What 15 minutes a day is actually worth: the arithmetic of everyday friction

The four numbers

Pick one task your business does every working day: invoicing a job, confirming a booking, re-typing an order into a second system. Time it honestly. Then ask what it would be worth if it took fifteen minutes less, or ten, or five, or one.

Time here is valued at $46.89 an hour. That is the average total cost of an hour of work in US private industry, wages plus benefits rather than the wage alone, from the Bureau of Labor Statistics release for June 2026. A working year is 260 days.

Then multiply it by your team

Friction is almost never one task for one person. It is a handful of small delays a day, for everyone who touches the process. The value recovered per year, at the same hourly cost:

Where the minutes actually go

Asana's Anatomy of Work Index, a survey of 9,615 knowledge workers, found leaders losing 62% of the workday to repetitive, mundane tasks and 3.6 hours a week to unnecessary meetings, while juggling around ten apps a day. The pattern in small businesses is the same shape, in miniature:

• The same customer detail typed into a booking sheet, an invoice and an email
• A phone call to confirm something a form could have captured
• A weekly report rebuilt by hand from three exports
• Chasing a status that no system records, so someone has to be asked
• Approvals that wait in an inbox because nothing tells anyone they are waiting
• A price or a policy that lives in one person's head, so every question routes through them

Why one minute is never one minute

A minute of friction looks too small to buy back. It rarely costs only a minute. In a detailed observational study of 24 information workers, Mark, González and Harris found people spent an average of 11 minutes on a working sphere before switching, and 57% of those spheres were interrupted. Interrupted work was picked up again the same day 77% of the time, and getting back to it took an average of 25 minutes and 26 seconds, after an average of 2.26 other activities in between.

That is the real cost of the one-minute problem: not the minute, but the break in concentration it forces, and the work that does not get finished because of it. The table above is the floor, not the ceiling.

Be skeptical of savings claims, including mine

Published productivity claims are usually marketing. The measured numbers are far more modest: in a nationally representative US survey of over 10,000 working-age adults, Bick, Blandin and Deming found that generative AI assisted between 1 and 5 percent of all work hours, and users reported time savings equivalent to just 1.4% of total work hours.

That is the honest frame. Nobody is giving you back half your week. What software does well is delete a specific, repeated step permanently, for everyone, without anyone remembering to do it. The way to know whether that is worth paying for is arithmetic on a task you have actually timed, which is why this guide does not contain a single projected benefit.

Measure your own minutes before you spend anything

• Pick one task that happens every working day and time three real runs of it with a stopwatch
• Count how many times a day it happens, across everyone who does it
• Write down which part is judgment and which part is copying, checking or chasing, because only the second part can be removed
• Multiply: minutes removed × times a day × 260 days, then value the hours at what an hour of that person's time costs you
• Compare the yearly figure with the price of fixing it. Custom work starts at $2,500, plus 15–20% of the build price a year to keep it running
• If the first year does not pay for itself, say so out loud and pick a different task

What removing the minutes usually looks like

• Entered once: systems that share data instead of asking someone to re-type it
• Booking and scheduling that confirms itself, so the phone tag stops
• Reports that build themselves on a schedule and land in an inbox
• Approvals and reminders that chase people automatically, with a record of what ran
• One place to look something up, so questions stop routing through one person

When the minutes are not worth buying back

• The task happens monthly rather than daily, so the arithmetic rarely clears the build price
• The slow part is the judgment, not the keystrokes; automating a decision you have not settled yet makes it worse
• The process is about to change anyway, so you would be paying to automate a draft
• The yearly value recovered is smaller than the care plan needed to keep the software alive (from $250 a month)
• An off-the-shelf tool you already pay for does it, and nobody has read its manual

Questions

Sources

1. U.S. Bureau of Labor Statistics: Employer Costs for Employee Compensation, June 2026
2. Asana: Anatomy of Work Index (9,615 knowledge workers)
3. Mark, González & Harris: No Task Left Behind? Examining the Nature of Fragmented Work (CHI 2005)
4. Bick, Blandin & Deming: The Rapid Adoption of Generative AI (NBER Working Paper 32966)
5. Standish Group CHAOS report (summary)