The hardest rule: do not explain
If I give someone Culture Atlas and immediately say, “First click this, then use this filter, and this colour means…”, I have not really tested the product.
I have tested my explanation.
So my first rule for a future user test is simple: watch before helping.
I would start with a few small tasks
I might ask someone to find a place that interests them. Then I could ask them to find a story, language or tradition connected to it. I could ask them to return to the map, or use a filter to discover somewhere else.
The goal would not be to see whether they are “good at” Culture Atlas. The goal would be to see where Culture Atlas makes them stop and think, “What am I supposed to do now?”
I would watch for questions
Where do they click first? Do they understand that clustered circles can be zoomed into? Do they notice the map/list switch? Do the filters make sense without an explanation? When they reach a place page, do they follow a connection?
I would also want to know what they ignore. A button nobody notices may be telling me something.
Confusion is useful evidence too
If a person clicks the wrong thing, I should not immediately decide that they made a mistake.
Maybe the interface gave them the wrong clue.
This feels similar to the research lessons from earlier in the journal. I should observe what happened before deciding why it happened.
This entry is a plan, not a result
I have not run this user test yet, so I do not want to write imaginary reactions from imaginary users.
The next useful evidence has to come from actual people using the Atlas.
The question can accidentally give away the answer
Even a simple task like “find a story connected to this place” tells the person that such a connection exists and that they should look for it.
So I would need some open tasks too, such as “Explore for a few minutes and tell me what you think this site is for.” That would show me what the interface communicates before I give away the answer.
A fair test should not contain the answer
This reminds me of checking a maths idea: if I design the test so that it can only confirm what I already believe, the result is not very useful. I also have to think about the audience, just as I do in English, and separate what I observe from what I think it means, as I do in I&S.
The next evidence has to come from users
A product can make perfect sense to the person who built it and still confuse everyone else.
The only way to find out is to let someone explore without putting my answers in their head first.
What I'm figuring out next
When I run the first proper test, I want to record what people actually do, where they hesitate, what they ask and what surprises them. Then I can compare my idea of Culture Atlas with their experience of it.
End of this phase
Entries 01–19 follow Culture Atlas from the first idea through research, coding and the plan for real-user testing. I am stopping here until there is real new evidence to write about.