Let people mark what they can’t explain yet
From a dashboard I designed for wholesale sales reps, who take retailers’ orders for manufacturers, and the agencies they work for.
At a session of our monthly council of agency owners and reps, I asked one agency owner why a number on their dashboard was violet. They said it meant the number was interesting, and when I asked what interesting meant, they couldn’t say.
The dashboard let people set their own threshold on a number’s card, so the card changed color when the figure crossed it. Thresholds were built for spotting trouble and started with warning and error colors. A few weeks later, people asked for colors that didn’t mean trouble, and we added a green for good news and a violet for news that was neither. Violet was meant for plain notices, where the person knew exactly what crossing the threshold meant, like open orders reaching the number where a rep set aside a day to follow up. Green got used as expected, and so, mostly, did violet.
Five or six agencies, the owner’s among them, used violet differently, on numbers where they didn’t know what crossing the threshold would mean. The owner’s card counted the retailers in one territory who had ordered in the last 30 days, with the threshold set just below where the count usually sat. All they could tell me was that they wanted to know if the count dropped below it. Choosing violet over a warning color suggested they couldn’t say whether a drop would be bad news.
The other violet cards I found looked much the same: a count of a territory’s retailers or orders, each with a threshold its owner couldn’t explain. My read is that people were watching a hunch about how a territory was doing, like whether it was cooling before any single retailer looked like a problem. No number on the dashboard measured that, so they watched the closest one they had.
The threshold feature ended up doing four jobs: flagging trouble, showing good news, giving plain notice, and keeping watch on a hunch. More colors raised a fair worry about cluttered, clashing cards, and in the months we watched, neither problem came up in research sessions or surveys. Violet didn’t tell me what to measure next, but it showed where to look. Violet cards became the first thing I asked about in research sessions, and “What made this one interesting?” got better answers than asking people which numbers they wanted.
A color stayed on a card until its number moved back across the threshold, even after someone had looked into it, and some people asked for a way to clear it by hand. I left it tied to the number on purpose: someone keeping watch on a hunch needs to know whether the number has changed, and a color they could clear by hand would only tell them they’d already looked.
People often start watching a number before they can say why. The most useful thing a dashboard can do then is let them watch it, keep the records behind it within reach, and not ask for a reason.