The Admin-Time Audit: Find the Hours Your Team Loses Every Week

Ask any small-business owner how much of the team's week goes to actual revenue work — building, selling, serving — and you'll get a confident, comfortable number. Ask where the rest goes and the confidence evaporates into a hand-wave: "you know… stuff. Admin." Nobody budgets for admin time, which means nobody measures it, which means it can quietly eat a startling share of a small company's week while remaining officially invisible.

The invisibility has a specific mechanism. Admin time almost never arrives as blocks; it arrives as crumbs. Six minutes forwarding a PTO question. Nine minutes answering the same how-do-I-reset question for the third customer this week. Twelve minutes hunting for the version of the onboarding doc that's actually current. Fifteen minutes walking the new hire through the thing nobody wrote down. No single crumb feels worth counting, so none get counted — and a team of eight can lose the equivalent of a full working day, every week, in pieces too small to see.

You cannot automate, delegate, or fix what you can't see. So before any tooling decision, run the audit. It costs one week of honest tallying and roughly zero dollars.

Why your estimate is wrong (in a predictable direction)

People are decent at estimating time spent on scheduled, named work — meetings, projects — and terrible at estimating interruption-shaped work, for a well-documented reason: memory encodes events, not durations. A ninety-second interruption that cost you eight minutes of refocus is remembered as ninety seconds, if it's remembered at all. Multiply that accounting error across every crumb, every person, every week, and self-reported admin time comes in dramatically under reality. The audit's whole job is to replace recall with capture.

The one-week tally

The method is deliberately primitive, because a method with any friction dies by Wednesday. Each person keeps one tally sheet — paper, a note on their phone, wherever — with four rows:

1. People & paperwork — anything HR-shaped: PTO questions, onboarding walk-throughs, policy lookups, form-chasing, "quick questions" about pay, benefits, or who-approves-what.
2. Answering the outside — customer questions that interrupt other work: the repeated how-tos, the status checks, the same five answers retyped in slightly different words.
3. Explaining the inside — ad-hoc teaching: showing a colleague the process again, the shadow session, the "let me just walk you through it" that replaces training that was never built.
4. Hunting & retyping — searching for documents, figuring out which version is current, re-entering the same information into a second system, reconstructing knowledge that exists but can't be found.

The rule: every time a crumb happens, one tick and a rough minute count. No descriptions, no justification, no app. Five seconds per entry, or people stop entering. On Friday, collect the sheets and add it up — total hours per category, across the team. Brace slightly. In the audits I've run and the ones owners have described to me, the total for a sub-20-person company almost always lands between "a full working day" and "a full working week" — and the owner's estimate was usually a third of the real number.

Reading the results: four buckets, one honest line

Every recurring item in the tally belongs in exactly one bucket, and the sorting question is the same each time: does this work require judgment and relationship, or is it retrieval and repetition?

Automate the retrieval and repetition — the answers that are the same every time, the document that should find itself, the walkthrough that should be a course, the form-chasing a system should do. This bucket is nearly always the biggest, and it's the one this suite exists for. Design out the crumbs that exist only because a process is broken — the version-hunting that disappears the day there's one source of truth, the retyping that disappears when two systems become one. Automating a broken process just breaks things faster; fix the design first. Delegate what genuinely needs a human but not that human — the owner answering questions a well-documented process would let anyone answer. And keep — deliberately — the crumbs that are actually relationship in disguise: the check-in conversation, the personal reply to an upset customer, the teaching moment with a struggling new hire. The audit isn't a purge. It's how you find out which interruptions were the job all along, and which were just the absence of a system.

From tally to plan

Rank the automate and design-out buckets by hours recovered, and take on one category per month, biggest first — not everything at once, for the same reason documentation sprints fail. A team that recovers even half of a lost day per week has effectively hired a part-timer for the cost of paying attention for five days. If you want the fuller version of that sequencing — which system first, what "done" looks like, how the pieces connect — the 90-day back-office plan below walks the whole arc.

The bottom line

Admin time is the only major cost at a small company that never appears in any report — unbudgeted, unmeasured, and paid every week in the team's scarcest currency. One week of five-second tallies makes it visible; four buckets make it decidable; one category a month makes it fixable. The hours were always there. The only thing the audit changes is that now you get to choose where they go.

— Tom