How fast were these markets growing? Were they growing at all? Is it pretty stable, and it's pricing and maybe competitive factors—how you got or lost market share?

In an operator model, you're looking at long-term growth. It's about protecting your maintenance revenue; it's slow and steady growth. Anything that was spiky or going after markets where it looked like there was a lot of potential but then suddenly had a lot of competitors—those would be avoided. It would be slow growth, tight control over the OPEX on that growth, and very smart around the pricing strategies and attrition. Those are the areas where they want to move. These operators aren't looking for markets they can rush into; they rely on other businesses who are able to prove that worth, prove where they're sitting, where they're competing, and prove sustainable growth. Then they will go in and buy it. They're not looking to try and outrun other competitors by bringing in more products. More products are risky; new products are risky; R&D for new things is risky and costly. It's about what's there, what's repeatable, and how do we go about that. Then they acquire companies that have proven a product or something sustainable and leave those to run autonomously.

You don't have to give an exact figure, but I'm curious about retention. What's the average length that a customer stays with this sort of business, and what would be the two reasons to change? Is it that every once in a while a contract comes up for bidding? How do you think about that?

Where we do see things being dislodged—I can talk to the NHS and some of the EPRs there—you've got the old systems like the Lorenzos of the world. Where those just didn't evolve, there was a process that could be in there sometimes for 20 years, and then there's a complete rip and replace. In higher education, we see a lot of that happening. You have legacy systems like Banner that have been in there for 25 years but just haven't modernized, and there you see universities looking for a rip and replace. Where it's modern architecture, modern software, modern solutions, where they've slowly evolved over time and kept up to pace, those systems can stay in there for 10, 15, 20 years; it's very sticky.