CONNECT IQ STORE: display the number of downloads not rounded

As they've showed the actual number for 9 years in the web based store, it's not an accuracy issue. I've seen cases where the number wasn't updated until the next day, but that's it

The timing of this change and monetarization mentioned here started me thinking they may be connected in some way.

jim_m_58 said:

As they've showed the actual number for 9 years in the web based store, it's not an accuracy issue.

There's absolutely no way to confirm that. It could've been a race condition on the backend, could've been a glitch in the stats collector, some sort of double counting in some cases but not others, etc. It just didn't matter before, so they never bothered fixing it. But now it does and here's a band-aid fix.

You had to be there.  How long have you been publishing?  Since the rounding started?  What kind of change in the numbers?

I've had apps in the store since 2015, and I know devs that started about the same time, and no one has mentioned the numbers are inaccurate, maybe just delayed updating, like I mentioned. 

If you're talking about what you see in statistics, I don't pay much attention to those numbers,,, The original stats (a few years back) made much more sense to me.

eau_de_vie said:

There's absolutely no way to confirm that. It could've been a race condition on the backend, could've been a glitch in the stats collector, some sort of double counting in some cases but not others, etc. It just didn't matter before, so they never bothered fixing it. But now it does and here's a band-aid fix.

That’s an interesting theory but it doesn’t fit the anecdotal evidence. The only part I agree with is “it matters now”. Not that the download count was glitchy before and it matters now for Garmin to hide that fact (*), but more like the download count was *correct* and precise before, and it matters now for Garmin to hide the previous precision.

(* I don’t understand the reasoning where an *incorrect* exact download count is bad now that monetization exists, but a *correct* exact download count isn’t? It seems to me that in either case, it’s the exact download count which would be objectionable to vendors of paid apps, bc it seems to leak information about app revenue to *everyone*.)

For almost 10 years the CIQ website showed an exact download number (which was received from the backend API). For as long as the CIQ app existed, it showed a “gently” rounded number, similar to how reviews are rounded now (e.g. 123,456 => 123k.)

Now all of a sudden, the backend returns a number that’s brutally rounded to the nearest power of 10 or 5 times power of 10 (e.g. 50,000, 100,000, 1,000,000, 5,000,000), and both front ends have been changed to truncate this number and display it as though it were an estimate (which it is): e.g. 1,234,567 => 1,000,000 => “1M+”.

Coincidentally this has happened at the exact same time official monetization for the store was announced.

Even more coincidentally, this is *exactly* what the Google Play store does.

Strangely enough, there is no problem displaying the exact or gently rounded number of reviews. (Nothing has changed in that respect in the backend API, the website or the app.)

Also, if you know Garmin, their lowest priority appears to be fixing bugs, especially bugs that only affect devs haha. Just look at all the bug reports in the discussion forum and the actual bug reports section. Look at the bug reports in the Fenix 8 forum.

Amazing that Garmin would coordinate a backend API change and 2 frontend changes to fix buggy stats at the exact moment monetization drops, and amazing that it looks just like what the Google Play store displays for download counts. Strange that they would care so much about some glitchy stats when so much broken functionality goes unfixed.

I’ll just say that in the Google Play store, it seems possible for devs to get exact install counts for their own apps. We can only hope that the same thing will come to the CIQ store eventually.

I know this doesn’t mean much, but if you look at the Quora topic on why Google Play doesn’t display exact download counts, its AI summary bot (:vomit:) guessed that one reason is to hide exact revenue numbers. This is exactly what I guessed (before doing any searches or asking any other opinions.) Doesn’t mean it’s correct, but I rly can’t think of a better reason.

As a counterpoint, does anyone know of any online store that shows *exact* purchase counts to customers? Like “exactly 5,675,123 people bought this item”

at least leave the developer the possibility of knowing the exact number.
maybe the app was downloaded 10000 times but in the end it was sold only once.
this would mean that it is not very appreciated and that something needs to be improved.
i have not opened the account as a seller yet.
maybe from there you can see the exact statistics.

dfaina said:

at least leave the developer the possibility of knowing the exact number.

Agree 100%

dfaina said:

i have not opened the account as a seller yet.
maybe from there you can see the exact statistics.

Afaik you don't get additional download stats as a seller, but you do get a sales report. (Someone has already posted a bug about sales reports tho.)

I don't think there is any developer that likes the change that the download count isn't visible any more. Not an improvement at all.

I haven't seen it mentioned but (another excuse for the  implementation could be that a ) static number could help caching as none of the info is changing...

peterdedecker said:

I haven't seen it mentioned but (another excuse for the  implementation could be that a ) static number could help caching as none of the info is changing...

If you mean caching of the web page itself, that can’t be the case since app info (including number of downloads) is returned in an API request that’s fetched asynchronously, not loaded as part of the initial page request itself, as mentioned above. In this thread, there’s literally a link to the app info API request for the GLANCE watch face — you can click on it (no auth required) and see what’s in there:

[https://apps.garmin.com/api/appsLibraryExternalServices/api/asw/apps/07ae0f49-7240-4475-a229-4507e8035fae]

If you mean caching of the app info API request, that also can’t be the case. First of all, no one would ever want a browser to cache that kind of request. Second of all, the response not only includes the (rounded) download count but also the the latest version numbers (internal and external), the timestamp that the app was last “changed”, and the exact review count. All of that other stuff will change fairly frequently. Maybe not as often as the download count, but often enough that you wouldn’t cache it. Besides, exactly what would be saved by caching that one request? Have you seen how bloated the average page is on the Connect IQ store website or Garmin Connect website? Caching that one tiny async request is not gonna make a noticeable difference to the end user or Garmin’s servers.

Idk why Garmin needs a (fake) “excuse” / pretext to make this change — it’s already done. They certainly didn’t ask anyone here for approval, and they certainly don’t care about the uniformly negative reactions in the forums. Maybe their “excuse” is they wanted to copy the Google Play store for absolutely no reason at all. Couldn’t have anything to do with monetization!!!!1!

What Garmin has is a *reason*, and they’re not telling us, so we can only speculate. But all this stuff like “maybe some designer thought ‘1M+’ looks nicer than ‘1.4M’ even though only the download count was changed and not the review count” and “maybe the download count was glitchy but no one cared for a whole 9 years, but now it’s important that that the (incorrect) exact number is hidden and replaced by an estimate — not bc showing the correct exact number would be a problem, but only bc showing an incorrect exact number would be an issue and they can’t fix it!” and “maybe their excuse is they want to make the website more efficient” just does not ring true to me. Since when did Garmin care about fixing bugs that nobody noticed and since when does Garmin care about having efficient websites? Why would a Garmin designer want to make the download count beautiful but keep the review count ugly?

Srsly, am I the only one who thinks Garmin just wants to hide the exact number of downloads (installs) bc they implemented built-in monetization and therefore they don’t want to accidentally disclose the exact (or even “gently” rounded) number of sales for any given app? Who knows, maybe they are even legally obligated to obscure this information. If not, they are probably doing it to avoid pissing off any big players which might want to get in on that sweet monetization action. Maybe they don’t want ppl to know exactly how much money they make off their GoPro app either (just as one example.)

Better yet, just ask yourself why the Google Play store does it this way. Or why the Apple app store doesn’t even have download estimates, just app rankings.

When I click on an item in the Amazon app, the app doesn’t tell me exactly how many people bought the item either. The closest they come is something like “200+ bought in the last month”. Again, note the “+” which indicates that it’s a rough estimate / minimum value. The exact number of ratings is disclosed tho. Hmmm, starting to see a pattern here….

Agree totally.  In fact the statistics section is worse that it was originally.  Originally you could see downloads each week.  Now you can only see % of users using each app version which is less useful.