Pilot

Running a successful clinic pilot

How to scope a four-week pilot, define success criteria before starting, choose the participants, and reach a decision you can defend at the end.

10 min readUpdated

A badly scoped pilot does not give a bad answer. It gives no answer at all, and after a month the clinic is left with contradictory impressions and nobody able to decide.

This guide is about running the pilot rather than about the product: what to settle before starting, who takes part, what to observe along the way, and how to reach a decision that holds up in front of your partners. It applies just as well if the final answer is no.

Three decisions before day one

These three questions are settled before installation. Postponing them is the most common reason a pilot ends without a conclusion.

Decision 1The scope

A specific subset of the caseload rather than the whole clinic. One condition, or one type of follow-up, or the patients of a single clinician.

A pilot extended to everyone tests your team's ability to change habits, not the tool.

Decision 2The duration

Four weeks suit most clinics: long enough to get past the novelty effect, too short to settle in by default.

Set the end date at the outset and put it in the calendar. A pilot with no end date becomes an adoption that was never decided.

Decision 3Who decides

A named person who will decide at the end, and the people they will consult.

Without this, the decision falls to whoever has the most opinions rather than to whoever has the most context.

Writing the success criteria before starting

This is the step everyone skips, and it is the one that determines whether the pilot serves any purpose.

A useful criterion can be verified without discussion. "The team likes it" is not a criterion: nobody can demonstrate that it has been met or missed. Two or three criteria are enough, and it is better that they be modest and verifiable than ambitious and vague.

They must be written down, dated and shared before the first session. A criterion formulated after the fact is always adjusted, unconsciously, to what has been observed.

Examples that work

  • Objective progress data recorded for ten patients
  • A progress report produced in under five minutes
  • Adherence to home programmes measured rather than estimated

Examples that do not work

  • "See whether it improves our practice"
  • "Check that the team is comfortable"
  • "Compare with our current system", without specifying on what

Add a stopping criterion at the same time as the success criteria: the condition that would make you stop before the end. A clinic that has not decided in advance what counts as failure almost always finds a reason to extend.

Choosing the participants

The reflex is to hand the pilot to the most enthusiastic. That is a mistake, because their verdict will convince nobody: it will be attributed to their enthusiasm.

The opposite reflex, handing the pilot to the most sceptical in order to win them over, fails for the symmetrical reason.

Take two or three clinicians in the middle, curious without being advocates, and include the front desk. The front desk sees the administrative friction clinicians do not notice, and its view carries more weight at the moment of decision than anyone expects.

Allow time for it. A pilot added on top of an already full week mainly measures the team's fatigue.

The week-by-week sequence

WeekWhat happensWhat you observe
0Kick-off meeting, written criteria, installation and trainingNothing yet. Week 0 exists so that nothing is improvised
1First use in real conditionsWhat blocks in use, not what people like
2Mid-point review on your own dataThe gap between the criteria and reality
3The team works unaidedWhat survives once the novelty has passed
4Decision meetingThe criteria, and nothing else

Week 2 is the moment a pilot is saved or lost. If a criterion is clearly out of reach, two weeks remain in which to understand why: a missing setting, a misunderstood step, or a genuine limitation of the product. These three causes lead to different decisions, and only the third is a reason to decline.

Keep a log of irritants over the four weeks, one line per incident, with its date. At the end, memory retains the most recent problem rather than the most frequent. A dated list corrects that distortion, and it is your best negotiating tool with the vendor.

The decision meeting

One hour, with the participants and the person who decides.

The agenda has three items, and it should be followed in this order. Take up the criteria written in week 0, one at a time, and rule on each: met, missed, or not measurable. Then go through the log of irritants and separate what is a matter of habit from what is a limitation of the product. Finally decide: adopt, extend with a modified scope, or stop.

The two forms of drift to watch for are predictable. The first is relitigating the criteria rather than applying them, which amounts to never deciding. The second is concluding on the general impression left by the meeting, which depends mostly on who speaks loudest.

A criterion that is not measurable counts as missed, and it indicates that it was badly written. Note it: that is useful for the next pilot, whoever the vendor is.

If the answer is no

A pilot that ends in a refusal has not failed. It cost four weeks and saved you a far more expensive decision.

Three things are worth doing before closing the file.

Write down the reason in one precise sentence. "Too expensive" and "the report does not replace our template" do not lead to the same place, and the second can be addressed.

Retrieve your data. What was entered during the pilot belongs to the clinic, regardless of the decision. Request the export before access is closed.

Send the log of irritants to the vendor. A documented refusal is worth more to them than a silent adoption, and it puts you in a better position if you reassess in a year.

Frequently asked questions

How many patients should be included?

Enough for the criterion to be verifiable, rarely more. Ten patients are enough to demonstrate that progress data is recorded correctly. Fifty prove nothing further and add to the team's workload.

Should we keep using the old system in parallel?

Yes, for the scope that is not in the pilot. No, for the scope that is: entering the same patients twice distorts the measurement of time, which is often one of the criteria.

What if a clinician disengages along the way?

Ask them why, privately and quickly. It is almost always a specific missing function rather than a rejection of the principle. The information is useful even if you cannot address it during the pilot.

Can the pilot be extended?

Once, with a modified scope and new written criteria. An extension with no change of framing does not produce a better decision, only a later one.

Who pays during the pilot?

This is agreed in advance and written down. At Bio6, the pilot is supported and carries no commitment, and the terms are set out on the Pilot programme page.

Further reading

More guides