A rank tracker whose job isn't done in the dashboard

6 min read

Most rank trackers are dashboards. You log in, you look at a chart, you close the tab. The product's value is defined as the thing you see while you're looking at it.

I think that's the wrong shape for this particular job, and the argument is narrow enough to state in one sentence: a rank is only interesting at the moment it changes, and you have no way of knowing when that moment is. So a tool that requires you to look is asking you to guess.

The thing you have to remember to open

Rankings move on their own schedule. A competitor publishes, you push an update, Google adjusts something. None of that is on your calendar.

So checking a dashboard has exactly two failure modes, and you're always in one of them:

• You check too often. Most days nothing moved, so you've spent attention to learn nothing, and after two weeks you stop.
• You check too rarely. Something dropped eleven days ago and you're finding out now, which is ten days after it was cheap to fix.

In practice everyone lands in the second one, via the first. The habit decays because the dashboard is boring precisely when it's telling the truth — and it's boring most of the time, because most days nothing happens. Rewarding you for looking is a bad deal when there's usually nothing to see.

The inversion is obvious once you say it out loud. Don't ask the user to poll. The tool already checks every day; it already knows the moment something changed. It should be the one to speak.

What that means concretely

So the interface is an email that arrives when a position moves, and nothing at all the rest of the time. The unit of value isn't a screen, it's a notification you didn't have to earn by remembering.

To be clear about what I'm not claiming: there is a dashboard. Keyword history, movement over time, who else is on page one. I'm not proud of removing features. The claim is narrower — the dashboard is where you go when you've decided to investigate something, and that's a rare, deliberate act. It's not where the product does its job. Being clear about which one is which changed a lot of decisions downstream.

The three things this makes harder

This shape has real costs, and I'd rather write them down than pretend the trade was free.

You lose browsing. A dashboard is good at questions you didn't know you had. You go in to check one keyword, notice something odd about another, and follow it. An email can't do that — it can only tell you the thing it already decided was worth saying. If your work involves genuine exploration across hundreds of keywords, you want the dashboard, and you probably want a suite. I'm not the tool for that, and pretending otherwise would just produce a bad version of both.

The threshold is the whole product. An alerting tool lives or dies on what it decides is worth interrupting you for. Alert on every one-position wobble and you've built a noise generator — people filter you to a folder within a week, and a filtered alert is worse than no alert, because now you believe you're covered. Alert too conservatively and you miss the drop you needed. There's no neutral default here; the threshold is an opinion about what matters, and getting it wrong is the main way a product like this fails.

Silence has to be trustworthy. This is the one I underestimated. In a dashboard product, if the checker breaks, you open the page and see stale data — annoying, but visible. In an email product, a broken checker looks exactly like good news. No email means nothing moved. No email also means the scan died three weeks ago.

Those are indistinguishable to the user, and the failure is silent in the worst direction: the user is calm, and shouldn't be. So an email-first tool takes on an obligation a dashboard doesn't — it has to know when it failed, and it has to be more suspicious of its own health than a dashboard ever needs to be. Most of the unglamorous work in this product is there rather than in the feature list.

The business tension worth naming

There's an awkward consequence. Dashboard products measure engagement — daily actives, session length, and the graph that goes up when people visit more. Those numbers are how you tell your product is working.

This design deliberately drives them to zero. A user who never logs in and gets an email once a month when something moves is a completely satisfied user. By dashboard metrics they look like they're churning.

So you have to pick a different measure of success, and be honest that it's harder to read. Mine is whether people keep paying while never showing up — which is a slower, quieter signal than a usage chart, and easier to fool yourself about in either direction.

I don't think that's a reason to build the dashboard instead. But it does mean the shape of the product commits you to a shape of company: no engagement loop to optimise, no reason to add a feature because it increases visits, and no way to grow by making the tool stickier. The only lever left is whether the thing is genuinely useful when it speaks.

Which is roughly where I wanted to end up anyway. The best outcome for a tool like this is that you forget it exists until the day it tells you something you needed to know.