A note on gathering less data

Yesterday, I said that you can lower costs by decreasing the amount of data you gather.

A note on that, you should not compromise on what interactions you gather, but where.

To restrict the data intake, restrict the pages that actually interaction data.

You should log that a customer visited a page, just skip logging interactions on that page.

The reason is consistency, you can remove unimportant pages for a while from your analysis and report that to the organization.

But going from few to many interactions on the same page means missing out on what customers were doing on that page before.

You can’t compare its current state to an earlier state, which means you can’t tell if any changes to that page has had an effect.

Also, when you make an analysis of historical data, you’ll have to remember what interactions where gathered at that point.

It’s easier to reason with some pages having their interactions not logged, you will not see any data for those pages on the dates they were restricted.

Decide on what pages you think are important and log everything there.

To help you decide what pages are important, figure out what customer journeys they take that you value.

– Samy