Decide who does the work
From a dashboard I designed for wholesale sales reps, who take retailers’ orders for manufacturers, and the agencies they work for.
Reps and agencies used to exporting everything to Excel pushed hard for a download on the dashboard, so I included one, and in the first six months after release fewer than one active user in twenty used it. Research suggested why: the dashboard had taken over the work they’d expected to keep doing in spreadsheets. Every design decides who does work like that, the product or the person.
The work was keeping track of orders after they were taken. An order could sit in pending because a retailer’s payment hadn’t reached the manufacturer, and the rep might not find out until the retailer, who didn’t know either, called asking where the order was. An order is a promise the rep makes to a retailer, so that call is the worst way to learn one is in trouble. Power users exported order tables and dropped them into a prepared spreadsheet workbook, whose other sheets updated to show which orders were held up. They had to export again whenever they wanted a current answer, and reps without that setup had little warning at all.
Larry Tesler’s point about software applies here: some complexity can’t be removed, only moved, and the platform had left this complexity with the reps. I moved it into the dashboard, so a rep no longer had to export, import, and search a workbook to find held-up orders. Cards showed an up-to-date count of the orders in each held-up status, pending among them, and each count opened into the orders behind it. A count showed where orders sat, not which ones were in trouble: a rep might know from a conversation that a retailer had agreed to a wait, while another order in the same status needed a call that day. The rep still read the orders behind a count and decided which ones to act on.
Usage showed most people saving a table of orders set up the way they wanted and pinning it so it was there every time they came back, rather than downloading anything. In research sessions after release, reps said they were finding held-up orders before retailers called about them. The download stayed in, mostly unused, for the few who still wanted a file.
Inside the company, I had to make the same case: that the dashboard, not a spreadsheet, should do this work. One product executive expected reporting to mean printable reports, and what changed their mind was watching a rep in a research session go from a count of pending orders to the few that needed a call, without exporting or printing anything.
When a product leaves people to export and maintain a spreadsheet, the work hasn’t gone away; it has landed on them. I count the spreadsheet people keep beside a product as part of that product, whether or not anyone designed it.