Reports! Every organization believes their reporting requirements are unique. They are willing to pay a premium to get their reports built out either in your product or in a data warehouse / business intelligence project.
I resisted handing over raw data because I felt that this way it would force us to build the right reports and visualization package in our product. I was worried about training about the constant data model changes to our customers would slow us down.
Data and insights from your product make the adoption easier. Either it’s data that they didn’t have access to (including speed, frequency of refresh, accessibility etc) or it’s new data.
However, getting the reports right on Day 1 is very hard. Many requirements are overdone or underdone. In any case it’s a lot of effort wasted.
Exposing the internal data is the best way forward. In hindsight I would spend a disproportionate time to train the BI/Analytics/MIS teams at the client on our data model and data sources and the common queries. I would have given them Jupyter Notebooks and SQL queries.
You can charge a premium for training. You win over the skeptics or the analytics teams that’s always trying to prove it’s worth. If you win this team, they will force all other teams to adopt the product. Because they want the data.
They will also be your first responders when data quality suffers. You can very quickly prototype their requirements and wishlist at their effort. Once they narrow in on the reports, you can “productionize” it.
To be honest, Enterprise SaaS products tend to stick until they are forced out of the system. It takes time to get sticky and the price doesn’t guarantee stickiness. But what does, Data! It make stickiness move forward or help you latch on even when clearly you are well past your prime time.