Honest Marketing Metrics in the AI Era

This is early, seed writing for the site, and it grew out of a longer piece of analysis I did on where marketing measurement is heading. The short version: AI did not just change how content gets made. It quietly broke a lot of the numbers we used to trust, and most dashboards have not caught up.

What actually broke

Two things happened at once.

First, content became nearly free to produce. When anyone can generate a thousand articles in an afternoon, the volume of stuff on the web exploded, and a lot of metrics that assumed content was scarce and effortful stopped meaning what they used to.

Second, traffic itself got noisier. Bots, scrapers, and AI agents move through the web now, and they show up in your numbers looking a little like people. Impressions and raw pageviews were already soft signals. They got softer.

The result is that a lot of the dashboard is now measuring activity, not interest. Big numbers that feel like progress and correlate with almost nothing you care about.

The signals I still trust

When I strip it back, the metrics I believe are the ones that are hard to fake and close to real human intent.

• Branded search. People typing your name into a search box did not happen by accident. They heard about you somewhere and went looking. That is genuine demand, and it is very hard to inflate.
• Genuine engagement. Not likes. Saves, shares, replies with substance, people quoting you to other people. The actions that cost the audience something.
• Word of mouth. "How did you hear about us" answers, direct traffic from real referrals, inbound that mentions a specific thing you made. The oldest signal there is, and still the best.
• Return behavior. People coming back. One visit is noise. A pattern of returning is a person who found something worth their time.

The thread connecting these is that they all cost the audience something, attention, effort, or memory, so they are hard to manufacture. That is exactly what makes them trustworthy.

The numbers I stopped leading with

I did not throw everything out. But I stopped treating these as the story:

1. Raw impressions and reach, on their own.
2. Follower counts as a headline.
3. Total pageviews without any sense of who or why.
4. Any single number going up, presented without context.

These are not useless. They are just easy to move without moving anything real, which makes them dangerous when someone is deciding what worked.

The deeper point

There is a structural reason this matters. When a metric is easy to inflate, it stops measuring the thing and starts measuring how hard people are trying to inflate it. The most gameable numbers become the least informative ones, precisely because everyone games them.

So the move is not to find one new magic metric. It is to shift your attention toward the signals that are expensive to fake and honest about human interest, and to hold your numbers with a little more skepticism than the dashboard invites.

The uncomfortable version of this: a lot of marketing reporting is optimized to look good in a meeting, not to tell the truth. AI made that gap wider. Closing it means being willing to report a smaller, realer number instead of a bigger, softer one.

That is a harder conversation, and it is the right one.

If you want the structured version, Honest Marketing Metrics in the AI Era goes through the whole framework, worksheet included.