Rethinking Chrome Extension Reviews

Update, September 8, 2026: Good news! Oliver Dunk from the Chrome Extensions team replied that they’ve already made changes to focus more on recent reviews. He’s also going to ask the team to look at Weather’s rating spike. Nothing is confirmed for Weather yet, but I’m glad they’re taking a look. My original post is below.

Should a review from eight years ago carry the same weight as one left today? As a Chrome extension developer, I’d like to see older ratings gradually count less. An extension can change a lot in that time, and its rating should reflect how well it works now.

I’ve been building Weather Extension since 2015. It started with a simple idea: show the temperature in my browser toolbar so I could check the weather without opening another website.

Weather grew slowly at first. Then Lifehacker featured it, and more people started using it. Over the years, I’ve added more forecast details, multiple locations, radar, alerts, and plenty of fixes. Eleven years later, I’m still working on it.

You can see how much it has changed in Weather’s visual history. But the ratings from 2017 and 2018 still have a big impact on its overall score.

I believe Weather was hit with review spam. In 2018, it received 2,239 one-star ratings in just 41 days. The overall rating dropped, and afterward I saw installs drop too. It’s been frustrating trying to recover from that while continuing to improve the extension.

Recently, I went back through the Chrome Web Store ratings export to get a better look at what happened. It contains 6,428 ratings from June 24, 2015 through August 20, 2026, including ratings with and without written reviews. The export has the counts, but not the review text.

I mostly remembered the 2018 wave. Looking through the daily numbers, I found another large spike in 2017.

Two periods totaling 72 days account for 2,984 one-star ratings. That’s 91.5% of all the one-star ratings in more than eleven years of data.

Each bar shows one month of one-star ratings. The spikes in late 2017 and 2018 stand out against years with very few.

In the 90 days before each of those periods, there were just five one-star ratings. Then, on October 13, 2018 alone, Weather received 125 ratings. Every one was one star.

The larger surge continued through November 20. Over the next 30 days, there were zero one-star ratings, while 38 other ratings came in.

What stands out to me is how suddenly the ratings started and stopped. There wasn’t a similar increase in the other ratings. Two-to-five-star ratings averaged about 1.3 per day in the 30 days before the 2018 surge and 1.2 during it. One-star ratings averaged more than 54 per day during the surge.

The counts alone can’t prove those ratings were spam, but that pattern is why I believe they were.

The frustrating part is how hard it is for the average to recover. From February 2019 through the end of the export, 1,531 ratings averaged 4.76 stars. The average across the full export is still only 2.87.

Using those totals, it would take roughly 7,300 additional five-star ratings to reach an average of 4.0, assuming no other ratings came in. That’s more ratings than the extension received in the entire eleven-year export.

Even if every one-star rating was legitimate, I’d still like to see this change. If a bad update caused a problem in 2018 and it was fixed, how much should those reviews matter in 2026?

In my experience, unhappy users have a reason to stop and leave a review. Someone who uses an extension every day without any problems may never think to rate it. Fixing a bug doesn’t mean people come back to update their old reviews, either. You can keep improving the extension for years without much change in the score.

I’d like Google to give more weight to recent ratings and gradually reduce the weight of older ones. Keep the old reviews available for anyone who wants to read them. This should work both ways. If an extension gets worse, its rating should reflect that too.

Apple already offers another option. Developers can reset their overall App Store rating when releasing a new version, while keeping the written reviews visible.

Personally, I’d prefer older ratings to count less over time, but I’d like to see Chrome offer a reset option too. I’m still building Weather and fixing issues as they come up. I want someone looking at its rating to get a useful idea of what they’re installing today.