I run a small, free planning poker tool — Scrum Poker Online. Teams open a room, share a link, and vote on story-point estimates. No signup, no setup. It has run quietly for years, funded entirely by ads. Not a business with a roadmap and a team — a side project that happened to cover its own costs and a little more.

In April 2024, the ad revenue roughly halved overnight. Same traffic, same users, half the money.

It took me a while to understand what had happened. The tool’s pages are exactly what you’d expect from an estimation app: a room, some cards, a results panel. Almost no text. Google’s ad system looked at those pages, saw screens with ads but barely any editorial content, and decided the inventory was low value. In ad-network language this shows up as a “smart pricing” adjustment — the system quietly pays out far less per impression because it doesn’t trust the page. For the same ad slot, another network was paying several times more than Google was. That gap was the tell.

So I had a tool that worked well, that people used daily, and that an algorithm had decided wasn’t worth much — because a collaborative tool is mostly interface, not articles.

The things that didn’t work

I did the obvious things first. Opened support threads. Asked for manual reviews. Looked into switching ad partners and escalating through intermediaries. Months of that produced nothing I could measure. The penalty wasn’t a bug to be reversed by the right support ticket; it was a judgment about the content on the pages, and none of my messages changed the content.

Treating it as a content problem, not an ad problem

The reframe that actually moved things: the ad system wasn’t wrong that my pages had no content. So I stopped trying to argue with it and started giving it something to value.

Two prongs.

First, stop being judged by the empty screens. The room pages — the part of the app that is genuinely just UI — got pulled out of Google’s index entirely, with robots rules plus noindex headers. And the screen people land on before a room got a real, substantial explanation of what the tool is and how estimation works, in every language the app supports. If a page is going to carry ads, it should carry something worth reading too.

Second, build the publication that a tool like this never bothered to have. A proper guide to the technique. Honest comparisons with the other tools in the space, including where mine is the wrong choice. An about page with a real name and a real face behind it, marked up so search engines and ad systems can see there’s an actual person and history here, not a content farm. Over months this turned into a small library of articles around the tool.

One concrete example, because it surprised me how much it mattered. The main landing page had the title “Free planning poker tool.” It ranked, but almost nobody clicked — the click-through rate from search was about 0.2%. The word “free” was doing nothing, and the title matched no real search intent. I rewrote it to describe what people actually search for — planning poker, story-point estimation — and the click-through rate on that page went from about 0.2% to about 15%. Same ranking, same page, roughly fifteen times the clicks. I had left that on the table for years.

Where it stands

About thirteen months after the penalty hit, the ad rate showed its first real upward movement — still well below where it was before April 2024, but clearly above the floor it had been stuck on. Organic traffic is up too, because the content now answers questions the tool alone never did.

I want to be honest that this is a recovery in progress, not a comeback story with a bow on it. I can’t prove the content directly lifted the ad pricing rather than other things moving at the same time. What I have is a tool that is now also a small publication, which is a better thing to be either way.

If you monetize a tool with ads, you are a publisher whether you act like one or not. The ad systems will treat thin, contentless pages accordingly — don’t wait for a penalty to find that out.

A penalty like this isn’t a support ticket. It’s feedback about your pages, and the fix is on the pages.

Identity and honesty are content. An about page with a real person, and comparisons that admit where your tool loses, aren’t fluff — they’re signals.

And check your search titles. Ranking and click-through are different problems, and the second one is often a five-minute fix you have been ignoring for years.