The Working Hypothesis: Auditing the Model Your Adviser Has Already Built of You

Specialist advisers are paid to be quick. That speed comes from pattern recognition: a sector archetype, a handful of comparable mandates, a set of public filings, and a first conversation with the sponsor. From these they construct a working hypothesis of the organisation — how it earns money, where its costs sit, how much risk the board will tolerate, who really decides, and which subjects are politically live. The hypothesis is rarely written down and almost never tested. It then quietly conditions everything that follows.

That is not a criticism of advisers. A hypothesis is what makes an early conversation useful rather than merely exploratory. The difficulty is that the model is invisible to the client, and by the time recommendations arrive it has usually hardened into something that reads like fact. Boards that want genuine specialist insight, rather than well-presented inference, need a way to inspect the model early.

How the model forms before the brief is finished

Three sources do most of the work. The first is the sector template: assumptions about margin structure, capital intensity and regulatory posture that hold broadly across a peer group and imperfectly for any single member of it. The second is public material — annual reports, credit ratings, press coverage, tribunal or regulatory outcomes — which is accurate as far as it goes and usually a year or two behind the operating reality. The third is the commissioning conversation, in which a sponsor describes the problem in the language of the outcome they want.

Each source is reasonable. Together they can produce a coherent picture of a company that does not quite exist. An adviser may assume a centralised cost base in an organisation that has devolved spending to divisional boards. They may read a cautious board as risk-averse when the caution is the residue of one specific event. They may assume a single executive sponsor with clear authority when the reality is a committee of near-equals and an incoming chair.

None of this surfaces in a steering group. It surfaces eighteen weeks later, in a recommendation that is internally logical and addressed to the wrong organisation.

An audit that fits on a single page

What follows is deliberately modest. It is a question set, not an interrogation, and the best advisers will treat it as a sign that the client intends to be a serious counterparty.

Ask the lead adviser, in writing, before substantive work begins, to answer the following:

1. In no more than five sentences, describe the organisation you believe you are advising: its economics, its principal constraints, and where decision rights sit.
2. Which elements of that description are sourced? Name the source for each.
3. Which elements are inferred? What evidence would overturn each inference?
4. What would typically be true of an organisation like yours that might not be true here?
5. What have you been told that you have not yet been able to verify?
6. Which recommendation in your draft approach would you withdraw if the answer to one of the above turned out differently?

Question four does most of the diagnostic work. It invites the adviser to name their own generalisations, which is a far more revealing exercise than asking whether they have any assumptions.

There is a companion exercise for the client side. The sponsor should record, separately and before reading the adviser's answers, what they actually conveyed in the commissioning conversation — what was said, what was implied, and what was left out for reasons of time or sensitivity. Comparing the two documents is usually instructive and occasionally uncomfortable.

Reading the responses

Three patterns matter.

Provenance or fluency. A response that is confident, well-written and entirely unsourced tells you the adviser is reasoning from archetype. That may still be right, but the board should know it is buying a hypothesis rather than an assessment.

Borrowed certainty. Watch for assertions that clearly derive from another client's situation. Sector experience is valuable; transplanted facts are not. An adviser who can distinguish the two is worth considerably more than one who cannot.

Willingness to be corrected. The strongest responses are the ones that nominate specific facts that would change the adviser's view. That is the mark of an analytical position rather than a sales position.

Equally, the exercise will occasionally reveal that the board's own self-description was the weaker document. Boards describe themselves aspirationally — the operating model as designed rather than as practised, the risk appetite as stated rather than as revealed by recent decisions. The audit is therefore a check on both parties, which is precisely why it belongs in the governance record rather than in a private email.

A cadence, not a one-off

An assumption audit performed once at commissioning is useful and incomplete. The model drifts as the engagement proceeds and as the organisation changes around it. Four triggers are sufficient for most mandates:

• At commissioning, before the brief is finalised, so the answers can shape the scope.
• On any change of sponsor, chair or chief executive, because the adviser's picture of decision rights will be out of date within days.
• After a material event — a profit warning, a restructuring, a regulatory intervention, a change of ownership — where the prior hypothesis may now be actively misleading.
• Before final recommendations are tabled, as a short re-test of the two or three assumptions on which the advice most depends.

A brief note recording each re-test, appended to the engagement file and summarised in a single page of the board pack, is enough. Directors are entitled to rely on specialist advice, but that reliance is better founded when the information base beneath it has been examined rather than assumed.

What it changes

The immediate effect is a better brief. The second effect is a better question from the board table: not whether the adviser is credible, but which version of the organisation their advice is addressed to. The third is a useful sorting mechanism. Advisers who welcome the audit tend to be the ones whose conclusions survive contact with the operating business. Those who find it intrusive have usually told you something about the confidence they place in their own model.

Correcting an assumption at week two costs a conversation. Correcting it at the point of implementation costs a programme. The audit exists to move that correction as far forward as it will go.