Six Questions To Ask Before Hiring A Data Consultant

Founders who hire data consultants often end up with reports nobody reads. These six questions separate a useful engagement from an expensive one.

Most founders who hire a data consultant get something back. A dashboard. A model. A pipeline. Something they can point to. What they rarely get is a decision they made differently because of it.

That gap is not a mystery. Firm-level productivity research shows that data work raises output only when it enters actual decisions, not when it produces outputs that sit outside the decision flow. The reports get filed. The dashboard gets bookmarked. The consultant leaves. Nothing changes.

The question that should come before all others

Before you evaluate a consultant's portfolio or ask about their tools, ask yourself what decision you are trying to make better. Not a category of decisions. One specific decision, with a clear owner, that you will face in the next 60 to 90 days.

If you cannot name it, you are not ready to hire a consultant. You are ready to think harder about your business first.

What the six questions actually test

Management consulting research identifies clearly defined goals and active client participation as the primary determinants of project success. Not technical quality. Not the consultant's credentials. The six questions below test whether a proposed engagement meets those conditions.

Ask the consultant: what decision does this output change? If they describe a deliverable instead of a decision, the engagement is scoped wrong.

Ask: who at your firm will own the output after the consultant leaves? If nobody can name a person, the output will not survive contact with a busy week.

Ask: does this require new infrastructure, or does it work with data you already collect? Infrastructure engagements that precede a defined question fail at high rates. Practitioner evidence on business intelligence projects shows the pattern repeatedly.

Ask: can your current staff maintain this without calling the consultant back? The OECD's SME analytics research is specific here: smaller firms gain from analytics when outputs align with existing workflows and skills, not when they require specialist upkeep.

Ask: what does success look like in 30 days? Not six months. Thirty days. If the answer is vague, the engagement has no feedback loop and no natural stopping point.

Ask: what happens if the data does not support the answer you expect? A consultant who cannot answer this has not thought about your business. They have thought about their deliverable.

The infrastructure argument deserves a fair hearing

The strongest objection to decision-first scoping is real. The productivity gains documented in Brynjolfsson et al. and in Tambe's manufacturing plant study accrued to firms that had already built broad data-driven practices across multiple functions. Those firms did not get there by answering one question at a time. They invested in data infrastructure and built the organisational habits that let data enter decisions at scale.

That argument breaks down for small firms on one specific point: it assumes infrastructure investment produces capability that the firm then uses. The OECD SME paper does not support that assumption for firms with thin internal data skills and ad hoc processes. A small firm that builds infrastructure before it has a decision to answer does not automatically develop the habits required to use it. The management consulting research is unambiguous: without a clear goal and an active client sponsor, the project fails regardless of technical quality. Infrastructure engagements without a defined decision fail both conditions at once.

The legitimate concern embedded in the infrastructure argument belongs in your scoping conversation, not as a reason to abandon decision-first thinking. Ask the consultant whether the outputs from a narrow engagement leave something you can build on. That is a different question than asking them to build the foundation before you know what you are building toward.

Where this leaves you

The OECD data shows analytics users grow labour productivity 5 to 10 percent faster than non-users. That number describes firms where data entered decisions. It does not describe firms that commissioned reports.

The six questions above are not a checklist. They are a filter. An engagement that cannot answer all six is an engagement scoped around a deliverable, not a decision. For a small firm with one shot at this before the budget runs out, that distinction is the whole game.