It arrived on a Tuesday afternoon, tagged low priority. A user reported that they had "completed the setup" but couldn't find their data anywhere. The message was short. No screenshots. No steps to reproduce. One sentence that sounded like a confused new user who hadn't finished onboarding.
Our support engineer replied with a link to the getting-started docs and moved on.
Three days later, the same user wrote again. Same complaint, different words: "I did everything the guide said. Nothing is showing up." This time the engineer dug in, confirmed the account was active, and found that setup had been completed — every step, in order, no errors. The data was there. It just wasn't where the user expected it.
That gap — between what the product did and what the user believed it did — turned out to be the most important signal we received all quarter.
When we designed the core workflow, we had a mental model. You configure your project, push data in, go to one place to see results. The flow was obvious to us because we built it. The screens connected in our heads before they connected on screen.
The user had a different model. They expected results to appear where they entered their configuration. They weren't wrong. They were operating on a reasonable assumption our product contradicted without explanation.
We pulled support logs for the previous two months. The same confusion appeared in eleven other tickets. Different words each time. Some users figured it out after clicking around. Some wrote in. Some — and this is the part that stung — probably just left. We'll never know how many.
Eleven tickets across two months. Not enough to trigger an alert. Not dramatic enough to escalate. Easy to file under "user education."
There are good reasons teams miss signals like this. The ticket didn't describe a bug. The product was working as designed. The fix seemed like "better documentation," which never feels urgent.
More honestly, we had a bias. We treated confusion as a user problem, not a product problem. If the system returned no error, the system was fine. The user just needed to read more carefully.
This is a comfortable lie. It lets you close tickets fast and keep shipping features. But it trades short-term velocity for long-term attrition. Every confused user who doesn't write in is a quiet departure you never get to diagnose.
Once we accepted the confusion was real and widespread, the question shifted from "how do we explain this better?" to "why does this need explaining at all?"
We reorganized the post-setup experience so the outcome appeared in context — right where the user finished their last configuration step. No extra navigation. No mental leap required. The product now matched the assumption most people carried into it.
The change was small in scope. A few screens rearranged. Copy rewritten. One transition removed entirely. It took less than two weeks to ship. It didn't appear on any roadmap before that ticket.
Within a month, support tickets about "where is my data" dropped to near zero. But the number that mattered more was activation rate — the percentage of new users who completed setup and actually used the product within 48 hours. It went up. Not by a trivial amount.
We had been losing people at a step we didn't think of as a step. No button to click, no form to fill out. Just a moment of disorientation where the user had to figure out where to go next. That moment was enough to stop some of them cold.
The insight wasn't about that specific workflow. It was about how we categorized feedback. We had drawn a hard line between "bug" and "confusion" and treated the second category as lower priority by default. That hierarchy was wrong.
A bug means the product doesn't work. Confusion means the product doesn't communicate. Both cost you users. But confusion is harder to detect because the logs look clean. No errors. No crashes. Just a person sitting in front of a screen, uncertain what to do next, deciding this product isn't for them.
Now, when a support ticket describes confusion — even a single ticket, even from a brand-new user with no screenshots — we treat it as a product signal, not a documentation problem. We ask: what assumption did this person carry in, and why did our product break it?
The most important feedback we received that quarter looked like a low-priority ticket from someone who hadn't read the docs. It would have been reasonable to close it. It would have been wrong.
If you build products, you're surrounded by signals disguised as noise. Vague complaints. One-line tickets. Users who "just don't get it." The temptation is to filter them out and focus on the clear, reproducible, well-documented issues.
Resist that. The users who describe their confusion clearly are doing you a favor most people won't. The real cost isn't the twelve tickets you received. It's the hundreds of silent exits from people who felt the same friction and never said a word.
Be the first to comment.
0 comments
Loading comments...