The SnapStore team has been working recently on improving the download statistics for snaps.
The statistics are visible in the dashboard only to their publishers (or collaborators) via the Stats menu-link on the snap page as you can see in the example below:
Our goal is to make it possible for developers to identify successful initiatives and existing trends regarding their user base and react accordingly.
Downloads, as a reflect of activity, is only one (the first) aspect of the data we would like to provide about snaps in the Store, there is more to come.
We count on your collaboration and feedback to continue to improve this experience for snap developers and surface mindful and relevant information that trully helps everyone to have better software.
I like the look of this update but this is effectively hiding information that we used to have access to before.
The old design used to let us (albeit rather painfully) see up to the top 20-25 countries and versions whereas now we only get the top 10 with everything else hidden in the “Others” title.
Could the number of rows be made configurable somehow?
Or even better, can we get a REST API or similar endpoint where we can access the entire dataset?
The reasons for reducing the initial number of categories (territories and versions) is exactly what you already hinted, the previous limit cluttered the UI and instead of seeing more you end up seeing nothing
We can certainly try design a more compact UI capable of showing more categories at once, but If the goal is to have access to the entire dataset, let’s jump straight to the API and allow authenticated users to fetch the pristine data. I think it’s a great idea and will enable publisher to do detailed analysis if they need.
When talking to developers there’s certain things they’ve asked for from the dashboard.
More flexible x axis - only showing the last 30 days is insufficient. Let the developer choose the horizontal scale -
years/months/days.
Profile of releases - how many people are sticking on one track / release, and not updating. How can I tell which releases people have, and how many have uninstalled vs keeping it installed, vs people upgrading.
Personally I’d like to see a raw data download option. While you could develop a plethora of graphs, having the raw data for each application I have, would enable me to track trends, compare application performance and otherwise flexibly graph the data. An API which let me grab that data would be awesome.
My interpretation of the old statistics was that most people downloading my app were deleting it shortly afterwards. It would be nice to know how many people were actually using the app. Probably too much to ask for as this might infringe privacy.
Yes, it would not be reasonable to show usage details for the application, and we don’t collect that anyway, but it sounds fine to show some sort of refresh statistics, which will at least provide a hint of how many systems have kept it around.
@jsmolka can you elaborate on your needs for download statistics ?
The current weekly active devices metrics offer a much more accurate picture of the snap growth and trends, that’s the general feedback. Nonetheless, please walk us through your case so we can figure something to help you.
Meanwhile, snap download metrics are regularly collected and available from the snap publisher API as linked at the bottom of the snap stats page. You can easily fetch and display them locally.
@jsmolka, your statement is not right per graph description:
“Weekly active devices” tracks the number of unique devices that have contacted the Snap Store using a 7 day window, broken down by channel.
In your particular case, 3 days offline will not show a decrease in the installed-base corresponding to the ‘offline PC’, but a 7d-long offline period will.
We have deliberately chosen one week period, because we wanted to smooth the “daily churn” (it’s perfectly normal for a laptop be off for few consecutive days but rarely longer than a week) but eventually account for lost installed-base when it’s the case.
Perhaps you have a concrete example of a snap of yours where the active installed base trends displayed do not reflect your expectations. I would be more than happy to analyze the data with you.
