Leave Yourself a Way Out

I.

I'm less worried about picking a tool we'll eventually replace than picking one we can't learn our way out of.

Those sound like the same problem. They're not quite.

Replacing software takes work. People have to get comfortable again. Some of what we built won't come with us. That's part of making a commitment and it isn't automatically a reason to avoid one.

But there's another cost that doesn't show up until we try to leave. We can't separate what we learned about the client's marketing from the way this particular tool taught us to describe it.

The export has all the rows. The reasoning is missing.

II.

Take a hypothetical agency using a platform to review search performance. The team can save annotations beside a chart. Over time those notes explain why certain pages were changed and why other pages were left alone.

Some notes describe facts. Some record a best guess. A few document an idea the team decided against.

When the agency moves platforms it brings over the performance data. The new dashboard looks complete. But the notes don't fit anywhere obvious so they stay in the old account until nobody wants to keep paying for it.

Now the next person sees a page that underperformed and a date when the team changed it. They can make a pretty convincing story out of that. They just can't see that the people doing the work had a different explanation at the time.

We didn't lose the data. We lost some of our ability to disagree with the story the data seems to tell.

That's the part I'd want to protect at TMG before a tool becomes the only place our thinking lives.

III.

An export button doesn't settle this.

You can export a score without knowing how it was calculated. You can save a report without preserving which parts the team didn't trust. The file exists... but somebody still has to figure out whether it means what it appears to mean.

So I'd test the exit with one actual recommendation. Give another capable person the exported evidence and the notes that would survive the move. Can they understand why we made that call? Can they tell what would have made us choose differently?

If they need the person who built the report to explain everything from memory, the export hasn't really carried the work forward.

That doesn't mean documenting every conversation. A few lines about an important decision can be enough. The question is whether those lines belong to the account or disappear with the software subscription.

IV.

There is a reasonable argument for keeping everything together.

People are more likely to leave a useful note where they're already working. Send them to three other places and the record might get worse. A beautifully independent knowledge system nobody updates doesn't solve much.

I'd rather let the team work where the work makes sense. Then agree on which decisions need a record that can survive the tool.

A change in campaign direction probably qualifies. Every small adjustment doesn't. The distinction should follow what another person would need to understand later, not how much we can persuade people to document.

That gives the platform room to be useful without asking it to become our permanent memory.

V.

It also changes how I think about trying new software.

A pilot can show us that a tool saves time. It takes a different kind of test to find out whether the time saved leaves us more capable or just more dependent on that tool's version of the work.

Suppose the system makes a certain recommendation easy to produce. The team gets good at producing it. Eventually we have more examples supporting that approach because it's the one we've had the most chances to try.

Leaving now feels risky. We have experience with this method and less experience with the alternatives. But part of that difference was created by the tool we picked.

I don't think we can remove that influence completely. We can keep enough of the underlying evidence to revisit it. Especially the cases that didn't fit neatly.

VI.

A way out gives us something useful even if we never take it.

It makes staying a decision we can still examine. We can look at a new platform without having to defend every judgment stored in the old one. We can keep the work that holds up and reconsider the parts that depended on how the software happened to organize it.

Before the next renewal I'd try moving one account's reasoning. Not the whole account. Just one decision with enough context for somebody else to pick it up.

That should tell us more than another tour of the export settings.