Strong Products Create Their Own Signal

If you need a complex dashboard to know whether something is working, it probably isn't.

That sounds harsh. It's not. The products that actually matter don't need a twelve-tab analytics setup to prove it. They surface signal on their own. Through behavior. Through pull that nobody had to manufacture.

The ones that don't? They hide behind dashboards. And the dashboards always look fine. Dashboards are great at looking fine. That's sort of the problem.

The tell

You can feel when something is working. Not gut instinct. Behavior.

People come back without being reminded. They tell other people without being asked. Support tickets shift from "how do I do this" to "can it also do this." Usage grows in places you didn't market to. Nobody on the team can explain why signups spiked in Brazil but here we are.

Sean Ellis built a framework around this. Ask users one question: "How would you feel if you could no longer use this product?" If 40% or more say "very disappointed" you're probably in good shape. If they don't, no dashboard is going to save you. He benchmarked nearly a hundred startups and the ones below 40% almost always struggled.. regardless of what their other metrics said.

One question. No dashboard. Just whether people actually care. Almost suspiciously simple.

What this looks like

Slack grew almost entirely through word of mouth in its first six months. People used it at one company, changed jobs and brought it with them. Teams adopted it bottom-up. Not because a CIO mandated it but because one person on the team said "just try it" and then suddenly everyone was in there arguing about lunch.

By 2015, a survey of 731 Slack users found 51% would be very disappointed without it. Over 90% of their website traffic was direct. People just typing the URL. No funnel. No retargeting. Just people who already knew what they wanted. That's not a marketing win. That's a product win.

Superhuman had a different path. When Rahul Vohra first ran the Ellis survey in 2017, only 22% hit "very disappointed." Not great. But instead of building more dashboards (tempting, I know) he focused on the users who loved it most, figured out what they valued and what held everyone else back. Three quarters later the score was 58%. He didn't find signal in a dashboard. He found it by talking to people who cared and ignoring the ones who didn't. Revolutionary stuff.

The dashboard trap

Here's where teams go wrong. They build the dashboard first and the product second. They instrument everything, set targets and optimize toward numbers instead of listening to what users are doing.

MAU doesn't mean people care. Session time doesn't mean value. You can make almost any dashboard look healthy if you pick the right window and the right metric. I've seen dashboards that could make a dying product look like it was having the best quarter of its life.

Tobi Lutke at Shopify avoids rigid KPIs entirely. His take is that no single metric can represent the full picture of a complex business. Which is a polite way of saying most dashboards are expensive fiction.

If the dashboard is the only place where things look good.. and nobody can point to a single organic signal that users care.. you don't have a product. You have a really convincing spreadsheet.

What to actually watch for

Pull over push. Are users coming to you or are you dragging them back? If retention depends on lifecycle emails and push notifications, that's not retention. That's life support.

Organic expansion. Is usage growing in ways you didn't engineer? Are people sharing it without a referral incentive? Or do you need to bribe them with credits to tell a friend.

Emotional response. Do people have opinions about your product? Indifference is the worst signal. If nobody loves it and nobody hates it, nobody cares. You want people arguing about your product in a Slack channel somewhere. That's the dream.

Retention without intervention. What happens when you stop sending emails? If the curve drops, the emails were doing the work. Not the product.

The point

Strong products don't need you to go looking for signal. The signal comes to you.