At the c-level and above, the idea of leadership is to either outsource things or personally nitpick everything. There is no vision for what tripadvisor should become and no feeling of building towards something cohesive. Despite this we still get hours of content-free all-hands type meetings where the high ups pat each other on the back.
The size of the company means that there are many non-good people working there, both in the engineering org and elsewhere. This manifests in the form of many engineers essentially just punching the clock and not willing to question obviously pointless projects which now come down from an excessively huge PM organization which acts as a game of telephone between the ceo and engineering.
In the team I was on, there was a drive to report "wins" on A/B experiments -- the implication being that this would affect bonus / promotion etc. This incentivized some of the worst statistical analysis that I have seen -- at this point the so-called analysts have basically adopted their own language for reporting results which justifies ignoring the p-values (e.g. , "revenue is directionally up with 80% confidence" as opposed to "revenue did not change significantly"). This combined with the overall shoddy state of logging at TA and the use of a pot-pourri of team-specific vanity metrics make it hard to believe any of the analysis. This kind of a pro and a con since you can evidently get ahead by adopting these techniques, you just cant take yourself seriously as a statistician anymore.
