Most disappointing consulting engagements were doomed at the brief. The consultant delivered what was asked. What was asked was the wrong thing, and nobody noticed until the invoice arrived.
Which means most of the work in hiring a consultant well happens before you speak to one.
Step one: write the problem, not the solution
The instinct is to brief a solution. “We need a content strategy.” “We need our website redone.” “We need someone to fix our SEO.”
Those are diagnoses dressed as briefs, and if the diagnosis is wrong you'll get a competent solution to a problem you didn't have.
Write the problem instead. What is happening, what should be happening, and what it's costing you. “We generate forty inquiries a month, three become qualified opportunities, and we need twelve. Sales says the leads are poor quality. Marketing says sales doesn't follow up. Neither of us can prove it.”
That brief lets a good consultant tell you the problem is your lead definition rather than your lead volume. The solution-shaped brief doesn't.
Step two: work out what kind of help you need
Four different purchases hide under “marketing consultant.”
Diagnosis. You don't know what's wrong. Short, bounded, produces a recommendation. Two to four weeks. Cheapest and often the highest return, because it stops you spending on the wrong fix.
Strategy. You know roughly what's wrong and need a plan. Four to eight weeks, produces a document and a sequence.
Implementation. You have a plan and need help executing it. Months rather than weeks, and closer to a contractor arrangement.
Ongoing leadership. You need someone to own marketing over time. That isn't consulting. That's a fractional or interim arrangement and it should be structured differently.
Buying implementation when you needed diagnosis is the expensive mistake. You spend six months executing a plan built on a wrong assumption, and everyone involved does good work.
Step three: sourcing
Referrals from founders in your sector are the strongest signal, because the person recommending has to face you afterwards if it goes badly. Ask specifically: who helped you with something like this, and what did they actually change?
Investors have watched several portfolio companies solve versions of your problem and will have opinions about who was useful.
Targeted LinkedIn beats searching for consultants. Find companies two or three years ahead of you, identify who ran marketing during their growth period, and see what those people do now. That finds operators with relevant experience rather than people who are good at describing themselves.
Industry communities and peer groups. Slower, high quality, worth running in parallel.
Marketplaces. Fast, and you do all the vetting. Useful for calibrating rates even if you hire elsewhere.
One caution about publicly prominent consultants. Being visible means being good at marketing yourself. That correlates weakly with being good at your specific problem, and it correlates strongly with being expensive and busy.
Step four: vetting
Two calls, and a specific structure for each.
First call: do they diagnose or do they pitch?
Describe your problem in three minutes and then say nothing.
What you want: questions. About your sales process, your economics, what you've already tried, who your customers actually are, and what happened the last time you tried to fix this.
What you don't want: a description of their methodology. Anyone who reaches for their framework before understanding your situation will apply that framework regardless of whether it fits.
The strongest signal is a consultant who challenges your framing in the first call. “You've described this as a lead generation problem. From what you've said it might be a qualification problem. Can I ask about your close rates by source?” That's someone thinking rather than selling.
Second call: a specific point of view
Ask for their initial read on your situation. Not a proposal. A view.
Good consultants will give you something real and be clear about their uncertainty. Weak ones will either say they can't comment without more research, which is a dodge, or produce a confident answer that would apply to any company, which is worse.
Ask two more things: what would you need to be true for you to tell me not to hire you? And tell me about an engagement that went badly and what your part in it was. Both answers tell you more than any reference.
Step five: the contract
Keep it short and get these things in it.
- The problem statement, in your words, agreed by both sides. This matters more than the deliverables list, because it's what you'll both refer back to when the work turns out to be different from expected.
- Specific deliverables with dates. “A written positioning document by 30 September,” not “strategic support.”
- Days and who does them. Named person. If a firm, get the actual person named and a clause about substitution.
- IP ownership. Everything produced is yours, including research, frameworks applied to your business, and any tooling built.
- A definition of success you both signed up to before starting.
- An exit. Thirty days' notice. If the first two weeks reveal a mismatch, both of you should be able to stop cheaply.
On pricing: prefer project fees to day rates where you can. Day rates price the consultant's time, which gives neither side an incentive to finish efficiently. Project fees price the outcome. Retainers are appropriate only for genuinely ongoing work, and a retainer that renews without anyone reviewing what it produced is how companies end up paying for something nobody can describe.
Step six: making it stick
Implementation is where consulting engagements die, and the causes are consistent.
Assign an internal owner on day one. Someone senior enough to make decisions, who is in every session, and who will own the output afterwards. Without this, the recommendations arrive with no advocate and lose the first argument they encounter.
Insist on something changing in the first fortnight. Even something small. Engagements that produce nothing visible for six weeks lose the organization's attention and never get it back.
Buy a few implementation hours. Two or three hours a month for the first quarter after delivery costs very little and roughly doubles the chance any of it survives. Consultants rarely offer this. Ask.
Book a review at ninety days. Not to assess the consultant. To assess whether you did anything with it. That review is usually uncomfortable and always useful.
Measuring whether it worked
Set this before you start, because after the fact everyone's memory becomes flexible.
For a diagnosis: did it change what you were about to do? A diagnosis that confirms your existing plan may be correct and may also be a consultant telling you what you wanted to hear.
For a strategy: is anyone using the document three months later? Can two people in the business state the positioning without reading it?
For implementation: the number moved, or it didn't.
And a question worth asking in all three cases: did they tell you anything you didn't want to hear? An engagement where the consultant agreed with you throughout was probably an expensive way to feel reassured.