3 ms·

Have you controlled for all factors? This seems like circumstantial evidence at best; just because facebook customers are new doesn't imply that the improvement

• the_watcher 13y ago
  You are right, we could definitely do more work on controlling all factors here. However, the numbers are so staggering (80% of a not insubstantial group of customers - something we haven't seen in any of our other channels), coupled with the anecdotal evidence we've gathered from surveys and monitoring the comments we get on Facebook (I know, not data, but in the absence of data, I make do) that it seems likely. This is also something that AdRoll has seen - they only see ~19% overlap between people who click through their general display and FBX ads (I haven't seen the study, but they imply that they did some extensive research here).