What admin actually costs a professional firm

What counts as admin?

Admin is any work the client would be surprised to be billed for. That is a more useful test than any category list, because it separates the work that carries your professional judgement from the work that merely surrounds it.

In practice it usually falls into five buckets:

• Movement. Copying information from one system to another: an enquiry form into the CRM, a signed quote into the accounting file, a job number into a spreadsheet.
• Chasing. Following up quotes, invoices, documents, signatures and information requests.
• Assembly. Building reports, matter summaries, status updates and month-end packs out of scattered inputs.
• Triage. Reading a shared inbox and deciding who each message belongs to.
• Re-explaining. Answering the same internal question repeatedly because the answer is not written down anywhere.

Note what is not on that list: advice, drafting that requires judgement, client conversations, supervision and review. Those are the work. Counting them as admin is how firms end up with a scary number and no credible plan.

How do you measure the hours without a time and motion study?

Sample rather than census, and do it for two weeks. You are not trying to produce accurate accounting, you are trying to establish an order of magnitude, and two weeks of honest sampling beats a year of estimates made from memory.

Three approaches, in order of how much you should trust them:

1. A running log. Every person in the firm keeps one line per admin task: what it was, and how long it took, rounded to five minutes. Do it live rather than in a Friday catch-up, because reconstruction from memory systematically understates the small interruptions that do the most damage.
2. Calendar and inbox archaeology. Go back four weeks and count the recurring pieces: how many quote follow-ups you sent, how many invoice chases, how many times the Friday report was rebuilt. Multiply by a per-instance time you can actually justify.
3. Ask the team, then discount. People are reasonable at ranking which tasks eat their week and poor at estimating duration. Use their ranking to decide what to log, not as the measurement itself.

Two cautions. First, count the interruption, not just the task: a two-minute question that arrives eleven times a day costs far more than twenty-two minutes, because each one lands in the middle of something else. Second, log the cycle time as well as the hands-on time. A quote that takes fifteen minutes to prepare and sits for six days before it is sent has a fifteen-minute cost and a six-day problem, and the six days may be worth more.

Which hourly rate should you use?

There are three defensible rates and they answer different questions. Choose deliberately, because this single decision moves the final number more than anything else in the calculation.

The fully loaded cost rate. What it costs the firm to have that person in the chair for an hour. This is real cash, it is always available, and it gives you a floor.

`` Loaded hourly cost = (salary + on-costs) / genuinely productive hours per year ``

On-costs generally include superannuation, paid leave, workers compensation insurance, payroll tax if your firm exceeds the threshold in your state or territory, plus equipment, software licences, training and supervision. Rates and thresholds change and vary by jurisdiction, so use your own current figures from your accountant rather than a rule of thumb from an article.

The charge-out rate. What the hour would have earned if it had been sold. This is the right rate only when the firm is genuinely capacity constrained: you are declining work, quoting long lead times, or capping intake. If there is slack in the diary, the charge-out rate measures an opportunity you were not going to take, and using it produces a number your own partners will not believe.

The replacement rate. What it would cost to have the work done by someone else: a bookkeeper, a part-time administrator, an offshore assistant, or the monthly running cost of a workflow. This is the most useful rate for deciding between options, because it is the actual alternative on the table.

The honest approach is to run the calculation at the loaded cost rate for a conservative floor, and at the charge-out rate for the ceiling. The truth sits between them and the gap tells you how much of the case rests on being able to sell the freed time.

How do you turn hours into an annual number?

Multiply, then discount. The multiplication is trivial and the discount is where most firms are dishonest with themselves.

``` Gross annual cost = H x W x R

H = admin hours per week (from your log) W = your actual working weeks per year, after leave and quiet periods R = the rate you chose above ```

Then apply the recovery factor, which is the share of freed hours that would turn into something valuable rather than into more email:

``` Recoverable value = H x W x R x C

C = the proportion of freed hours you would genuinely convert ```

C is never one. An hour returned to a principal who is already at capacity is worth close to its full value. An hour returned to an administrator who then does other administration is worth very little in cash terms, though it may still be worth having for accuracy and turnaround. Pick C yourself, write down why you picked it, and be prepared to be wrong on the low side.

Run the whole calculation twice, once at the loaded rate with a conservative C, once at the charge-out rate with an optimistic C. You now have a range. A range you can explain is worth more than a single figure you cannot.

How do you know whether the number justifies a fix?

Compare it to the total cost of the fix, including what it costs to keep running, and look at the payback period rather than the headline saving.

`` Payback (months) = build cost / (monthly recoverable value - monthly running cost) ``

Running cost is the part firms forget. Any automated workflow has an ongoing bill: software subscriptions, per-run or per-token charges where AI is involved, and the small amount of human attention it takes to check that it is still working. If those are not in the sum, the payback is fiction.

A few judgements worth applying to the result:

• If the conservative version does not pay back within a year, be sceptical. The optimistic version is not the one to plan on.
• If the fix is a rule, a template or turning on a feature you already pay for, do that first and re-measure. Plenty of admin disappears without anything being built.
• Weigh the second-order effects qualitatively rather than pretending to price them. Faster quote turnaround, fewer things falling through, invoices going out the same week the work finishes, and less rework from copy-paste errors are all real, and none of them belong in a spreadsheet as invented percentages.
• Watch out for the fix that moves the admin rather than removing it. If someone now spends their week checking a workflow's output, you have bought a different job, not fewer hours.

What to do next

Start the log on Monday. Two weeks, one line per task, everyone in the firm, and no attempt to make it precise. When it is finished, rank the tasks by total hours, take the top three, and run the arithmetic above on those three only.

That gives you a defensible number and a shortlist in the same fortnight, which is enough to decide whether to fix something yourself, hire, or have it built. If you want the ranking done with you rather than by you, that is exactly what a Shift Flow Map is: a structured interview that turns the same two weeks of evidence into a one-page plan.