At startups, we all brag about being data-driven, based our decisions solely in the data, and making no assumptions or as little assumptions as possible about our users. To succeed in doing this you need to have a data platform that allows you to answer the questions you come up with.

Every data platform is made of a few parts:

• Getting the data
• Processing the data
• Visualizing the data

Getting the data

Users use the app, and they create lots of entries in our database in doing so. This is the more straight forward data we can get, but that is probably not enough. Some of the actions the user perform don’t need to store anything in the DB. Getting stuff like when a user visited a screen, or what are the funnels that led to conversions is very important for a startup. It helps to identify friction point and to prioritise the work on them.

I’d probably suggest using a third party for this, tools like Heap, Mixpanel or Amplitude will help you implementing all sort of events.

Processing the data

Processing the data can be very tricky. I’m a huge fan of Redash. This little tool can be installed from a Bitnami image in AWS and GCP (and probably in other cloud providers as well). The Bitnami image works decently for small organizations and it allows us to connect to the different databases we could have. It supports lots of different databases, although some of them don’t really work that well (Kibana, for instance, gave me some headaches in the past).

If your DB is SQL you can start writing SQL queries and plotting some charts straight away. At Wager, I created a dashboard with metrics like the number of users signed up, conversions, etc. almost when starting, and I literally check it every day.

A few mind-blowing features Redash have are:

• Ability to run queries against many types of DBs
• Ability con combine the results of different queries, potentially from different sources
• Ability to run python code as part of the query
• Different charts options
• Export the data tables to CVS and Excel
• Ability to create a dashboard with different charting in minutes

More processing of the data

Sometimes we need to do things that are a bit more complex. Another incredibly helpful tool is Jupyter.