From your own hours and your own cost, not from a published benchmark, because no credible one exists. Compliance is benchmarked because it is comparable between firms. Advisory is not.
Ignition's 2026 Australian benchmark surveyed 160 firms and is genuinely useful on compliance: individual returns most commonly above $300, company and trust compliance clustering at $2,000 to $3,000, 78% of firms planning a rise, most commonly 5% to 10%.
It says nothing about advisory rates. That is not an omission. It is because an hour of advisory at one practice and an hour at another are not the same product, so an average of them would be a number with no meaning that firms would nonetheless price against.
Here is what to use instead.
Start from the four numbers you control
| Number | Where it comes from | What goes wrong without it |
|---|---|---|
| Hours the last comparable engagement took | Your own timesheets | You price from memory, which omits the calls, and memory is always low |
| Your cost per hour for whoever delivers it | Salary plus on-costs, divided by productive hours | You price off the partner rate for work a manager does, or the reverse |
| Hours already given away to this client | Time logged against the client with no rate attached | You price the new work and keep funding the old |
| What this client currently returns | Gross profit per client, after direct cost | You raise the fee 8% on a client who needs 40% |
Only the second of those is an accounting exercise. The other three are a record-keeping exercise, and they are the reason most advisory pricing is guesswork: the practice has never held the hours, so there is nothing to price from.
The three shapes, and when each is wrong
Fixed price. Right for anything with a defined output: a three-year plan, a cash flow model, a board pack, a restructure recommendation. The client knows what they get and you know what you are pricing. It fails when the scope is "and then we will see", because you have fixed the price and left the delivery open.
Retainer. Right for an ongoing relationship. Fails, reliably, when the fee is stated and the hours are not. A retainer of $2,000 a month with no hours attached has no definition of too much, so there is no month in which anyone notices it has become too much. It just gets slightly worse for two years and then somebody resigns the client.
Hourly. Honest, and right for genuinely unscoped work. The problem is that it caps your practice at the number of hours in the week, which is the argument in pricing a fractional engagement. Use it as a fallback, not as the model.
The common failure across all three is the same: the engagement is priced once, at the start, and never checked against what it took.
A method you can run this month
- Pick the last three advisory engagements you delivered. Not the ones you are proud of. The last three.
- Reconstruct the hours as honestly as you can, including the calls that were not logged. If you cannot, that is your first finding and it is more important than the pricing question.
- Cost them at the rate of whoever actually did the work, not at the rate you would like to charge.
- Compare that to what you invoiced. The gap is your current advisory margin, and for most practices seeing it for the first time it is negative or close to it.
- Price the next one from step 2 plus your target margin, and attach hours to it so you find out within a month if you were wrong rather than within a year.
Step 2 is the whole exercise. Everything else is arithmetic.
Why repricing conversations are easier than firms expect
The fear is that clients will leave. The 2026 data does not support how large that fear usually is: of the Australian firms that raised prices, 45% said most clients accepted it without issue and 3% reported a negative impact on profitability.
That does not mean repricing is free. It means the practices that handled it well were the ones who could say what they were charging for. "Our fees are going up 8%" invites an argument. "This engagement took 26 hours last year and was priced for 14, here is what it covers" is a different conversation, and you can only have it if the 26 exists somewhere.
What has to be true in your systems first
You need hours against clients, categories that separate compliance from advisory, rates attached to people, and a profit and loss per client that counts the direct cost of delivery.
That is what Empiraa GPS holds for professional services firms: timesheets logged against the client goal the work moved, fixed-price projects with an hours budget that flags when the work has gone past what the progress has paid for, retainers showing hours used against hours given, and a gross profit and loss per client. Where an hour has no rate on it, GPS counts the hours beside the figures and withholds the gross profit rather than showing one that is too high, which matters when the number is about to become a fee conversation.
The short answer
There is no number to copy. There is a method, and it starts with knowing what last year took.
If you only do one thing before setting 2027 fees, make it this: get the hours recorded by category from 1 December, so that when you price in June you are pricing from six months of evidence instead of from the same guess you made last year.
Related reading: 93% of firms say they offer advisory, so why does compliance still pay the bills, and Tranche 2 added hours to your practice, who is paying for them.
Compare GPS Professional Services plans, priced by client account with unlimited practice and client users on every plan.
