Pilot Success Criteria: How to Define Them Before You Start

Pilot success criteria are the written, measurable thresholds that decide whether a pilot turns into a paid contract. Good ones name one primary business metric and two or three secondary ones. They carry a deadline and a named decision-maker, and they are agreed with the buyer before the pilot begins. If you are a founder or early sales rep selling B2B software, this is the most valuable hour of the whole pilot.

This guide is about customer pilots: the paid or unpaid trials you run with a prospect to earn a contract. It is not about airline pilots or clinical pilot studies, though the logic of setting a bar before you test is the same.

TL;DR

• Pick one primary outcome metric and two or three secondary metrics, all measurable, all tied to the buyer’s business, not to “we liked it.”
• Get the economic buyer, not only the daily user, to agree to the criteria in writing before the pilot starts.
• Set the threshold a little below your typical results, and set an activation milestone with a date so you can intervene early.
• Define what happens next (price, term, signature) before day one, or the pilot will stall.
• Judge the result against the criteria on the decision date. Do not let the bar move afterward.

Table of Contents

• Why the criteria matter more than the pilot itself
• What good pilot success criteria look like
• How to write the criteria, step by step
• Benchmarks to sanity-check your plan
• Define what happens next
• Why successful pilots still fail to scale
• A worked example: a pilot scorecard
• Our perspective: criteria are a sales conversation
• Get the criteria right on the call
• Common mistakes to avoid
• FAQ
• Sources

Why the criteria matter more than the pilot itself

A pilot is a sales stage, not a technical experiment. The Edmired validation guide argues that a missing co-authored agreement is a leading reason “successful” pilots never convert: the buyer redefines success afterward to avoid making a purchase decision. You have probably seen it. The users are happy, the dashboard looks fine, and the champion says “let’s give it another month.”

The numbers back the discipline. A 2025 study of enterprise software companies reports that pilots with documented success criteria, set timelines and pre-negotiated conversion terms converted at 68%. Free pilots with vague metrics and no conversion plan converted at 23%. Treat those as one source’s figures rather than a law of nature. The gap is large enough to act on.

There is also a cost argument. If a buyer will not co-author measurable criteria, GTM Skills notes that you have learned something cheaper than a three-week evaluation. A prospect who will not define success was probably never going to buy.

What good pilot success criteria look like

Strong criteria share a few traits. SaasDash describes them as specific (a number or verifiable outcome, not an impression) and achievable within the pilot window. Add two more from our own experience: they are owned by the buyer, and they can be checked by someone who was not in the room.

A decision gate, not a sentiment survey

Traction Technology frames criteria as a decision gate. “Users liked it” is a vanity metric. A pilot succeeds when the tool clears an objective threshold that justifies its cost, such as cutting report generation from four hours to 45 minutes, whatever people thought of the interface.

One primary metric, two or three secondary

DoWhatMatter recommends one primary business outcome metric and two to three secondary adoption or value metrics. Their example of a primary metric is reducing manual reconciliation from five steps to two. The primary metric answers “is this worth paying for?” The secondary metrics answer “will it stick?”

A weak criterion and a strong one, side by side:

How to write the criteria, step by step

1. State the question the pilot answers. One sentence, such as: “Can this tool cut our onboarding time by a third with our current team?” If you cannot write it, you are not ready to start.
2. Interview the economic buyer. The daily user tells you what is annoying. The person who signs tells you what is worth paying for. DoWhatMatter stresses that the economic buyer, not only the user, must agree to the criteria. If you are mapping who that is, our guide to how enterprise B2B buying works covers the buying group.
3. Choose the metrics. One primary, two or three secondary. Each needs a baseline, a target, a data source and an owner.
4. Set the threshold with a margin. More on this below.
5. Add an activation milestone and a date. The smallest observable action that proves the buyer saw value on their own data.
6. Fix the timeline and decision date. Put the review meeting on the calendar before kickoff.
7. Write the conversion path. Price, term, who signs, and what happens if the thresholds are met.
8. Get it signed. An email reply of “agreed” from the economic buyer beats a verbal nod. A one-page document is enough.

Set the bar a little below your average

A new environment adds friction: messy data, a skeptical team, a security review that eats week one. The Rework playbook suggests setting thresholds slightly below your typical performance. If your tool usually improves efficiency by 40-60%, a 35% threshold leaves room for a rough start without killing a deal that still delivers real value. Do not set it so low that hitting it proves nothing to the buyer’s CFO.

Use an activation milestone

Broad feature testing wastes pilots. GTM Skills describes an activation milestone as the smallest observable action, such as one workflow finished on the buyer’s own data, that confirms they felt the core value. If it is not hit by a set day, step in at once. The same page puts conversion for un-activated pilots at only 2-8%. That is why we suggest a milestone in the first third of the pilot, not at the end.

Benchmarks to sanity-check your plan

Do not compare your pilot to a generic conversion average. GTM Skills reports 2025-2026 medians that differ by motion: roughly 14% for self-serve trials, 44% for card-required trials, and 55% for sales-assisted proofs of concept. Your own motion’s range is the fair yardstick. For early-stage teams selling to a handful of accounts, the sales-assisted figure is the relevant one, and it only holds when someone is actively running the evaluation.

A related question is whether to charge. One practitioner’s view on paid versus free pilots is blunt: unpaid pilots are extended demos. You do not have to agree, but a paid or at least resourced pilot, with the buyer committing people and time, does filter out tire-kickers. If you are still shaping the wider motion, see our early-stage B2B sales motion guide for founders.

Define what happens next

A pilot that meets its metrics but has no commercial path usually gets extended instead of bought. DoWhatMatter and Traction Technology both say to define the scale pathway before launch. In practice, write a short “if/then” into the agreement:

• If the primary metric is met and at least two of three secondary metrics are met, the buyer moves to the agreed plan at the agreed price.
• If the primary metric is met but adoption is low, you run a defined 30-day adoption sprint, with a named owner on their side.
• If the primary metric is missed, you hold a post-mortem and either stop or redesign. Both sides know this up front.

That third branch is not pessimism. It makes the buyer take the criteria seriously, and it protects your time. For the mechanics of keeping a deal moving between meetings, our piece on deal advancement techniques goes deeper, and the B2B sales cycle explainer shows where a pilot sits in the sequence.

Why successful pilots still fail to scale

Sometimes the software hits every number and the deal still dies. Dan Ruggles, writing about AI pilots, argues that failure usually comes from governance, not the model. His point is that teams bolt a new tool onto old processes instead of redesigning the workflow around it. His guidance also calls for a business KPI tied to a concrete financial or operational metric before a pilot advances.

For your criteria, that means adding an operational question: who owns this after the pilot, which existing step does it replace, and who approves the change? A pilot that proves one workflow end to end, before anyone tries to scale it, avoids most of this. A good illustration of that mindset is this write-up on proving one workflow before you scale, written for SEO automation but applicable to any tool rollout.

A worked example: a pilot scorecard

Say you sell a workflow tool to a 40-person operations team at a Series B company. Here is a one-page scorecard you could send for signature:

The targets here are placeholders. The structure is the point: every row has a number, a source and a name.

Our perspective: criteria are a sales conversation

We build for founders and reps at seed and Series A companies, and we see the same pattern on their calls. Pilot criteria get set in a rushed ten minutes at the end of a good demo, when everyone is friendly and nobody wants to be precise. The result is criteria like “see how it goes.”

The fix is to treat criteria as part of discovery, not paperwork. The questions that produce a good primary metric are the same ones that qualify the deal: what is the pain costing today, who feels it, who owns the budget, what happens if nothing changes, and what would make this a clear yes? If you run a methodology like MEDDIC, the primary metric and the economic buyer’s sign-off map directly onto the metrics, economic buyer and decision criteria you are already trying to fill in. When those answers are thin, the pilot will be thin too.

We built OffBook around that idea: the moment to get a precise answer is during the call, not in a follow-up email three days later. Live cues remind a rep to ask for the baseline, to name the decision-maker and to confirm the date while the buyer is still on screen. If you want to see how we think about the surrounding conversations, our guides on running demo calls and moving from discovery to demo are good starting points.

— Neil

Get the criteria right on the call

If you are a founder running your own pilots, the hardest part is remembering every question while also listening. OffBook is real-time sales coaching software that shows on-screen cues during video calls, without a bot joining the meeting. It structures prompts around methodologies like MEDDIC and MEDDPICC, and it generates pre-call briefs on the people you are about to meet. Plans are published on the pricing page: Power is $59 per month or $590 per year, and Team is $1,000 per year per seat. You can also see how it works for AI call coaching for sales.

Common mistakes to avoid

• Too many metrics. Seven criteria mean none of them decides anything. Stay at three or four.
• Criteria owned only by the champion. If the person who signs has not seen them, they are not criteria yet.
• Measuring features instead of outcomes. “Used the dashboard” is activity. “Cut cycle time” is value.
• No baseline. Without a before number, an after number means nothing. Capture it in week zero.
• Letting the pilot run open-ended. Put the end date and decision meeting in writing.
• Changing the bar midstream. If something truly changes, amend the document and have the buyer re-confirm. Silent edits are how pilots turn into debates.

If you want a sense of how to track the result on your own side of the table, our explainer on the sales conversion rate and the sales scorecard guide help you measure the pilot stage like any other.

FAQ

How many success criteria should a pilot have?

One primary metric and two or three secondary metrics is a sensible ceiling. The primary one decides whether the product is worth buying, and the secondary ones show adoption and fit. More than that dilutes the decision.

Who should approve pilot success criteria?

The economic buyer, the person who controls budget, plus the day-to-day owner of the workflow. Criteria approved only by a user often get reopened when procurement or finance joins. Get written agreement before kickoff.

How long should a B2B pilot last?

It depends on how long the buyer needs to see the primary metric move on real data. Many practitioners suggest several weeks, and one 2021 guide recommends at least a month and a half. Pick a length tied to your metric, set a decision date, and add an activation milestone early.

What if the pilot meets the criteria but the buyer still hesitates?

This usually means the next step was never defined. Return to the signed document and the conversion path, then ask what has changed since the buyer agreed to it. If nothing has, ask for the decision. If something has, you may have missed a stakeholder.

Sources

• Pilot Customer Success Criteria for B2B Startups, DoWhatMatter
• How to Run a Successful Startup Pilot, Traction Technology
• Trials and POCs that convert, GTM Skills
• B2B Pilot Program Validation Guide, Edmired
• Deal Closing Pilot Programs: A Proven Playbook, Rework
• SaaS Pilot to Enterprise Conversion, SaasDash
• Your AI Pilot Didn’t Fail Because of the Model, Dan Ruggles
• B2B SaaS Pilots: A Disciplined Approach, Medium

Recommended

• 30–60 Day Pilot to Test Live Sales Coaching for Founder Led Startups
• 60–90 Day Pilot to Prove Live Coaching for Sales and CS Leaders
• Pilot 4 Live Call Coaching Techniques in 60–90 Days, Founder Led SaaS