Summary
The best way to prove that your database is capable and fast is to have users try it. Yet, when we looked at our data, a large percentage of new users would lose interest before they even loaded data into a database. By developing clear tutorials to guide users, we were able to improve data loading by 14% and also increased the ratio of users loading their own data vs sample data.
Design Details
A landing page to direct users
After users signed up, we originally did not have any call to action. To remediate that, we created a home page where we could recommend next steps based on their interests. For existing users, we could use this space to announce new features and upcoming product releases.
Pre-packaged, personalized tutorials
Notebook tutorials allowed us to internally create and crowdsource personalized demos for users. Using jupyter notebooks allowed us to explain the recommended dataset and visualize the sample queries.
Improved load data for personal files
Previously loading data was an API only action and setting up the connection would cause our users to lose interest. By providing a wizard for quick file uploads, we retained users within the platform and could direct them to test queries after the files loaded.
