Most marketing data is noise. Here's how I find the signal.

Every dashboard tells you that something happened. Almost none tell you why.

Traffic went up 30% last week. Cool. Was it the tweet? The newsletter mention? The Slack community where someone shared your link? If you can’t answer that, you can’t repeat it, and marketing you can’t repeat isn’t a strategy, it’s weather.

I’ve spent the last year building linkutm because I kept hitting this wall myself. Here’s what I’ve learned about telling signal from noise.

The vanity trap

Most metrics feel good and mean nothing. Impressions, page views, follower counts: they move, they make charts go up, and they tell you almost nothing about whether your marketing worked.

The test I use now: would I change what I do next week based on this number? If the answer is no, it’s not a metric, it’s decoration.

Impressions went up? I don’t change anything. Decoration.
Signups from one specific post went up? I write more like that post. Signal.
A channel sent 500 visitors who all bounced in 10 seconds? I stop investing there. Signal, the uncomfortable kind.

Three questions that actually matter

When I open analytics now, I’m only trying to answer three things:

Which channel sent people who did something? Not visited: signed up, replied, came back. A hundred visitors who convert beat ten thousand who scroll past.
Which content made them come back? First visits are luck. Return visits are product-market fit for your writing.
What would happen if I doubled down on only one thing? This is the scary one, because the honest answer usually means dropping four channels you’ve been comforting yourself with.

Everything else in the analytics tab is entertainment.

Tag everything, or you’re guessing

Here’s the boring, unglamorous truth: most founders can’t answer question one because their links are naked. A link shared in a newsletter looks identical to the same link shared on X, in a DM, or in a community. The traffic all arrives as “direct”, the analytics equivalent of a shrug.

UTM tags fix this, and almost nobody uses them consistently, because tagging every link by hand is tedious. So people do it for one campaign, feel organized, and quietly stop.

My rule now is absolute: no link leaves the house without a tag. Every post, every reply, every newsletter, every DM where I share something I made. It takes seconds and it’s the difference between “traffic went up” and “that one reply in a founder community sent 40 signups.”

(Yes, this is the exact itch linkutm scratches. I built the tool because I broke my own rule too many times.)

Small experiments, honest verdicts

The last piece: signal only appears when you give it a chance to. If you post everywhere, all the time, in different formats, you’ve built a system where nothing is measurable: too many variables moving at once.

What works for me is boring: one variable per week. One new channel, or one new format, or one new topic, tagged, measured, and given an honest verdict on Friday. Keep what moved signups. Kill what didn’t, even if it was fun.

Weather becomes climate when you measure it long enough. That’s the whole game: turn “something happened” into “I know what I did, and I can do it again.”

This is post two of writing in public. If you’re fighting the same dashboard fog, email me. I collect these stories now.