You can't analyse what you've never captured.
A configurator is generating data the whole time. The only question is: did you build it in a way that lets you actually use it?
Good reporting on your CPQ data goes further than just seeing how much customisation gets sold. It also shows what was offered to customers within a given period — and that's a different figure from what was ultimately sold.
That difference is exactly where the valuable information sits. A particular piece of customisation might turn out to have been decisive in winning an order. Another piece gets offered routinely but is never chosen — time and effort that delivers little. Without laying that data side by side, you don't see the difference.
The same applies to trends over time: which model or item gets offered a lot in one period, and barely at all in another? Patterns like that say something about where the market is heading — often before you'd see it in revenue figures alone.
"This kind of insight doesn't appear on its own. It requires thinking about it while the configurator is being built — not the first time you try to produce a report."
It requires the right data to actually be there, in the right form, and over a longer period than you might expect beforehand. Miss that, and you'll only notice the first time you try to report — when the data turns out not to be there, or not to be usable.
In your organisation, is the data you'll later need for reporting already considered while a configurator is being built, or does that question only surface afterwards?