worth buying at all. Some consulting engagements exist to build a company's first working review, not to run a one-time diagnostic on top of a review that never existed — and knowing which one an organization actually needs is itself information a current review would have already surfaced.