I disagree. Reports and dashboards provide decision makers with information that they require to make decisions. This is always step one.
Randomly exploring data for "insights" is why so many companies are turning against "Data Science"; it rarely bares fruit. This work should be focused on dialing in business processes.
Agreed. A lot of my background is in banking / fintech. So much scrambling on "adhoc analytics" can be avoided if you have really well constructed, standard reports and dashboards.
The department I started my career at didn't have this and every exec request was a bespoke, adhoc request, tailored to answer that one question. I spent many late nights writing SQL and building excel files / powerpoints to answer a single question.
When I started at a fintech as the first data person, we built a few primary dashboards where you could drill down for detail. Those answered over 80% of the questions people asked.
I don't think we disagree (Report and dashboards are useful decision making tools), but in practice hiring external people to come in and build them causes more problems then it solves. In my experience processes on the ground can and will change and all of these pre-written reports and dashboards quickly become ossified and outdated.
If you take steps to make sure the data is properly stored and queryable stakeholders can maintain and develop their own reporting. Make the data understandable and accessible and reporting will follow. Human friendly schemas and such help a lot.
Randomly exploring data for "insights" is why so many companies are turning against "Data Science"; it rarely bares fruit. This work should be focused on dialing in business processes.