# Understanding statistics

**URL:** <https://discourse.mozilla.org/t/understanding-statistics/18538>\
**Category:** addons.mozilla.org\
**Created:** [August 29, 2017, 1:19pm UTC](https://discourse.mozilla.org/t/understanding-statistics/18538 "2017-08-29T13:19:00Z")\
**Posts on this page:** 9\
**Page:** 1

<div class="post-metadata">

**Author:** ![naber](https://sea1.discourse-cdn.com/flex001/user_avatar/discourse.mozilla.org/naber/32/21062_2.png) [@naber](https://discourse.mozilla.org/u/naber)\
**Post date:** [August 29, 2017, 1:19pm UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/1 "2017-08-29T13:19:00Z")

</div>

Hi, I wonder why not all downloads of my add-on help increase the daily user count. Thanks to being featured, the add-on has about 6,000 downloads per day, but the daily user count increases only about 1,500 per day. There are some uninstalls of course, but those are less than 500 per day. So are most of the installations from users who don’t use Firefox on a daily basis? What else might explain the difference?

---

<div class="post-metadata">

**Author:** ![erosman](https://sea1.discourse-cdn.com/flex001/user_avatar/discourse.mozilla.org/erosman/32/15447_2.png) [@erosman](https://discourse.mozilla.org/u/erosman)\
**Post date:** [August 29, 2017, 6:16pm UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/2 "2017-08-29T18:16:41Z")

</div>

One recent feature is that the daily users now show those who have the addon enabled.

In the past, it showed people who have the addon installed (Enabled or disabled) but now it only shows the ones that are enabled.

---

<div class="post-metadata">

**Author:** ![Pierre-Adrien](https://sea1.discourse-cdn.com/flex001/user_avatar/discourse.mozilla.org/pierre-adrien/32/22680_2.png) [@Pierre-Adrien](https://discourse.mozilla.org/u/Pierre-Adrien)\
**Post date:** [August 30, 2017, 8:27am UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/3 "2017-08-30T08:27:20Z")

</div>

> [@erosman](#):
>
> One recent feature is that the daily users now show those who have the addon enabled.
> 
> In the past, it showed people who have the addon installed (Enabled or disabled) but now it only shows the ones that are enabled.

Still, it is weird that he gets 6000 DL a day, but an increase of only 1500 users per day (it would mean 75% of the people installing the addon uninstall or disable it right away?).

Maybe there something more in the user count calculation?

---

<div class="post-metadata">

**Author:** ![DaveRo](https://avatars.discourse-cdn.com/v4/letter/d/838e76/32.png) [@DaveRo](https://discourse.mozilla.org/u/DaveRo)\
**Post date:** [August 30, 2017, 9:22am UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/4 "2017-08-30T09:22:43Z")

</div>

I’ve also wondered how these stats were derived. I assume ‘users’ is actually instances of Firefox - e.g. if, as was recently the case, I have release, beta, and nightly on both a desktop and Android then that’s 6 users.

Are updates perhaps counted as downloads? Does syncing have any odd effects?

One of my addons has 430 downloads and 12 users - 2 of which are me. My guess is that these downloads were bots or Mozilla processes but it’d be nice to know. I cited it in a bug report so that will account for a few.

FYI Here’s the change erosman referred to: [https://github.com/mozilla/addons-server/issues/3632](https://github.com/mozilla/addons-server/issues/3632)

---

<div class="post-metadata">

**Author:** ![naber](https://sea1.discourse-cdn.com/flex001/user_avatar/discourse.mozilla.org/naber/32/21062_2.png) [@naber](https://discourse.mozilla.org/u/naber)\
**Post date:** [August 30, 2017, 8:12pm UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/5 "2017-08-30T20:12:26Z")

</div>

Could it be related to the way my add-on is featured? It’s on `about:addons`. Anything special about that page?

---

<div class="post-metadata">

**Author:** ![jorgev](https://avatars.discourse-cdn.com/v4/letter/j/71c47a/32.png) [@jorgev](https://discourse.mozilla.org/u/jorgev)\
**Post date:** [September 6, 2017, 7:58pm UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/6 "2017-09-06T19:58:50Z")

</div>

We haven’t looked into it (recently), but there are many possible situations where a download doesn’t add to your user count:

- Failed downloads.
- Downloads using other browsers.
- Users downloading because they think that’s how they get the latest version.
- Old users uninstalling and new ones replacing them in the stats.
- Web bots.

---

<div class="post-metadata">

**Author:** ![naber](https://sea1.discourse-cdn.com/flex001/user_avatar/discourse.mozilla.org/naber/32/21062_2.png) [@naber](https://discourse.mozilla.org/u/naber)\
**Post date:** [September 6, 2017, 8:28pm UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/7 "2017-09-06T20:28:25Z")

</div>

Thanks, I didn’t know that web bots are counted. These reasons apply to all add-ons, so I’ve checked the numbers of some other add-ons whose stats are publicly available. It turns out that what I see is quite common. It’s actually common for one download to cause less than 0.1 daily users (looking at all-time downloads).

- AdBlocker Ultimate: 4.6m downloads, 0.38m daily users = 0.08 daily users per download
- The Camelizer - Price Tracker: 0.42m downloads, 0.038m daily users = 0.09 daily users per download
- Google search link fix: 1.3m downloads, 0.06m daily users = 0.04 daily users per download
- Youtube Best Video Downloader 2 - MP3 MP4 1080P: 3.7m downloads, 0.1m daily users = 0.02 daily users per download
- Status-4-Evar: 1.8m downloads, 0.1m daily users = 0.05 daily users per download

---

<div class="post-metadata">

**Author:** ![naber](https://sea1.discourse-cdn.com/flex001/user_avatar/discourse.mozilla.org/naber/32/21062_2.png) [@naber](https://discourse.mozilla.org/u/naber)\
**Post date:** [September 7, 2017, 7:39am UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/8 "2017-09-07T07:39:21Z")

</div>

> [@jorgev](#):
>
> Failed downloads.  
> Downloads using other browsers.  
> Users downloading because they think that’s how they get the latest version.  
> Old users uninstalling and new ones replacing them in the stats.  
> Web bots.

While these are valid reasons, I don’t think they can explain why the downloads started to grow massively when our add-on became featured on `about:addons`. Why should bots or users of other browsers suddenly start downloading an add-on when it’s featured on some internal Firefox page (I don’t think `about:addons` is available on the web)?

> [@jorgev](#):
>
> Old users uninstalling and new ones replacing them in the stats.

Uninstallations aren’t a huge issue for us, we can count them because users get sent to an uninstall page. To me, it still feels like either the daily user count is too low or the download count is too high. Maybe you could look into this?

---

<div class="post-metadata">

**Author:** ![DaveRo](https://avatars.discourse-cdn.com/v4/letter/d/838e76/32.png) [@DaveRo](https://discourse.mozilla.org/u/DaveRo)\
**Post date:** [September 7, 2017, 8:07am UTC](https://discourse.mozilla.org/t/understanding-statistics/18538/9 "2017-09-07T08:07:17Z")

</div>

While these anomalous download numbers are interesting, I think the number of actual users is more important. Could sync’d devices be counted as one user? Or, if a sudden drop in the number of users would be unacceptable, could the number of ‘duplicate users’ be captured?

I assume that the average user has two devices - but maybe that’s way out.