Insights / People

Change management is the project

What the research says

Prosci surveyed more than 2,600 change practitioners and compared how well each project managed the people side of change with whether it met its objectives. With excellent change management, 88 percent met or exceeded objectives. With good, 73 percent. With fair, 39 percent. With poor, 13 percent, about one in eight.^[1] Prosci summarizes the gap as projects with effective change management being seven times more likely to meet their objectives.^[2]

Why it matters more in a small company

In a 60-person company there's no training department and no change team. One respected salesperson who keeps using a personal spreadsheet can quietly decide whether the whole team follows. Every person's habits count, which cuts both ways: resistance spreads fast, and so does enthusiasm once people see the system saves them time.

Research roundups also flag months three to six after go-live as the riskiest period, when initial training fades and old habits return.^[3] Most project plans have already ended by then.

A practical change plan for 20 to 300 people

1. Name an owner. One leader who uses the system and asks about it in meetings.
2. Design with the people who do the work. Bring service, sales and finance staff into discovery sessions, not just managers.
3. Remove, don't add. Set a date when the old spreadsheet or shared inbox is retired. Two systems in parallel means nobody uses the new one.
4. Train by role, with real data. A service agent needs a different hour than a sales manager, using records they recognize.
5. Measure use weekly for 90 days. Who logged in, which records were updated, where people get stuck.
6. Fix friction fast. A field nobody fills in, or a screen that takes too many clicks, should be changed within days.

How this shows up in our process

Pre-discovery sessions include the people who do the work, not only owners. The proof of concept is tested by real users on real records. Adoption is measured after go-live, because a system that's running but not used hasn't delivered anything.