I Asked for Feedback After I'd Finished

I showed a client a polished search feature and asked what they thought. We spent the meeting discussing a border.

I once showed a client a search feature I'd spent most of the week building. It had filters, tidy results, and a little message for when nothing matched. I'd even made the empty state look good, which is a questionable use of time when you haven't yet established whether anyone wants the full state.

At the end of the demo, I asked what they thought.

They asked whether the border around the search box could be darker. Someone else wanted to change a label. I wrote both things down and left the call feeling pretty good. A whole feature, two small changes. Development was going suspiciously well.

A few days later, we walked through how someone would actually use it. The client started with a question I hadn't designed the search to answer. They wanted visitors to describe what they needed help with and get pointed toward the right service. I'd built a very nice way to find a service if you already knew what it was called.

I could explain every choice I'd made. The filters matched the categories they'd given me. The results used their existing service names. None of it was random. I'd just made a fairly large assumption about the person arriving at the page, then spent the week making that assumption look finished.

The client hadn't hidden the requirement during the demo. We'd been looking at different things. I was asking whether my understanding of the task was right, but what I'd put in front of them looked like a task that had already been understood and completed. A border was an easy thing to comment on. Questioning the entire search meant potentially sending me back to the beginning.

And I wasn't exactly inviting that. I had walked them through all the things it could do, including a couple I was particularly pleased with. Then I asked a broad question and waited for feedback.

I wanted useful criticism. I also wanted them to like what I'd made. Those interests had become difficult to tell apart.

We went back to the question a visitor would bring to the site. The client described a few common inquiries, and we tried to follow each one through the search. It became obvious where a visitor would get stuck: choosing a service was the part they needed help with. Asking them to select it from a filter didn't do much for them.

Some of the work was still useful. We could keep the result cards and the service information. But the way into those results needed another attempt, and I had to let go of a screen that looked better than the idea behind it.

For the next review, I brought something much less impressive. There were a few rough screens with sample questions and the destinations they would lead to. I explained that we were checking whether those paths made sense before I built them. Then I asked the client to pick a common inquiry and walk through it.

They stopped at the second screen. Two choices sounded too similar. We changed the wording, tried the same inquiry again, and found another missing option. This was considerably less pleasant than hearing that the border needed work. It was also the conversation we should have had before I wrote the search.

Showing unfinished work can feel like arriving unprepared. You know which parts are ugly, and there's a temptation to fix them before anyone sees. But you can explain what you're showing. A rough draft with a specific question gives someone a way to help. An unexplained mess mostly gives them concern for your wellbeing.

The question makes a difference even when the work is nearly done. Asking whether a visitor can find help without knowing the service name gives the reviewer something to try. Asking what they think leaves them to guess which decisions are still open, how much trouble a change would cause, and whether you mostly want permission to ship.

The second version still needed visual polish. By then, changing the border was a reasonably safe way to spend an afternoon.

For occasional notes on work and deliberate decisions, you can join my newsletter.