Count decisions, not competitor reports

How many competitor briefs did your team publish last quarter? Now the harder question: how many of them changed something?

If you can answer the first one instantly and the second one not at all, your competitive intelligence program is measuring the wrong thing. And it is not a small mistake. The metric you pick decides what the program turns into.

Output metrics reward the wrong behavior

The usual CI scorecard looks productive. Twelve briefs. Forty alerts routed. Six battlecards refreshed. A Slack channel with 300 posts.

Every one of those numbers goes up when the program works harder. None of them goes up when the program works better. A team graded on briefs writes more briefs. A team graded on alerts lowers the threshold until the alert count looks healthy. The scorecard turns into a production quota, and production quotas are easy to hit with material nobody reads. We have written before about the #competitors channel nobody reads; output metrics are how that channel gets built.

Worse, output metrics make the program look strongest exactly when it is drifting. A quiet month with three sharp findings scores lower than a noisy month with thirty forgettable ones.

A decision is the only unit that matters

Intel has value at one moment: when somebody does something different because of it. A pricing page change that made you hold your own price increase. A competitor's new integration that moved a roadmap item up a quarter. A removed free tier that sales used in four deals that week.

Those are decisions. They have an owner, a date, and a "before" and "after." You can write them down. That makes them countable, which is the whole point.

So keep a decision log. One line each:

• What changed on the competitor side, with the date and the URL.
• Who acted on it.
• What they did differently.

That is the scorecard. If the log has two entries at the end of the quarter, the program produced two decisions. That is a true number, even if it is an uncomfortable one.

"But most intel is context, not decisions"

This is the strongest objection, and it is half right. Plenty of good intel only sharpens a mental model. A PM who reads competitor docs every month makes better calls without ever pointing at one brief.

The answer is not to drop context work. It is to stop counting it as the result. Context is a cost you pay so that decisions get better. If a year of context produces zero logged decisions, the context did not land, or it went to the wrong people. Either way you want to know.

The log also exposes something output metrics hide: which sources feed decisions. In most teams, a handful of surfaces — pricing, plan limits, integrations, hiring — carry nearly every entry, and the rest is noise. That is a good argument for tracking fewer competitors and fewer pages, with more attention on each.

What this changes about tooling

Once you count decisions, the job of a monitoring tool gets narrower. It does not need to produce more. It needs to put a specific, dated change in front of the person who owns the decision, fast enough to matter. That is the part Seeto is built for: it watches competitors' public pages continuously and reports each diff as a discrete change event, so the decision log can cite exactly what changed and when. It does not make the decision, and it does not replace someone reading the change and thinking about it. Nothing should.

The same logic applies to rituals. A meeting that exists to review output is a meeting that exists to justify output. That is why we argued for cancelling the weekly competitive review and routing changes to owners instead.

Here is the uncomfortable test. Open your last CI report and ask, for each item, "who did what because of this?" If the honest answer is mostly blank, the program is not failing for lack of effort. It is succeeding at the metric you gave it.

Change the metric.