What sources/referrers are bringing people to the app?
What download channels generate the most monthly active users? The most daily active users?
For list of all questions see: https://docs.google.com/document/d/1jlgPYsMYSGKMvAc85kjtzWWibZCyHF9ycQP2zDTjJys/edit?ts=5a37eadd
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | SNowick_WMF | T184098 [EPIC] Analytics baseline for Android app | |||
| Resolved | SNowick_WMF | T184094 What are the most productive referrers/channels for Android? | |||
| Resolved | Dbrant | T188146 Android app acquisition channel is not logged in MobileWikiAppInstallReferrer |
What sources/referrers are bringing people to the app?
Already answered as part of T184089: https://github.com/wikimedia-research/App-Android-Baseline_Metrics/tree/master/T184089#retained-installs
What download channels generate the most monthly active users? The most daily active users?
Note to future self: will need to join MobileWikiAppInstallReferrer events with wmf.webrequest data in a way that makes it easier to do T184095 also.
Here's some data until a dedicated analyst can pick this up:
| month | utm_source | utm_medium | n_installs | prop_installs |
|---|---|---|---|---|
| 2019-08 | google-play | organic | 525527 | 98.4% |
| 2019-08 | organic | 4317 | 0.8% | |
| 2019-08 | (not set) | (not set) | 3279 | 0.6% |
| 2019-08 | wmf | centralnotice | 507 | 0.1% |
| 2019-08 | portal | button | 275 | 0.1% |
| 2019-08 | (other) | (other) | 41 | 0.0% |
| 2019-08 | (other) | 14 | 0.0% | |
| 2019-08 | (other) | organic | 8 | 0.0% |
| 2019-08 | portal | (other) | 1 | 0.0% |
| 2019-09 | google-play | organic | 580746 | 98.0% |
| 2019-09 | organic | 4867 | 0.8% | |
| 2019-09 | (not set) | (not set) | 3952 | 0.7% |
| 2019-09 | portal | button | 2759 | 0.5% |
| 2019-09 | (other) | (other) | 93 | 0.0% |
| 2019-09 | wmf | centralnotice | 77 | 0.0% |
| 2019-09 | (other) | 14 | 0.0% | |
| 2019-09 | portal | (other) | 11 | 0.0% |
| 2019-09 | (other) | organic | 4 | 0.0% |
| 2019-10 (partial) | google-play | organic | 276795 | 98.1% |
| 2019-10 (partial) | organic | 2292 | 0.8% | |
| 2019-10 (partial) | (not set) | (not set) | 1631 | 0.6% |
| 2019-10 (partial) | portal | button | 1374 | 0.5% |
| 2019-10 (partial) | (other) | (other) | 38 | 0.0% |
| 2019-10 (partial) | (other) | 6 | 0.0% | |
| 2019-10 (partial) | wmf | centralnotice | 4 | 0.0% |
| 2019-10 (partial) | portal | (other) | 4 | 0.0% |
| 2019-10 (partial) | (other) | organic | 2 | 0.0% |
Here's the query (for the future apps-focused analyst we're currently hiring for):
SELECT CASE month WHEN 8 THEN '2019-08' WHEN 9 THEN '2019-09' WHEN 10 THEN '2019-10 (partial)' END AS month, IF(event.utm_source IN('google-play', 'google', '(not set)', 'portal', 'wmf'), event.utm_source, '(other)') AS utm_source, IF(event.utm_medium IN('organic', '(not set)', 'centralnotice', 'button'), event.utm_medium, '(other)') AS utm_medium, COUNT(DISTINCT event.app_install_id) AS n_installs FROM event.mobilewikiappinstallreferrer WHERE year = 2019 AND month IN(8, 9, 10) AND event.utm_source IS NOT NULL GROUP BY CASE month WHEN 8 THEN '2019-08' WHEN 9 THEN '2019-09' WHEN 10 THEN '2019-10 (partial)' END, IF(event.utm_source IN('google-play', 'google', '(not set)', 'portal', 'wmf'), event.utm_source, '(other)'), IF(event.utm_medium IN('organic', '(not set)', 'centralnotice', 'button'), event.utm_medium, '(other)') ORDER BY month, n_installs DESC LIMIT 1000;