Hi all,
here is the weekly look at our most important readership metrics.
As laid out last week, the main purpose is to raise awareness about how these are developing, call out the impact of any unusual events in the preceding week, and facilitate thinking about core metrics in general.
We will iterate on the presentation (e.g. to better take seasonality into account, in particular including year-over-year comparisons) and eventually want to create dashboards for those which are not already available in that form already. Feedback and discussion welcome.
(All numbers below are averages for September 14-20, 2015 unless otherwise noted.)
Pageviews:
Total: 523 million/day
Context (April 2015-September 2015):
Pageviews were down 4.3% from last week, a somewhat noticeable drop (for a more dramatized view, see the Vital Signs dashboard https://vital-signs.wmflabs.org/#projects=all/metrics=Pageviews).
Desktop: 58.1%
Mobile web: 40.7%
Apps: 1.2%
Global North ratio: 77.4% of total pageviews
Context (April 2015-September 2015):
Unique app users
Android: 1.303 million /day
Context (January 2015-September 2015):
A slight rise in the last two weeks, but we’re not seeing a clear effect of the recent media promotion for the Android app (blog post https://blog.wikimedia.org/2015/09/10/wikipedia-app-new-navigation-features/ on September 10, media coverage https://meta.wikimedia.org/wiki/Communications_committee/Press_clippings/2015#September on September 10, 11 and 13) - compare the bump in April around the time when the app was featured in the Play store and the media promotion for the iOS app happened (with much wider media coverage https://meta.wikimedia.org/wiki/Communications_committee/Press_clippings/2015#April ).
iOS: 319k / day
Context (January 2015-September 2015):
No news here.
New app installations
Android: 44,066/day (installs per device, from Google Play, September 14-19)
Context (Aug 21, 2015 - Sep 21, 2015):
One is tempted to see a bit of a bump after the start of the media promotion on September 10 (concretely, average daily installs in the time from September 10-19 were 2.5% higher than in the time from August 21-September 9), but in the end we can’t really be certain whether it had an effect. (Compare again very visible peak in April in the year-over-year graph https://commons.wikimedia.org/wiki/File:Wikipedia_Android_app_-_daily_installs_by_device,_Sep_11,_2014_-_Sep_11,_2015_(Google_Play).png I posted last week.)
The drop on Sunday September 20 seems to be a problem with Google Play’s statistics (several other metrics like uninstall numbers show a similar drop) and not something we need to worry about; I left it out from the average above as an outlier.
iOS: 4,839/day (download numbers from App Annie)
Context (September 2014-September 2015):
----
For reference, the queries and source links used are listed below (access is needed for each). Most of the above charts have been updated on Commons, too https://commons.wikimedia.org/w/index.php?title=Special:ListFiles&offset=20150923010000&user=Tbayer+%28WMF%29&ilshowall=1&limit=4 .
hive (wmf)> SELECT SUM(view_count)/7000000 AS avg_daily_views_millions FROM wmf.projectview_hourly WHERE agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20";
hive (wmf)> SELECT year, month, day, CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, sum(IF(access_method <> 'desktop', view_count, null)) AS mobileviews, SUM(view_count) AS allviews FROM wmf.projectview_hourly WHERE year=2015 AND agent_type = 'user' GROUP BY year, month, day ORDER BY year, month, day LIMIT 1000;
hive (wmf)> SELECT access_method, SUM(view_count)/7 FROM wmf.projectview_hourly WHERE agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20" GROUP BY access_method;
hive (wmf)> SELECT SUM(IF (FIND_IN_SET(country_code, 'AD,AL,AT,AX,BA,BE,BG,CH,CY,CZ,DE,DK,EE,ES,FI,FO,FR,FX,GB,GG,GI,GL,GR,HR,HU,IE,IL,IM,IS,IT,JE,LI,LU,LV,MC,MD,ME,MK,MT,NL,NO,PL,PT,RO,RS,RU,SE,SI,SJ,SK,SM,TR,VA,AU,CA,HK,MO,NZ,JP,SG,KR,TW,US')
0, view_count, 0))/SUM(view_count) FROM wmf.projectview_hourly WHERE
agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20";
hive (wmf)> SELECT year, month, day, CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")), SUM(view_count) AS all, SUM(IF (FIND_IN_SET(country_code, 'AD,AL,AT,AX,BA,BE,BG,CH,CY,CZ,DE,DK,EE,ES,FI,FO,FR,FX,GB,GG,GI,GL,GR,HR,HU,IE,IL,IM,IS,IT,JE,LI,LU,LV,MC,MD,ME,MK,MT,NL,NO,PL,PT,RO,RS,RU,SE,SI,SJ,SK,SM,TR,VA,AU,CA,HK,MO,NZ,JP,SG,KR,TW,US')
0, view_count, 0)) AS Global_North_views FROM wmf.projectview_hourly
WHERE year = 2015 AND agent_type='user' GROUP BY year, month, day ORDER BY year, month, day LIMIT 1000;
hive (wmf)> SELECT SUM(IF(platform = 'Android',unique_count,0))/7 AS avg_Android_DAU_last_week, SUM(IF(platform = 'iOS',unique_count,0))/7 AS avg_iOS_DAU_last_week FROM wmf.mobile_apps_uniques_daily WHERE CONCAT(year,LPAD(month,2,"0"),LPAD(day,2,"0")) BETWEEN 20150914 AND 20150920;
hive (wmf)> SELECT CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, unique_count AS Android_DAU FROM wmf.mobile_apps_uniques_daily WHERE platform = 'Android';
hive (wmf)> SELECT CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, unique_count AS iOS_DAU FROM wmf.mobile_apps_uniques_daily WHERE platform = 'iOS';
https://play.google.com/apps/publish/?dev_acc=18062848292902787645#StatsPlac...
https://www.appannie.com/dashboard/252257/item/324715238/downloads/?breakdow...
+ Analytics
Tilman, this is great. I find myself enjoying the stats/graphs and analysis/commentary equally and so far I don't see anything that isn't very helpful. Two considerations:
- Two app-level metrics I would like to see moving forward would be a number showing a daily/monthly visit count divided by the install base (if it is achievable): 15% of Android users visit every day, 45% visit every month. Along those lines, I am very interested to see the retention data that Android is now collecting.
- Ultimately, I think 30 day views (or 28 day views) are less subject to noise than 7-day.
- On a related note, so far the weekly cadence is great. It might also be that after 3 weeks we realize that a biweekly or monthly cadence is all we really need.
At some point, we will have to solidify this so that we can automate some/most of it. Perhaps we should chat offline about what your tolerance is in that regard.
-J
On Wed, Sep 23, 2015 at 12:26 AM, Tilman Bayer tbayer@wikimedia.org wrote:
Hi all,
here is the weekly look at our most important readership metrics.
As laid out last week, the main purpose is to raise awareness about how these are developing, call out the impact of any unusual events in the preceding week, and facilitate thinking about core metrics in general.
We will iterate on the presentation (e.g. to better take seasonality into account, in particular including year-over-year comparisons) and eventually want to create dashboards for those which are not already available in that form already. Feedback and discussion welcome.
(All numbers below are averages for September 14-20, 2015 unless otherwise noted.)
Pageviews:
Total: 523 million/day
Context (April 2015-September 2015):
Pageviews were down 4.3% from last week, a somewhat noticeable drop (for a more dramatized view, see the Vital Signs dashboard https://vital-signs.wmflabs.org/#projects=all/metrics=Pageviews).
Desktop: 58.1%
Mobile web: 40.7%
Apps: 1.2%
Global North ratio: 77.4% of total pageviews
Context (April 2015-September 2015):
Unique app users
Android: 1.303 million /day
Context (January 2015-September 2015):
A slight rise in the last two weeks, but we’re not seeing a clear effect of the recent media promotion for the Android app (blog post https://blog.wikimedia.org/2015/09/10/wikipedia-app-new-navigation-features/ on September 10, media coverage https://meta.wikimedia.org/wiki/Communications_committee/Press_clippings/2015#September on September 10, 11 and 13) - compare the bump in April around the time when the app was featured in the Play store and the media promotion for the iOS app happened (with much wider media coverage https://meta.wikimedia.org/wiki/Communications_committee/Press_clippings/2015#April ).
iOS: 319k / day
Context (January 2015-September 2015):
No news here.
New app installations
Android: 44,066/day (installs per device, from Google Play, September 14-19)
Context (Aug 21, 2015 - Sep 21, 2015):
One is tempted to see a bit of a bump after the start of the media promotion on September 10 (concretely, average daily installs in the time from September 10-19 were 2.5% higher than in the time from August 21-September 9), but in the end we can’t really be certain whether it had an effect. (Compare again very visible peak in April in the year-over-year graph https://commons.wikimedia.org/wiki/File:Wikipedia_Android_app_-_daily_installs_by_device,_Sep_11,_2014_-_Sep_11,_2015_(Google_Play).png I posted last week.)
The drop on Sunday September 20 seems to be a problem with Google Play’s statistics (several other metrics like uninstall numbers show a similar drop) and not something we need to worry about; I left it out from the average above as an outlier.
iOS: 4,839/day (download numbers from App Annie)
Context (September 2014-September 2015):
For reference, the queries and source links used are listed below (access is needed for each). Most of the above charts have been updated on Commons, too https://commons.wikimedia.org/w/index.php?title=Special:ListFiles&offset=20150923010000&user=Tbayer+%28WMF%29&ilshowall=1&limit=4 .
hive (wmf)> SELECT SUM(view_count)/7000000 AS avg_daily_views_millions FROM wmf.projectview_hourly WHERE agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20";
hive (wmf)> SELECT year, month, day, CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, sum(IF(access_method <> 'desktop', view_count, null)) AS mobileviews, SUM(view_count) AS allviews FROM wmf.projectview_hourly WHERE year=2015 AND agent_type = 'user' GROUP BY year, month, day ORDER BY year, month, day LIMIT 1000;
hive (wmf)> SELECT access_method, SUM(view_count)/7 FROM wmf.projectview_hourly WHERE agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20" GROUP BY access_method;
hive (wmf)> SELECT SUM(IF (FIND_IN_SET(country_code, 'AD,AL,AT,AX,BA,BE,BG,CH,CY,CZ,DE,DK,EE,ES,FI,FO,FR,FX,GB,GG,GI,GL,GR,HR,HU,IE,IL,IM,IS,IT,JE,LI,LU,LV,MC,MD,ME,MK,MT,NL,NO,PL,PT,RO,RS,RU,SE,SI,SJ,SK,SM,TR,VA,AU,CA,HK,MO,NZ,JP,SG,KR,TW,US')
0, view_count, 0))/SUM(view_count) FROM wmf.projectview_hourly WHERE
agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20";
hive (wmf)> SELECT year, month, day, CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")), SUM(view_count) AS all, SUM(IF (FIND_IN_SET(country_code, 'AD,AL,AT,AX,BA,BE,BG,CH,CY,CZ,DE,DK,EE,ES,FI,FO,FR,FX,GB,GG,GI,GL,GR,HR,HU,IE,IL,IM,IS,IT,JE,LI,LU,LV,MC,MD,ME,MK,MT,NL,NO,PL,PT,RO,RS,RU,SE,SI,SJ,SK,SM,TR,VA,AU,CA,HK,MO,NZ,JP,SG,KR,TW,US')
0, view_count, 0)) AS Global_North_views FROM wmf.projectview_hourly
WHERE year = 2015 AND agent_type='user' GROUP BY year, month, day ORDER BY year, month, day LIMIT 1000;
hive (wmf)> SELECT SUM(IF(platform = 'Android',unique_count,0))/7 AS avg_Android_DAU_last_week, SUM(IF(platform = 'iOS',unique_count,0))/7 AS avg_iOS_DAU_last_week FROM wmf.mobile_apps_uniques_daily WHERE CONCAT(year,LPAD(month,2,"0"),LPAD(day,2,"0")) BETWEEN 20150914 AND 20150920;
hive (wmf)> SELECT CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, unique_count AS Android_DAU FROM wmf.mobile_apps_uniques_daily WHERE platform = 'Android';
hive (wmf)> SELECT CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, unique_count AS iOS_DAU FROM wmf.mobile_apps_uniques_daily WHERE platform = 'iOS';
https://play.google.com/apps/publish/?dev_acc=18062848292902787645#StatsPlac...
https://www.appannie.com/dashboard/252257/item/324715238/downloads/?breakdow...
-- Tilman Bayer Senior Analyst Wikimedia Foundation IRC (Freenode): HaeB
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
Couple comments: on Jon's comments:
- Two app-level metrics I would like to see moving forward would be a number showing a daily/monthly visit count divided by the install base (if it is achievable): 15% of Android users visit every day, 45% visit every month. Along those lines, I am very interested to see the retention data that Android is now collecting.
Apple now provides App Retention info on iTunes Connect Analytics page, so we can do this for both platforms. Let me know if I can help show you where that live in our Apple account.
- Ultimately, I think 30 day views (or 28 day views) are less subject to noise than 7-day.
Some kind of moving average may be helpful to reduce the noise. Maybe as a trend line laid on top of the raw data?
On Wed, Sep 23, 2015 at 3:24 PM, Jon Katz jkatz@wikimedia.org wrote:
- Analytics
Tilman, this is great. I find myself enjoying the stats/graphs and analysis/commentary equally and so far I don't see anything that isn't very helpful. Two considerations:
- Two app-level metrics I would like to see moving forward would be a
number showing a daily/monthly visit count divided by the install base (if it is achievable): 15% of Android users visit every day, 45% visit every month. Along those lines, I am very interested to see the retention data that Android is now collecting.
- Ultimately, I think 30 day views (or 28 day views) are less subject
to noise than 7-day.
- On a related note, so far the weekly cadence is great. It might
also be that after 3 weeks we realize that a biweekly or monthly cadence is all we really need.
At some point, we will have to solidify this so that we can automate some/most of it. Perhaps we should chat offline about what your tolerance is in that regard.
-J
On Wed, Sep 23, 2015 at 12:26 AM, Tilman Bayer tbayer@wikimedia.org wrote:
Hi all,
here is the weekly look at our most important readership metrics.
As laid out last week, the main purpose is to raise awareness about how these are developing, call out the impact of any unusual events in the preceding week, and facilitate thinking about core metrics in general.
We will iterate on the presentation (e.g. to better take seasonality into account, in particular including year-over-year comparisons) and eventually want to create dashboards for those which are not already available in that form already. Feedback and discussion welcome.
(All numbers below are averages for September 14-20, 2015 unless otherwise noted.)
Pageviews:
Total: 523 million/day
Context (April 2015-September 2015):
Pageviews were down 4.3% from last week, a somewhat noticeable drop (for a more dramatized view, see the Vital Signs dashboard https://vital-signs.wmflabs.org/#projects=all/metrics=Pageviews).
Desktop: 58.1%
Mobile web: 40.7%
Apps: 1.2%
Global North ratio: 77.4% of total pageviews
Context (April 2015-September 2015):
Unique app users
Android: 1.303 million /day
Context (January 2015-September 2015):
A slight rise in the last two weeks, but we’re not seeing a clear effect of the recent media promotion for the Android app (blog post https://blog.wikimedia.org/2015/09/10/wikipedia-app-new-navigation-features/ on September 10, media coverage https://meta.wikimedia.org/wiki/Communications_committee/Press_clippings/2015#September on September 10, 11 and 13) - compare the bump in April around the time when the app was featured in the Play store and the media promotion for the iOS app happened (with much wider media coverage https://meta.wikimedia.org/wiki/Communications_committee/Press_clippings/2015#April ).
iOS: 319k / day
Context (January 2015-September 2015):
No news here.
New app installations
Android: 44,066/day (installs per device, from Google Play, September 14-19)
Context (Aug 21, 2015 - Sep 21, 2015):
One is tempted to see a bit of a bump after the start of the media promotion on September 10 (concretely, average daily installs in the time from September 10-19 were 2.5% higher than in the time from August 21-September 9), but in the end we can’t really be certain whether it had an effect. (Compare again very visible peak in April in the year-over-year graph https://commons.wikimedia.org/wiki/File:Wikipedia_Android_app_-_daily_installs_by_device,_Sep_11,_2014_-_Sep_11,_2015_(Google_Play).png I posted last week.)
The drop on Sunday September 20 seems to be a problem with Google Play’s statistics (several other metrics like uninstall numbers show a similar drop) and not something we need to worry about; I left it out from the average above as an outlier.
iOS: 4,839/day (download numbers from App Annie)
Context (September 2014-September 2015):
For reference, the queries and source links used are listed below (access is needed for each). Most of the above charts have been updated on Commons, too https://commons.wikimedia.org/w/index.php?title=Special:ListFiles&offset=20150923010000&user=Tbayer+%28WMF%29&ilshowall=1&limit=4 .
hive (wmf)> SELECT SUM(view_count)/7000000 AS avg_daily_views_millions FROM wmf.projectview_hourly WHERE agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20";
hive (wmf)> SELECT year, month, day, CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, sum(IF(access_method <> 'desktop', view_count, null)) AS mobileviews, SUM(view_count) AS allviews FROM wmf.projectview_hourly WHERE year=2015 AND agent_type = 'user' GROUP BY year, month, day ORDER BY year, month, day LIMIT 1000;
hive (wmf)> SELECT access_method, SUM(view_count)/7 FROM wmf.projectview_hourly WHERE agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20" GROUP BY access_method;
hive (wmf)> SELECT SUM(IF (FIND_IN_SET(country_code, 'AD,AL,AT,AX,BA,BE,BG,CH,CY,CZ,DE,DK,EE,ES,FI,FO,FR,FX,GB,GG,GI,GL,GR,HR,HU,IE,IL,IM,IS,IT,JE,LI,LU,LV,MC,MD,ME,MK,MT,NL,NO,PL,PT,RO,RS,RU,SE,SI,SJ,SK,SM,TR,VA,AU,CA,HK,MO,NZ,JP,SG,KR,TW,US')
0, view_count, 0))/SUM(view_count) FROM wmf.projectview_hourly WHERE
agent_type = 'user' AND CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) BETWEEN "2015-09-14" AND "2015-09-20";
hive (wmf)> SELECT year, month, day, CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")), SUM(view_count) AS all, SUM(IF (FIND_IN_SET(country_code, 'AD,AL,AT,AX,BA,BE,BG,CH,CY,CZ,DE,DK,EE,ES,FI,FO,FR,FX,GB,GG,GI,GL,GR,HR,HU,IE,IL,IM,IS,IT,JE,LI,LU,LV,MC,MD,ME,MK,MT,NL,NO,PL,PT,RO,RS,RU,SE,SI,SJ,SK,SM,TR,VA,AU,CA,HK,MO,NZ,JP,SG,KR,TW,US')
0, view_count, 0)) AS Global_North_views FROM wmf.projectview_hourly
WHERE year = 2015 AND agent_type='user' GROUP BY year, month, day ORDER BY year, month, day LIMIT 1000;
hive (wmf)> SELECT SUM(IF(platform = 'Android',unique_count,0))/7 AS avg_Android_DAU_last_week, SUM(IF(platform = 'iOS',unique_count,0))/7 AS avg_iOS_DAU_last_week FROM wmf.mobile_apps_uniques_daily WHERE CONCAT(year,LPAD(month,2,"0"),LPAD(day,2,"0")) BETWEEN 20150914 AND 20150920;
hive (wmf)> SELECT CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, unique_count AS Android_DAU FROM wmf.mobile_apps_uniques_daily WHERE platform = 'Android';
hive (wmf)> SELECT CONCAT(year,"-",LPAD(month,2,"0"),"-",LPAD(day,2,"0")) as date, unique_count AS iOS_DAU FROM wmf.mobile_apps_uniques_daily WHERE platform = 'iOS';
https://play.google.com/apps/publish/?dev_acc=18062848292902787645#StatsPlac...
https://www.appannie.com/dashboard/252257/item/324715238/downloads/?breakdow...
-- Tilman Bayer Senior Analyst Wikimedia Foundation IRC (Freenode): HaeB
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
On Wed, Sep 23, 2015 at 4:21 PM, Joshua Minor jminor@wikimedia.org wrote:
Couple comments: on Jon's comments:
- Two app-level metrics I would like to see moving forward would be a
number showing a daily/monthly visit count divided by the install base (if it is achievable): 15% of Android users visit every day, 45% visit every month. Along those lines, I am very interested to see the retention data that Android is now collecting.
Apple now provides App Retention info on iTunes Connect Analytics page, so we can do this for both platforms. Let me know if I can help show you where that live in our Apple account.
As an experiment, I added 1-day retention to today's report after discussion with Josh. Still need to look into the Android data.
- Ultimately, I think 30 day views (or 28 day views) are less subject
to noise than 7-day.
Some kind of moving average may be helpful to reduce the noise. Maybe as a trend line laid on top of the raw data?
Yes, trendlines or generally a better handle on seasonality would be good. For some metrics like the app uniques where we don't yet have a full year of data, this might be trickier.
On Wed, Sep 23, 2015 at 3:24 PM, Jon Katz jkatz@wikimedia.org wrote:
- Analytics
Tilman, this is great. I find myself enjoying the stats/graphs and analysis/commentary equally and so far I don't see anything that isn't very helpful. Two considerations:
...
- On a related note, so far the weekly cadence is great. It might
also be that after 3 weeks we realize that a biweekly or monthly cadence is all we really need.
Yes, that's been on my mind too - we should loop you into existing
conversations about the cadence. My thinking was that to keep such a report worthwhile reading, it should offer some interpretation and context, and I also think there is value in regularly nudging us to examine the data for the impact of unusual events or regressions.
At some point, we will have to solidify this so that we can automate some/most of it. Perhaps we should chat offline about what your tolerance is in that regard.
Indeed, as noted in the very first announcement and the earlier Phabricator task, dashboarding most of these is the goal. Looking into options currently. Then again we wanted to keep it flexible at first, and open to feedback - no use building a dashboard for metrics that people tell us they don't find useful. Fortunately, based on the reactions so far, that doesn't seem to be the case.