How Do You Price a Pilot So It Becomes a Contract?

Should you discount a pilot to win the deal?

Rarely, and almost never to zero. A pilot priced at nothing teaches the buyer that your work is free to try, and the conversation that follows is about whether to start paying rather than about whether to continue. Those are very different conversations and only one of them converts reliably.

I price pilots at a real number, and I lose some deals because of it. The ones I win convert at a rate that makes the trade obviously correct, and the ones I lose were usually never going to buy the full engagement anyway.

What follows is how I think about sizing, pricing and structuring a pilot so that it leads somewhere rather than ending politely.

What is a pilot actually for?

It is a paid way to remove the buyer's biggest specific doubt. Not to demonstrate everything you do, not to prove your credentials in general, and not to give a discounted taste of the full engagement. One doubt, named out loud, tested in a bounded piece of work.

That framing does most of the useful work, because it forces you to ask what the doubt actually is before you propose anything. Sometimes it is whether the approach works on their data. Sometimes it is whether you are easy to work with. Sometimes it is whether the internal champion can get budget, in which case a pilot is a political instrument rather than a technical one.

If you cannot name the doubt in one sentence, you are not proposing a pilot, you are proposing a small project and hoping. Small projects end when they end. Pilots end with a decision, and the difference is entirely in how you set them up.

What does a free pilot cost you?

More than the hours. A free pilot arrives with no internal sponsor, because nobody had to defend a budget line for it, and work that nobody defended internally gets deprioritised the moment something else goes wrong. You end up chasing a client for the inputs you need to do the work you are giving them.

It also sets the reference price. Everything you do afterwards is compared against a version of you that was free, and the buyer's mental anchor for the value of your work is now zero. You can recover from that, but you are working uphill for the rest of the relationship.

The third cost is selection. Free pilots attract people who want a free thing, and paid pilots attract people with a budget and a problem. If your pipeline is thin, the first group feels like progress. It is not, and the month you spend servicing them is the month you did not spend finding the second group.

How should you size a pilot?

Small enough to finish in weeks, large enough to be genuinely useful on its own. If the deliverable would be worthless without the follow-on engagement, you have built a demo rather than a pilot, and the buyer can feel the difference immediately.

My test is whether the client would be glad they did it even if they never hired me again. An audit they can act on, a Zapier or Make workflow that runs, a Webflow page that performs. If the honest answer is that the output only makes sense as the first chapter of something bigger, I make it smaller and more complete instead.

Time-boxing matters as much as scope. A pilot with no end date drifts, and a drifting pilot never reaches the conversion conversation because there is no natural moment to have it. I would rather cut the scope to hit a date than extend a date to fit a scope, because the date is what creates the decision.

What should the pilot price actually be?

Enough that somebody had to approve it, and low enough that approval does not require a committee. The exact figure depends on your market, but the principle is that you are pricing for the approval process rather than for the hours, and those two numbers are often very different.

I work on fixed fees, with most projects falling between one thousand and ten thousand dollars, and a pilot sits at the bottom of that range rather than outside it. Pricing a pilot below the level that requires any thought at all removes the very signal you wanted, which is that somebody chose to spend money on this.

Whatever the number, make it a fixed fee rather than an hourly estimate. A pilot billed hourly invites the buyer to watch the clock rather than the outcome, and it converts the trust-building exercise into a procurement exercise. I have written before about why I charge fixed fees instead of billing by the hour, and it applies doubly here.

What has to be agreed before the pilot starts?

Four things in writing. The doubt being tested, what the output will be, the date, and what happens next if it goes well. That last one is the one everybody skips, and skipping it is why so many successful pilots end in a polite thank you.

Agreeing the next step in advance costs you nothing and changes the entire shape of the ending. It does not have to be a commitment to buy. It can be as simple as agreeing who will be in the room for the decision and roughly when. What matters is that the conversation has a scheduled home rather than needing to be manufactured by you afterwards.

Write down the inputs you need from them too, whether that is access to their HubSpot, their Airtable or their analytics, along with what happens if those inputs are late. A pilot that overruns because the client took three weeks to send access is still a pilot that overran, and without the clause you will end up absorbing it. That is the same discipline as knowing what to do when a fixed-fee project goes over scope.

How do you handle the conversion conversation?

Hold it on the date you agreed, present what you found, and make a specific recommendation including the option of not continuing. The recommendation being genuinely open is what makes the rest of it credible, and buyers can tell the difference between advice and a sales pitch wearing advice as a costume.

Come with a number and a scope, not with a promise to send a proposal. The energy of a pilot is highest on the day you present it, and the week you take to write a document is the week a competing priority arrives. If the full engagement is genuinely hard to scope until the pilot is done, prepare two versions in advance and pick one in the room.

If the answer is no, ask what would have had to be different. That answer is worth as much as the deal you did not get, because it is the only reliable way to find out whether your pilot is testing the right doubt. A pattern of no answers with the same reason is a pilot design problem, not a sales problem.

When should you refuse to run a pilot at all?

When the buyer cannot name what would count as success, when the real decision-maker will not be involved, and when the pilot is being used to compare you against two competitors doing the same unpaid work. In that last case you are not a candidate, you are free labour in a procurement exercise.

The no-success-criteria case is the most common and the most tempting to accept. It feels like a client who is flexible. It is usually a client who has not decided what they want, and a pilot cannot resolve that, because there is nothing to test. Send them away to decide and offer to help them decide, which is itself a piece of paid work.

Refusing is easier when your pipeline is healthy, which is the real argument for keeping it healthy. It is also worth knowing your own economics well enough to say no with confidence, which is why I think reading CAC payback properly when you sell services matters more for a small practice than most people assume.

What should you do next?

Take your last three pilots or trial engagements and ask which doubt each one was testing. If you cannot answer for one of them, that is the one that did not convert, and now you know why rather than blaming the client's budget cycle.

Then write your standard pilot as an actual document. The doubt, the output, the date, the fixed price, the inputs you need, and the agreed next step. Having it written means you propose it in a day instead of a week, and speed at that moment matters more than polish.

Finally, put a real number on it and hold that number for the next three opportunities. You will lose one of them and learn more from the other two than a year of discounting would teach you. If you want a second pair of eyes on how you have structured a pilot before you send it, reach out.