Short answer: Log every manual task for two weeks with its frequency and duration, multiply out to an annual figure, then value founder hours at your effective billing rate and team hours at fully-loaded cost. Most founder-led businesses under thirty people find between six and twelve recoverable hours per week for the founder alone.
The reason this calculation matters isn't guilt. It's that it converts a vague feeling — "everything takes too long" — into a number you can compare against the cost of fixing it.
Keep a simple log: task, how long it took, how often it happens. Two weeks catches weekly and fortnightly cycles; note monthly tasks separately from memory. Include the small things. Copying a client's details into a second system takes ninety seconds and happens forty times a month — that's an hour, and it's invisible until you write it down.
Be specific about who does each task. Founder hours and admin hours have different values and different opportunity costs.
Split the log into three buckets:
The third bucket is usually the largest and the most commonly overlooked.
Multiply each recoverable task by its annual frequency. A ten-minute task done daily is roughly forty hours a year. A three-day monthly reporting cycle is thirty-six days a year. These numbers tend to surprise people, because the individual instances feel trivial.
Use two different rates:
Founders routinely undervalue their own time by using the salary figure. That systematically understates the case for fixing the system.
Manual processes generate errors, and errors have costs that don't appear in a time log: an invoice sent to the wrong address delays payment; a missed follow-up loses a deal; a duplicated record means a client gets contacted twice. Estimate conservatively — even a rough figure changes the total materially.
Across the businesses we've audited, founders typically recover around eight hours a week and error rates drop substantially once data has one home. The honest caveat: that figure comes after the system is built, and building takes time first. A realistic model assumes three to six months to net positive, not three weeks.
Only the portion that exists because information isn't visible in a system — status meetings, reconciliation calls, "where are we on X" check-ins. Those are reporting failures wearing a calendar invite.
Then don't do the project. That's a legitimate and reasonably common outcome for very small or very low-volume businesses.
Within about twenty percent. You're making a go/no-go decision, not filing accounts.
Want the version of this done properly, with someone else counting? Book a free systems audit.