M360 Compass · Management insights
Pilot project for a new tool in a small business

A pilot project helps a small business see whether a new tool improves a task before rolling it out. Pick one workflow, measure how it works today and agree in advance on the conditions that would lead you to expand, adjust or stop the trial.
Imagine an equipment repair shop that takes requests at a service counter. Staff sometimes have to call a customer back when the workshop discovers that a piece of information is missing. A digital form might help. To find out, watch what happens after someone submits it. If a request is easier to open but the technician needs a second screen to find the equipment reference, the work has shifted from one colleague to another. Follow the request to the point where the technician can act on it.
How do you define the scope of a pilot project?
Start with the two people who do the work. Ask the person taking requests what information they need to get from customers. Ask the technician what they need to pick up a job. Follow an ordinary request and one that needed a callback. Identify where the information was missing and how someone spotted the gap. The Government of Canada’s digital service guidance recommends understanding the workflow, testing with users early and setting baseline measures. Government of Canada, guidance on frequent iteration (in French)
In this fictional shop, the manager chooses routine requests received at the counter during the day. Urgent requests and jobs already open go through the usual process. She names one person at the counter and one technician for the trial. She sets a review date and protects two short blocks of time for practice and review. If the team cannot free up that time, she agrees on a smaller trial with the people doing the work. The workload management guide can help you check what fits in the working week.
What should you measure before and during the trial?
Measure a few comparable requests using the current tool first. In the fictional example, a ticket is ready when it includes contact details, the equipment reference and a description of the symptom. The team times the active work at the counter and in the shop, including callbacks. Time spent waiting while nobody works on the request stays out of that total. The team separately checks that every request received can be found in the log at the end of the shift.
| Agreed criterion | Before | During |
|---|---|---|
| At least 11 tickets ready at first handover | 8 of 12 | 11 of 12 |
| Median active work time of 9 min or less | 10 min | 7 min |
| Every request accounted for at closing | 0 missing | 0 missing |
The figures are fictional. The thresholds reflect what this shop needs to make a decision. Set your own before you see the results. Pair each measure with an observation. When a request takes less time, ask where the team saved it. A technician who can read one ticket and get to work may have avoided a trip to the counter. A colleague who stays late to complete records has incurred a cost the team should count.
How can you keep a fallback available?
Before the first live request, demonstrate the whole process with a fictional ticket. Show participants where the authoritative record for each pilot request lives and how to recognize a request that belongs to the trial. Give them a simple way to report a problem. Name the person who can pause the trial and the person who will tell the workshop. The UK government’s Service Manual describes starting with a limited group of users and keeping the previous service available during the transition. GOV.UK, How the beta phase works
The shop agrees to use the old log again if the technician cannot open requests or if an incoming request cannot be found. The manager checks the last confirmed request and reconciles the pilot records before the shift ends. Practise the fallback with sample tickets. Check whether someone else can take over if the designated person is away. The absence management guide walks through the information a colleague needs for that handover.
How do you decide what happens next?
At the review, ask what changed for counter staff and technicians. The Institute for Healthcare Improvement’s Model for Improvement suggests comparing what you observed with what you expected and using what you learned to design the next test. In the example, the team met its criteria on 12 daytime requests. It has yet to observe how the tool holds up at the counter during a rush. The manager adds one busier shift and uses the same measures. Institute for Healthcare Improvement, Testing Changes
If the form makes it easier to open a request but creates more corrections downstream, pick a possible cause you can check. Staff may be reading the field labelled “reference” differently. Clarify the label and try it again with a few comparable requests. If the fallback fails or a request goes missing, pause the trial and restore the agreed process. Whichever decision you make, tell the team what to do on the next shift and when you will review the decision again.
Where should you record the decision and follow-up?
With M360 Momentum, you can record issues and decisions, assign action owners and deadlines, and track actions between meetings. Record who will run the trial on the busier shift, who will reconcile the requests at closing and when the team will review what happened. Keep the measures in the working record your team agreed to use.
Before inviting the first participant, write down the result you want to see and the signal that will prompt a pause. To see how your team could connect decisions with follow-up, request a M360 Compass demo.
Frequently asked questions
How long should a pilot project last?
Allow time to observe several comparable cases and at least one variation that matters to the team. Set a review date. If volume is too low, extend the trial to answer a specific question.
Who should take part in the first trial?
Invite someone who enters the information and someone who uses it later in the workflow. Give both a way to report problems quickly and suggest changes at the review.
Should we keep two logs in parallel?
Choose a single authoritative record for each request. If you need a backup record, specify who keeps it, when they use it and how they reconcile it with the main record.
What if the results are mixed?
Look at which criterion the trial missed and what happened at that point in the workflow. If work can continue within the agreed scope, adjust one element and run another short trial.
