Is there other work on how ads affect traffic? This seems pretty significant if it's reproducible and generalizable gwern.net/Ads
2
35
2
105
Replying to @danluu @gwern
Most of the work on this stuff is proprietary. I ran such an experiment for a large content site which generated directionally similar results. I helped another major content site set up a similar test, but they didn't tell me the results.

11:57 PM · Nov 12, 2017

1
0
1
3
You’re unable to view this Tweet because this account owner limits who can view their Tweets.
Nothing that I know of. Definitely no one used an experimental methodology as interesting as yours. If you're running a real content site (i.e. you have an Omniture license) you'd just randomize visitors with hash(visitor cookie + testname) % 2 == {0 or 1}. 1/
1
0
0
0