If you've sat through a vendor demo in the last few years, you've heard it: "Our platform achieves a 94% resolution rate." Or 89%. Or 87. The number changes, but the emphasis stays consistent. Resolution rate gets top billing. It's the headline metric, the thing that goes on the slide with the biggest font, the number sales teams lead with.
There's a reason for that, and it's worth understanding because it says something genuine about how these vendors think about their products and, more importantly, how they think about you evaluating them.
Resolution rate measures the percentage of issues that get resolved without escalating to a human. Sounds straightforward. A user submits a problem, the system handles it, done. Higher number equals better product. Makes sense on the surface, which is exactly why vendors push it so hard.
The thing is, that metric is almost comically easy to manipulate. You can nudge the definition of "resolved" and suddenly your numbers improve. Did the system provide any response? Resolution. Did it provide a response the user accepted, even if it didn't actually fix anything? Resolution. Did it suggest something the user tried, regardless of whether it worked? You're probably calling that a resolution too.
More sophisticated games exist too. You can train your system exclusively on the easiest 20% of incoming issues because those are the ones where you'll get the highest success rate. You can weight your test data toward problems you know you handle well. You can exclude entire categories of complex issues from your resolution rate calculation by filing them under "out of scope." The vendor landscape is creative with this stuff.
Contrast that with metrics that are harder to fake. Mean time to resolution for actual resolved issues, where "resolved" means the user is back to normal productivity? Harder to game. First contact resolution rate, but only counting cases where you have genuine follow-up confirmation? Much harder. Customer satisfaction with the resolution? Basically impossible to manipulate without it becoming obvious. Those metrics are real, but they require actual competence. They require your system to genuinely solve problems.
So why do vendors lean so heavily on resolution rate specifically? Because it works in sales conversations. It's a number that sounds impressive. It's easy to compare across vendors (even though you're not actually comparing the same thing). And most importantly, it's something a vendor can claim to optimize without necessarily making their product better at what users actually care about.
There's a version of this in almost any market where measurement is hard. The metrics that get promoted are usually the ones that are simultaneously important-sounding and easy to manipulate. They become the metric of choice not because they're the best measure of value, but because they're the easiest to present confidently.
The vendors aren't doing anything illegal. They're not lying, technically. They're just running the game on the field that's been marked out for them. If resolution rate is what everyone's comparing on, then resolution rate is what you optimize. You'd almost be foolish not to.
The real insight here isn't that vendors are playing games with metrics. It's that if you're evaluating a tool and the vendor leads with a single number, especially one that's easy to define loosely, you're probably not looking at the most meaningful information. The questions worth asking are the ones that are harder to answer in a sales call: What do your users do when they get a resolution? Do they come back? What percentage actually solve their problem on the first interaction? How many escalations happen that look like resolutions in your report?
Those conversations are messier. The numbers are harder to produce. But they're the ones where you actually learn something about whether a tool will work for you.