13 karma · joined January 14, 2015
I have only 48gb of ram, so can fit only 80k context max, so good compaction is must.
We know that cases are one of the most important things in this game, but the problem is that the process is very slow (like starting a campaign could take up to 2 months), so this makes everything a little bit complicated.
Even if we received very positive feedback on product / idea and plans on integration - the process of generating hypothesis and creating different variations was pretty time-consuming and in the end people did not complete it.
Posting to StackThatMoney definitely sounds like something worth trying, and we’ll give it a shot. Thank you!
Maybe we did it all wrong - we tried to build a slick product, collect leads, carefully onboard them, nurture with content, optimize our funnels and so on. Perhaps we should just concentrate on collecting phone numbers and selling directly, but this is the thing that we tried to avoid all the time.
Usually it requires about 10k unique visitors per month to start with. Concrete answer on your question hugely depends on your audience structure and number of variations.
2) The idea here is that with ML you should not analyze every dimension separately. ML is taking into the account all available characteristics and making decisions based on all of them together (like if the guy on OS X, who came from NY from the Facebook campaign in the evening - prefer to watch product video instead of watching screenshots - no problem, we'll show him video).
3) The real power of ML comes out when you could not obviously split your traffic based on the ad link (like utm_campaign=dogs). Direct and search traffic on your homepage are great examples in this case. Also, manual targeting requires a bunch of analytic folks, who will continuously analyze your traffic, setup and adjust optimization campaigns. Even in this case - it's still difficult to adapt to dynamic changes in traffic (like a new type of visitors, season changes, etc). So ML could not only improve results but also decrease the amount of human resources which is currently required for solving such complex problems.
So what I'm trying to say is that your assumptions definitely make sense in some cases. But we believe that there are still plenty of cases when ML could drastically increase your results and save your time.
Regarding the question about multivariate - we're currently support only A/B personalization (create few variations and we will show the most convertible one) and Split URL personalization (when we're splitting traffic between multiple landing page URLs)