[quantified]

Improv's model was poor though, it was based on cell positions and it was easy to double-count. Basically was bad at various semantic aspects. I worked on a small conpetitor back in the day and we competed on modelling, there were a number of others too.

I've also worked on Anaplan and and its modeling is also much better, so please don't lower it to any version of Improv! It isn't, or at least wasn't a little while back, good at cross-metric ("line item" to them") calculations and presentations, but still easier to get something complex correct.

The real mistake of Improv was that it wasn't 1-2-3 so Lotus didn't know how to narket it or sell it.