Why does the question get asked on Tuesday instead of Friday — or not at all?
It's Friday afternoon. Someone wants to know how ad spend is tracking against actual purchases. Not a deep-dive — just a quick gut-check. But getting the answer means opening a spreadsheet, pulling data from two places, aligning columns, formatting the paste so it's readable. Twenty minutes, minimum. So the question doesn't get asked. It gets deferred to next week, or dropped entirely because the answer isn't worth the effort of the setup.
That's not a data problem. That's a friction problem.
What actually happened
Someone replaced that twenty-minute ritual with one question in a chat interface. Type the question, get the answer back — ad spend versus actual purchases, live, no spreadsheet, no file to download, no column-aligning.
The twenty minutes didn't go away because the task became easier to do manually. They went away because someone built a path that removed the manual work entirely. The question and the answer are now the whole interaction.
The part that matters more than the time saved
Saving twenty minutes on a Friday is a decent outcome. But it's not the interesting outcome.
The interesting outcome is that the question now gets asked on Tuesday. Or Wednesday. Or whenever it becomes relevant — not whenever the friction budget allows.
When answering a question costs twenty minutes of mechanical work, you unconsciously ration your questions. You batch them. You wait until you have enough to justify the setup cost. You skip the ones that feel like they might not be worth a one-liner answer. The questions that don't make the cut aren't logged anywhere. Nobody counts them. They just don't happen.
That's what friction does to curiosity: it applies a hidden tax. The tax is invisible because you never see the questions that didn't get asked. You only see the spreadsheet you opened.
Why this doesn't show up in any measurement
The cost of the twenty-minute column-alignment ritual is easy to see in retrospect — once someone removes it, you notice the time that used to disappear. But the cost of deferred or skipped questions has never appeared in any spreadsheet, because there's no row for "question that didn't get asked."
This is why the problem persists for so long in most teams. The friction is real, but it's invisible in the places where teams measure cost. No one is tracking how often an analyst check-in gets pushed from Tuesday to Friday. No one is tracking how often it gets pushed from Friday to never. The spreadsheet shows the work that happened. It can't show the work — the thinking, the curiosity, the small corrections — that the friction quietly killed.
What changes when the friction is gone
When a question costs one line in a chat instead of twenty minutes of prep, the calculus around asking it changes completely. A question that wasn't worth the setup cost is now worth asking on the spot. The threshold drops. Curiosity that used to get deferred or abandoned now just... happens.
And that's not a small shift. The difference between a team that checks ad spend vs. purchases once a week on a fixed schedule and a team that checks it whenever something feels off is the difference between catching a problem on Friday and catching it on Tuesday. Or catching it at all.
Where this logic breaks
This only works if the answer coming back is actually reliable. A fast wrong answer isn't better than a slow right one — it's worse, because it removes the friction without removing the doubt. If the team has to verify the chat answer against the spreadsheet anyway, the twenty minutes didn't go anywhere; they just moved.
The other failure mode: the question gets asked more often, but nobody acts on the answers. Friction was never the bottleneck there — something else is. Removing the friction in that situation just produces more ignored answers faster.
What to do with this
Look at the questions your team batches to a fixed weekly ritual. Ask which ones are batched because it makes sense to review them together — and which ones are batched because the setup cost makes asking them individually unreasonable.
The second category is friction wearing a schedule as a disguise. Those are the questions worth making cheap to ask. Not because saving the time is valuable on its own, but because the question will start getting asked when it's actually relevant — and that's a different kind of useful.