A pilot should help a client decide whether an approach is ready to expand. Agree on a limited scope and success criteria, then choose a sample that represents the conditions the wider rollout will face. For a migration, that means more than checking whether a few files copy: permissions, metadata, user validation, and the working environment matter too.
A connectivity test proves that systems can talk to each other. It does not establish how the work will perform with a representative volume of content or whether users can do their jobs afterward. Include enough real conditions to expose those gaps, and record what the pilot has not tested. A clean demonstration tells you little about the difficult cases it never touched.
Use what the pilot reveals to update the method and estimates for the next phase. Do not let additional departments, workflows, or cleanup work slip into the pilot simply because they are part of the eventual ambition. The client should leave with evidence of what worked and an honest view of what remains, including work that needs separate planning and funding.
This post is part of my 25 lessons from 25 years in consulting series, exploring what it takes to build a consulting firm that delivers for clients and supports the people doing the work.
From leadership and employee experience to sales and delivery, the series comes back to one idea: a consulting company should be its own best client.
Have a lesson of your own? Share what has worked, what hasn’t, or what you wish more consulting firms understood in the comments.