The problem
Since FIAT's discovery phase, users have told us it's important they can trust the data we show them. This includes knowing where it came from, and how recent it is.
Our previous design explained this inside a details component near the bottom of each page.
The always-visible link text said "Sources and updates".
Clicking on it revealed a short piece of content explaining where each item of data on the page came from, when it was last imported and when the next import would be.
Testing has since shown that some users did not:
- realise the information was there
- understand what it said
What we found
We set some time aside during our private beta research to learn how users were interacting with the component, and what they thought about the information it contained.
Some found it useful as they often add it to documents and like knowing who's responsible for the data:
- “That's really good. Because that's one of the most important things, like knowing how up to date and where things are found or where they're sourced from. So yeah, that's cool.”
- “I can see where it's all come from. That's good. That way if it's wrong, we know who to talk to.”
But there were other users that either did not spot it or did not understand what it meant:
- “I didn't sort of realise that was there.”
- “it wasn't immediately obvious”
- "I guess it's how often that source of data is updated though as well, isn't it? That's the thing. It might be pulling through every day, but if that's only updated like every few months..."
The link text
Details components can be a good way to provide information for those who need it without it getting in the way for those who don't.
But because they intentionally hide content, they also carry a risk that users will miss what's inside them.
Our link text said "Sources and updates" because it was short and described the content inside the component.
All links need to be understandable by the person reading them so that they can decide if clicking on them would be useful.
One user told us that they had been asking for information about the data we provide in FIAT but hadn't realised it was already there.
We suspected our copy may have:
- felt out of context
- seemed to relate to something technical
User understanding
Another problem we found was that the words we had chosen didn't always match the mental model of our users.
We wanted to let users know when the information they find in FIAT was last updated, but what we meant by "updated" wasn't what users understood by it.
We meant:
- When we last took the information from the source.
What users often thought it meant:
- When the source information was last updated.
Because users were including these details on documents that would be presented at advisory boards, it was important they understood what they really meant.
What we chose to do
We chose to change the link copy to "Where this information came from", as we felt it:
- removes potentially technical language
- more clearly associates it with the data on that page
- avoids the word "update"
We also tried to make the content within the component easier to read and understand by:
- using plain language sentences
- listing data sources with bullets
- replacing "last updated: [date]" with "taken from...on [date]"
- breaking content into headed sections
Where a page contains more than one data source, each source is now listed individually along with the date of the last import.
For pages where all of the data comes from the same place, we've reduced the copy to a single bullet item.
How it tested
Users said that they were happy with this language and found the new copy clearer to read.
They also wanted to know:
- why FIAT data doesn't update instantly when the source data changes
- why the data is captured from those sources at different times
We felt that explaining this would involve too much copy for the component area, so we've added it to our backlog as a separate user need.
What next
We made the following assumptions that we want to test and may need to iterate:
- The new component link copy uses simpler language but we still need to see how well users associate it with the content higher up the page.
- If feedback shows that users still assume the data they see in FIAT pages always matches the source data, we may need to be more explicit about this.
The team are also planning to investigate how we might provide information around our import cadence and where that information might sit.
Finally, because transparency around the data we provide is a common need, this work will help form the basis of a cross-cutting design pattern which other teams can reuse and improve on.