There's a sequence that runs through every chapter of Earn the Change: process first, people second, tools last. It's the most practical rule in the book, and the one most often broken.
The temptation is understandable. Tools are concrete. You can buy one this afternoon. Process and people are messier. They involve conversations, habits, and admitting that the way things work today isn't the way anyone designed them.
Why the order matters
Think about what a tool actually does. It takes a process and makes it faster, more consistent, or less manual. That's valuable when the process is right. When the process is wrong, the tool makes the wrong thing happen faster, more consistently, and with less human judgment in the way to catch it.
Research on failed technology projects tends to tell the same story. The software usually works. What fails is everything around it: unclear goals, a process nobody agreed on, and people who were never asked. Gartner, for example, expects many agentic AI projects to be canceled because of rising costs and business value nobody can pin down. Those aren't technology problems. They're process problems that showed up wearing a technology label.
Process first
Before anything else, get clear on the work itself. What's the recurring process? Where does it start and end? What does "done well" look like? Where, specifically, does it slow down or break?
Most of the time, the answer isn't "everything." It's one step. The handoff where nobody knows who owns it. The form that always comes back half filled. The approval that sits for three days because the approver is on site. Fixing that one step often matters more than anything a tool could add.
People second
Once the process is clear, the next question is who runs it and who decides. A change that makes sense on paper can still fail if the person doing the work wasn't part of shaping it, or if nobody knows who gets the final call when something unusual happens.
This is where trade-offs live. Every change makes someone's day a little different. Naming that honestly, and agreeing on it, is what keeps a good change from quietly being undone a month later.
Tools last
Only now does the tool question get interesting. With a clear process and agreed roles, you can ask a sharp question: which step would a tool make meaningfully better, and how will we know? That's a question a vendor can actually answer. "Can your product help our business?" is not.
Tools last doesn't mean tools never. It means tools earn their way in, the same way any change does.
Exhibit 1
Then, from Habit, loop back
- ↺ Back to ObstacleWhat holds it back now?
- ↺ Back to the listRescore, pick the next one, start at Why
- ↺ Back to WhyWhen tools, goals, or volume change
The loop needs an owner
A 30-minute review, monthly or quarterly: rescore, check, pick next.
Source: Earn the Change, Troy Green
A simple test
Next time someone suggests a new tool, ask three questions before anything else:
- Which process is this for, exactly?
- What step in that process is the problem?
- Who will own the result, including the mistakes?
If those have clear answers, the tool conversation will be short and productive. If they don't, you've just saved yourself a subscription.

