How would you know if website traffic has been reduced after adding a feature? what would you do?

Question

Zelman · Accepted Answer

I would be careful to answer it without investigating the core issue. What was the purpose of the feature (what were the goals), in what regard the traffic was reduced (regional, global, what about existing users, who did they react) what does traffic mean here? for e.g. if the new feature is SMS notification instead of an email, then perhaps the user is satisfied or has no way to get to the site from SMS. It maybe good or bad based on the objectives. If it is actually meeting the objectives of the feature but has a side effect as a whole, then it is a different solution altogether. Get the problem domain nailed before solution.

Satish · Answer

1) Scope the problem -> What is traffic? New users, unique users, MAU, retention? Can we tell how much?
2) Even before you roll out something -? Do A/B testing to ensure this is statistically significant
3) Make sure there are no seasonal effects or external elements such as network issues
4) Make sure there are no other substitutes that just became available in the market at the same time
5) If you can statistically conclude that there's indeed a correlation, take corrective action
6) Corrective action can be to roll back the feature or make modifications with A/B testing

Kobi · Answer

I would make sure that along with this feature, no other features are added. Then I would sample this feature along some kind of limited audience that should represent the total audience (call it B group). I would monitor traffic in comparison between A group and B group.

Google

Pergunta de entrevista da empresa Google

Respostas da entrevista

Empresas seguidas

Buscas de vagas