Ticket deflection measures how many customers never reached a human agent. Resolution measures how many customers actually got their problem solved. They sound similar, but they pull your support team in opposite directions. Deflection rewards keeping people out. Resolution rewards getting people answers. If you only track the first one, you can post great numbers while your customers quietly suffer.
This post breaks down what deflection really counts, how it became the default metric, why it can hide a bad experience, and why resolution is the healthier thing to measure and pay for.
What ticket deflection actually means
Deflection is the share of contacts that get handled before they become a human ticket. A customer opens a chat, reads a help article, or talks to a bot, and then doesn’t escalate to a person. That counts as deflected.
The logic is simple. Every ticket a human doesn’t touch is a ticket you don’t pay an agent to work. For a team drowning in volume, that math is appealing. Fewer human tickets means lower cost per contact, shorter queues, and a tidy chart for the next board meeting.
That’s why deflection took off. It maps cleanly to cost. It’s easy to calculate. And for years, the tools available, mostly searchable knowledge bases and basic chatbots, were good at one thing: stopping a customer before they reached an agent. The metric matched the technology.
The problem with measuring an absence
Here’s the catch. Deflection measures something that didn’t happen. The customer didn’t escalate. But “didn’t escalate” covers a lot of ground, and not all of it is good.
A customer counts as deflected when they:
- Got a clear answer and left happy. (Good.)
- Got a vague answer, gave up, and stewed. (Bad.)
- Couldn’t find the “talk to a human” button. (Worse.)
- Closed the tab in frustration and churned a week later. (Worst.)
All four look identical in a deflection report. The metric can’t tell a solved problem from an abandoned one. So a tool can drive deflection up by making the human handoff hard to find, padding bot replies, or burying the contact form. The number climbs. The experience sinks.
This is the quiet danger. Deflection can be gamed without anyone deciding to game it. The incentive does the work. When a team is told to raise deflection, the cheapest path is often to make leaving harder, not to make answers better.
Why resolution is the healthier metric
A resolution is different. It’s a customer who came with a question and left with an answer they could use. It measures presence, not absence. Something good happened, and you can point to it.
That distinction changes behavior across the board:
- For your team, the goal becomes “did we actually help,” not “did we avoid a ticket.” Those are not the same job.
- For your customers, the path to a human stays open. A handoff isn’t a failure. It’s part of a good system. Some questions need a person, and pretending otherwise just frustrates people.
- For your tooling, the pressure points the right way. Better answers, grounded in your real help docs, beat clever ways to make people stop asking.
This is the difference between resolving and deflecting. One solves the problem. The other moves it out of view. We built Fidiora around resolving, because deflection metrics quietly reward the wrong outcome. You can read how we define a resolution and what we deliberately don’t count.
How to tell the two apart in your own data
You don’t need a new platform to start auditing this. A few questions cut through fast:
1. What happens after a “deflected” chat?
Look at customers who didn’t escalate. Do they come back within a day or two with the same issue? Repeat contacts are deflection’s tell. A real resolution rarely needs a sequel.
2. How hard is it to reach a human?
Open your own chat as a customer. Count the steps to a person. If it takes effort to find, your deflection number is partly a measure of friction, not success.
3. Are you measuring satisfaction on the bot path, not just the human path?
Plenty of teams survey resolved tickets and never survey the deflected ones. That’s a blind spot the size of your whole automation. Ask the people who never reached an agent how it went.
How the metric shapes your bill
This isn’t only about customer experience. It’s about money, and specifically about how your vendor gets paid.
Many automation tools tie their value, and sometimes their price, to deflection or “automated resolutions” that include anything the bot touched. Read that carefully. If a vendor counts a deflection as a win when the customer actually gave up, your incentives and theirs have split. They’re paid to keep people out. You need people helped.
A cleaner model pays for genuine resolutions and nothing else. At Fidiora you pay per real resolution. Human handoffs are free. Abandoned chats, silent timeouts, and escalations are never billed, because none of those solved anything. That’s the point of defining a resolution in the customer’s favor: the metric you pay for is the same one your customers care about. You can see exactly how that works on our pricing page.
When the bill and the customer pull in the same direction, you stop having to choose between a clean dashboard and a happy inbox.
Key takeaways
- Deflection measures an absence (no human ticket). Resolution measures a result (a solved problem). They reward different behavior.
- Deflection can be inflated by making the human handoff harder. Resolution can’t be faked the same way.
- Audit your own deflection: check repeat contacts, time-to-human, and satisfaction on the bot path.
- Watch how your vendor counts wins. If they profit from deflection, they profit when customers give up.
- Pay for outcomes, not avoidance. Resolve, don’t deflect.
Want a support metric that lines up with your bill instead of fighting it? See how Fidiora prices per resolution.
Resolve, don't deflect.
See Fidiora resolve a ticket, capture a lead, and keep the bill predictable.