Greetings all, and happy 1 month anniversary of the tablet redirect :)
I've got some data to share on new tablet, phone, and app editors and how they stack up against new desktop editors. This is enwiki only for now; I await the glorious return of Dario to help us out with global data :)
*New active mobile editors*
Our metric for new active mobile editors is 5+ edits within 30 days of registration, so we don't quite have the full numbers on the first month of newly registered users yet (because people who signed up yesterday haven't had time to reach active status). But here's today's snapshot of the percent of newly registered users we've converted to active editors in the last 30 days across all out different platforms:
[image: Inline image 1] Note that the trend Dario observed still holds – both newly registered tablet and mobile site users are being converted to 5+ editors at a higher rate than desktop.
Slightly concerning is the conversion rate on the Android app – I'd have expected this to be higher, more in line with mobile site editors on phones. We'll need to figure out if this is due to users editing anonymously on the app instead of signing in and counting toward our new active user count.
*New user productivity*
I was also interested in the volume of contribution we're seeing across each of the platforms, to get a sense of how much new folks are contributing. So I ran the numbers for bytes changed (added, removed, and total changed) by newly registered users across all our different platforms:
[image: Inline image 2] Though mobile site editors (both tablet and phone) are reaching 5+ edits at a higher rate than desktop users, it looks like desktop editors make significantly bigger edits in terms of bytes. Not super surprising, given the smaller screen real-estate and greater difficulty in contributing longform text, but it's good to validate that with real data. Note again that Android app editors are on the lower end of bytes contributed.
To me, these two graphs raise the question of whether mobile site editors (on both tablets & phones) make a larger number of small edits as opposed to one big edit like they might on desktop. I don't know if we'll be able to answer that definitively anytime soon, but it's definitely something to think about and dig into more in the future...
Thanks for putting this together -- this is really interesting data. It's interesting that the 5+ activation rate for mobile web > desktop now -- just a few months ago, mobile was about half of desktop.
It could be the case that mobile web editors make a smaller number of small edits/spread a big edit out into smaller ones -- would be interesting at some point to see whether these initial edits happen in rapid succession. It'll also be interesting to compare the (short-term) retention rates of mobile vs. desktop.
The Activation rate for apps is weird. We know on desktop that about 70% of users create an account, but never edit -- this would imply that on mobile apps a far higher percentage of users create accounts for reasons unrelated to editing. We should look a little closer to the data . Something doesn't seem right.
Howie
On Thu, Jul 17, 2014 at 5:40 PM, Maryana Pinchuk mpinchuk@wikimedia.org wrote:
Greetings all, and happy 1 month anniversary of the tablet redirect :)
I've got some data to share on new tablet, phone, and app editors and how they stack up against new desktop editors. This is enwiki only for now; I await the glorious return of Dario to help us out with global data :)
*New active mobile editors*
Our metric for new active mobile editors is 5+ edits within 30 days of registration, so we don't quite have the full numbers on the first month of newly registered users yet (because people who signed up yesterday haven't had time to reach active status). But here's today's snapshot of the percent of newly registered users we've converted to active editors in the last 30 days across all out different platforms:
[image: Inline image 1] Note that the trend Dario observed still holds – both newly registered tablet and mobile site users are being converted to 5+ editors at a higher rate than desktop.
Slightly concerning is the conversion rate on the Android app – I'd have expected this to be higher, more in line with mobile site editors on phones. We'll need to figure out if this is due to users editing anonymously on the app instead of signing in and counting toward our new active user count.
*New user productivity*
I was also interested in the volume of contribution we're seeing across each of the platforms, to get a sense of how much new folks are contributing. So I ran the numbers for bytes changed (added, removed, and total changed) by newly registered users across all our different platforms:
[image: Inline image 2] Though mobile site editors (both tablet and phone) are reaching 5+ edits at a higher rate than desktop users, it looks like desktop editors make significantly bigger edits in terms of bytes. Not super surprising, given the smaller screen real-estate and greater difficulty in contributing longform text, but it's good to validate that with real data. Note again that Android app editors are on the lower end of bytes contributed.
To me, these two graphs raise the question of whether mobile site editors (on both tablets & phones) make a larger number of small edits as opposed to one big edit like they might on desktop. I don't know if we'll be able to answer that definitively anytime soon, but it's definitely something to think about and dig into more in the future...
-- Maryana Pinchuk Product Manager, Wikimedia Foundation wikimediafoundation.org
On Jul 17, 2014, at 6:33 PM, Howie Fung hfung@wikimedia.org wrote:
Thanks for putting this together -- this is really interesting data. It's interesting that the 5+ activation rate for mobile web > desktop now -- just a few months ago, mobile was about half of desktop.
Yeah, the jump happened in May, and we still don't know why. My theory is lastmodified, but it's a bit of a stretch for 5+.
It could be the case that mobile web editors make a smaller number of small edits/spread a big edit out into smaller ones -- would be interesting at some point to see whether these initial edits happen in rapid succession.
I can play around with some visualizations for that :)
It'll also be interesting to compare the (short-term) retention rates of mobile vs. desktop.
The Activation rate for apps is weird. We know on desktop that about 70% of users create an account, but never edit -- this would imply that on mobile apps a far higher percentage of users create accounts for reasons unrelated to editing. We should look a little closer to the data . Something doesn't seem right.
If it's weird, it's at least consistently weird -- I saw the same thing (a ~1% 1+ conversion rate) in the early post-release data. This is coming from Wikimetrics, which is pretty reliable. But yeah, let's get an independent third party to verify...
In the meantime, some wild speculation: maybe people really are used to signing up on all apps as part of first time ux without necessarily getting anything out of it, more so than on other platforms.
Howie
On Thu, Jul 17, 2014 at 5:40 PM, Maryana Pinchuk mpinchuk@wikimedia.org wrote: Greetings all, and happy 1 month anniversary of the tablet redirect :)
I've got some data to share on new tablet, phone, and app editors and how they stack up against new desktop editors. This is enwiki only for now; I await the glorious return of Dario to help us out with global data :)
New active mobile editors
Our metric for new active mobile editors is 5+ edits within 30 days of registration, so we don't quite have the full numbers on the first month of newly registered users yet (because people who signed up yesterday haven't had time to reach active status). But here's today's snapshot of the percent of newly registered users we've converted to active editors in the last 30 days across all out different platforms:
<image (21).png> Note that the trend Dario observed still holds – both newly registered tablet and mobile site users are being converted to 5+ editors at a higher rate than desktop.
Slightly concerning is the conversion rate on the Android app – I'd have expected this to be higher, more in line with mobile site editors on phones. We'll need to figure out if this is due to users editing anonymously on the app instead of signing in and counting toward our new active user count.
New user productivity
I was also interested in the volume of contribution we're seeing across each of the platforms, to get a sense of how much new folks are contributing. So I ran the numbers for bytes changed (added, removed, and total changed) by newly registered users across all our different platforms:
<image (20).png> Though mobile site editors (both tablet and phone) are reaching 5+ edits at a higher rate than desktop users, it looks like desktop editors make significantly bigger edits in terms of bytes. Not super surprising, given the smaller screen real-estate and greater difficulty in contributing longform text, but it's good to validate that with real data. Note again that Android app editors are on the lower end of bytes contributed.
To me, these two graphs raise the question of whether mobile site editors (on both tablets & phones) make a larger number of small edits as opposed to one big edit like they might on desktop. I don't know if we'll be able to answer that definitively anytime soon, but it's definitely something to think about and dig into more in the future...
-- Maryana Pinchuk Product Manager, Wikimedia Foundation wikimediafoundation.org
-- _______________________ Howie Fung Director of Product Development Wikimedia Foundation
hfung@wikimedia.org
On Thu, Jul 17, 2014 at 7:05 PM, Maryana Pinchuk mpinchuk@wikimedia.org wrote:
On Jul 17, 2014, at 6:33 PM, Howie Fung hfung@wikimedia.org wrote:
Thanks for putting this together -- this is really interesting data. It's interesting that the 5+ activation rate for mobile web > desktop now -- just a few months ago, mobile was about half of desktop.
Yeah, the jump happened in May, and we still don't know why. My theory is lastmodified, but it's a bit of a stretch for 5+.
Yeah, lastmodified is a stretch at 5+. . .I can't think of anything else. Let's keep an eye on this.
It could be the case that mobile web editors make a smaller number of small edits/spread a big edit out into smaller ones -- would be interesting at some point to see whether these initial edits happen in rapid succession.
I can play around with some visualizations for that :)
Cool -- look forward to seeing the results. A metric that comes to mind is velocity (e.g., distributions of time-to-5th edit). Lots of ways you can look at this.
It'll also be interesting to compare the (short-term) retention rates of mobile vs. desktop.
The Activation rate for apps is weird. We know on desktop that about 70% of users create an account, but never edit -- this would imply that on mobile apps a far higher percentage of users create accounts for reasons unrelated to editing. We should look a little closer to the data . Something doesn't seem right.
If it's weird, it's at least consistently weird -- I saw the same thing (a ~1% 1+ conversion rate) in the early post-release data. This is coming from Wikimetrics, which is pretty reliable. But yeah, let's get an independent third party to verify...
In the meantime, some wild speculation: maybe people really are used to signing up on all apps as part of first time ux without necessarily getting anything out of it, more so than on other platforms.
If that's the case, we should be able to see if in the data, right? If we're tagging account creations as first time ux vs. left nav, we should see tons of account creations coming in from first time ux.
Howie
On Thu, Jul 17, 2014 at 5:40 PM, Maryana Pinchuk mpinchuk@wikimedia.org wrote:
Greetings all, and happy 1 month anniversary of the tablet redirect :)
I've got some data to share on new tablet, phone, and app editors and how they stack up against new desktop editors. This is enwiki only for now; I await the glorious return of Dario to help us out with global data :)
*New active mobile editors*
Our metric for new active mobile editors is 5+ edits within 30 days of registration, so we don't quite have the full numbers on the first month of newly registered users yet (because people who signed up yesterday haven't had time to reach active status). But here's today's snapshot of the percent of newly registered users we've converted to active editors in the last 30 days across all out different platforms:
<image (21).png>
Note that the trend Dario observed still holds – both newly registered tablet and mobile site users are being converted to 5+ editors at a higher rate than desktop.
Slightly concerning is the conversion rate on the Android app – I'd have expected this to be higher, more in line with mobile site editors on phones. We'll need to figure out if this is due to users editing anonymously on the app instead of signing in and counting toward our new active user count.
*New user productivity*
I was also interested in the volume of contribution we're seeing across each of the platforms, to get a sense of how much new folks are contributing. So I ran the numbers for bytes changed (added, removed, and total changed) by newly registered users across all our different platforms:
<image (20).png> Though mobile site editors (both tablet and phone) are reaching 5+ edits at a higher rate than desktop users, it looks like desktop editors make significantly bigger edits in terms of bytes. Not super surprising, given the smaller screen real-estate and greater difficulty in contributing longform text, but it's good to validate that with real data. Note again that Android app editors are on the lower end of bytes contributed.
To me, these two graphs raise the question of whether mobile site editors (on both tablets & phones) make a larger number of small edits as opposed to one big edit like they might on desktop. I don't know if we'll be able to answer that definitively anytime soon, but it's definitely something to think about and dig into more in the future...
-- Maryana Pinchuk Product Manager, Wikimedia Foundation wikimediafoundation.org
-- _______________________ Howie Fung Director of Product Development Wikimedia Foundation
hfung@wikimedia.org
The first time UX isn't in any of our production builds right now. In the present build, there's only one place you can create an account, and that's in the left menu.
The onboarding was just pushed to beta yesterday, and should go live to the production build in around two weeks. Once that happens, the account creation schema does contain a source field [1], so we'll know what action they took to land them at the screen (onboarding, left menu, edit workflow).
Dan
[1]: https://meta.wikimedia.org/wiki/Schema:MobileWikiAppCreateAccount
On 17 July 2014 19:23, Howie Fung hfung@wikimedia.org wrote:
On Thu, Jul 17, 2014 at 7:05 PM, Maryana Pinchuk mpinchuk@wikimedia.org wrote:
On Jul 17, 2014, at 6:33 PM, Howie Fung hfung@wikimedia.org wrote:
Thanks for putting this together -- this is really interesting data. It's interesting that the 5+ activation rate for mobile web > desktop now -- just a few months ago, mobile was about half of desktop.
Yeah, the jump happened in May, and we still don't know why. My theory is lastmodified, but it's a bit of a stretch for 5+.
Yeah, lastmodified is a stretch at 5+. . .I can't think of anything else. Let's keep an eye on this.
It could be the case that mobile web editors make a smaller number of small edits/spread a big edit out into smaller ones -- would be interesting at some point to see whether these initial edits happen in rapid succession.
I can play around with some visualizations for that :)
Cool -- look forward to seeing the results. A metric that comes to mind is velocity (e.g., distributions of time-to-5th edit). Lots of ways you can look at this.
It'll also be interesting to compare the (short-term) retention rates of mobile vs. desktop.
The Activation rate for apps is weird. We know on desktop that about 70% of users create an account, but never edit -- this would imply that on mobile apps a far higher percentage of users create accounts for reasons unrelated to editing. We should look a little closer to the data . Something doesn't seem right.
If it's weird, it's at least consistently weird -- I saw the same thing (a ~1% 1+ conversion rate) in the early post-release data. This is coming from Wikimetrics, which is pretty reliable. But yeah, let's get an independent third party to verify...
In the meantime, some wild speculation: maybe people really are used to signing up on all apps as part of first time ux without necessarily getting anything out of it, more so than on other platforms.
If that's the case, we should be able to see if in the data, right? If we're tagging account creations as first time ux vs. left nav, we should see tons of account creations coming in from first time ux.
Howie
On Thu, Jul 17, 2014 at 5:40 PM, Maryana Pinchuk mpinchuk@wikimedia.org wrote:
Greetings all, and happy 1 month anniversary of the tablet redirect :)
I've got some data to share on new tablet, phone, and app editors and how they stack up against new desktop editors. This is enwiki only for now; I await the glorious return of Dario to help us out with global data :)
*New active mobile editors*
Our metric for new active mobile editors is 5+ edits within 30 days of registration, so we don't quite have the full numbers on the first month of newly registered users yet (because people who signed up yesterday haven't had time to reach active status). But here's today's snapshot of the percent of newly registered users we've converted to active editors in the last 30 days across all out different platforms:
<image (21).png>
Note that the trend Dario observed still holds – both newly registered tablet and mobile site users are being converted to 5+ editors at a higher rate than desktop.
Slightly concerning is the conversion rate on the Android app – I'd have expected this to be higher, more in line with mobile site editors on phones. We'll need to figure out if this is due to users editing anonymously on the app instead of signing in and counting toward our new active user count.
*New user productivity*
I was also interested in the volume of contribution we're seeing across each of the platforms, to get a sense of how much new folks are contributing. So I ran the numbers for bytes changed (added, removed, and total changed) by newly registered users across all our different platforms:
<image (20).png> Though mobile site editors (both tablet and phone) are reaching 5+ edits at a higher rate than desktop users, it looks like desktop editors make significantly bigger edits in terms of bytes. Not super surprising, given the smaller screen real-estate and greater difficulty in contributing longform text, but it's good to validate that with real data. Note again that Android app editors are on the lower end of bytes contributed.
To me, these two graphs raise the question of whether mobile site editors (on both tablets & phones) make a larger number of small edits as opposed to one big edit like they might on desktop. I don't know if we'll be able to answer that definitively anytime soon, but it's definitely something to think about and dig into more in the future...
-- Maryana Pinchuk Product Manager, Wikimedia Foundation wikimediafoundation.org
-- _______________________ Howie Fung Director of Product Development Wikimedia Foundation
hfung@wikimedia.org
-- _______________________ Howie Fung Director of Product Development Wikimedia Foundation
hfung@wikimedia.org
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
