Shadowing beats surveys for internal tools
Six afternoons with finance managers taught me more than 200 survey answers.
Note
Notes are where things go before they turn into a case study. Some never do, and that’s fine.
We started the approvals project with a survey. Two hundred managers told us approvals were slow because the app was slow. Load times were fine.
Then I asked six of them if I could sit next to them for an afternoon. Nobody was slow at clicking. They were slow at deciding, because they were afraid of approving something wrong.
They opened every receipt, checked the hotel name against the trip, and looked up whether this person usually spent this much. All of it by hand, in other tabs.
No survey question would have found that, because nobody thinks of their own checking as a problem. It’s just the job.
For internal tools, I now shadow first and survey second. The survey is good for counting how common something is. Shadowing tells you what to count.