3 ms·
> So now there are 4 independent quasi-/experimental datapoints (me, Pandora, LinkedIn, & Mozilla resp) showing that a standard load of ads vs no ads decreases
by everdev 8y ago
> So now there are 4 independent quasi-/experimental datapoints (me, Pandora, LinkedIn, & Mozilla resp) showing that a standard load of ads vs no ads decreases total consumption/activity by 10–15%.
From the tweet itself, this is 10% of activity, not users. Kind of suspect that someone misinterpreted their own data so significantly.
- gwern 8y agoI disagree about it being a misinterpretation. They are measuring different things because they have to, but if you read the papers, you see that that's the net effect and most reasonable way to summarize them: an 11% reduction in intention to visit, a 9% reduction in total traffic, a similar reduction in total web browsing per individual, and in the case of LinkedIn, you can see both a decrease in users and activity per user, which net to the >10% reduction in total activity. This is as it must be (what is a web browser user going to do - stop using web browsers completely and quit the Internet entirely...? Not likely! Reduction in web usage is their only possible response.) How else could one reasonably summarize that in a few words?
- everdev 8y agoIn analytics there are ways to discuss activity like: visits per month, pages per session, time on site, conversion rate, etc. It's possible for any of those activity metrics to go down 10% without losing any active users.