Research
Why we stopped asking support to tag tickets
Tags looked like a cheap way to sort feedback. In the eleven teams we studied, they were neither cheap nor sorted. Here is what we saw.
Most teams we talk to started the same way: a list of tags in the help desk, and a request to support to pick one for every ticket about a missing feature. It seems sensible. A year later, the list has two hundred tags and nobody trusts the counts.
What we looked at
Over the summer we read the tag lists of eleven small software teams, with their permission and with every customer detail removed. We counted how tags were used and asked the support leads how they chose them.
What we found
Three patterns came up in almost every team.
- The list keeps growing. New tags are added for new requests, old ones are never merged. One team had four tags for the same export problem.
- Busy days get fewer tags. On the days with the most tickets, the share of tagged tickets dropped by about a third. The feedback you most need to count is the feedback least likely to be counted.
- Tags lose the words. A tag says exports. The ticket says the export fails for projects over 10,000 rows and the customer has a board meeting on Thursday. Only the second one helps you plan.
Every support lead we spoke to said the same thing in different words: tagging is the first job to go when the queue gets long.
What we changed
Meridel reads the ticket itself. Support keeps answering customers, and the grouping happens without them. If your team still likes tags, Meridel imports them and shows them beside its own themes, so you can compare.
The goal was never to get rid of tags. It was to stop depending on a step that people under pressure skip.