Blog
7 min read
Effective hourly rate is the fee divided by the hours you actually spent delivering the work. Not the number on your rate card. The number the job earned. Learning how to calculate effective hourly rate on a fixed fee project takes a minute. Most agency owners get the single-job version right first time.
The roll-up is where it falls apart. One figure for a whole book of work hides the projects that are underwater, because a couple of disasters and a run of quiet successes produce exactly the blended number that reads as healthy.
Divide the fee by the hours delivered. That is the whole formula.
Everything after that is an argument about which hours count. Take a brand identity job quoted at $18,000. Your rate card says $150 an hour, so the estimate assumed 120 hours. Delivery booked 164 by the time the last round of amends closed.
$18,000 over 164 hours is $110 an hour. Nobody lost money here. The job returned $40 an hour less than the rate card claims. Quieter than a loss, and more expensive, because it repeats.
Three inputs. The second one causes all the trouble:
Runn's documentation frames the same metric as a reference rate per role based on the actual charge per billable hour worked. Same idea, cut by seniority rather than by job.
Every hour that touched the job. Delivery time, coordination, the call about the call, the two days of amends nobody thought to log.
This is where the sum quietly breaks. Agencies book design and development hours religiously, then leave account management out of the total on the grounds that servicing time is overhead and overhead belongs somewhere else. It is not overhead. It is the cost of one specific client, and stripping it out is how a job that ran at a loss gets filed as fine.
Bennett Financials, writing on how agencies lose money on profitable clients, puts the indirect cost a single client generates at 30 to 40% of total client servicing costs. Account management time, internal meetings about the client, proposal revisions, onboarding, offboarding. Almost none of it gets allocated back to the account that caused it.
Then set the floor. Parakeeto's agency rate guidance recommends a target of at least 2.5 times your average cost per hour. An effective rate under that is not a rounding error. It is a job you priced wrong or delivered badly, and knowing which one matters more than the number itself.
If your cost base carries licences, hosting or third-party tooling, fee over hours is too blunt on its own. Our piece on which implementation projects are actually profitable once licences and rework are counted works through that version.
Because you already knew about that project.
Every agency owner can name the job that went wrong last quarter. Computing its effective rate confirms something the delivery lead told you in week five. Useful for a post mortem. Useless as a warning system. That is how the metric earned its reputation: interesting, never actionable.
It earns its keep only when you run it for every project closed in a period and read the whole set at once, ranked, with the hours attached. Then it stops being a score and starts being a map.
Sum every fee, sum every hour, divide once. Weight by hours rather than by project, or six small jobs will outvote the one that consumed half your delivery capacity.
Here is an invented six-project quarter at an agency with a $150 rate card. The figures are illustrative. The shape of them is what repeats.
| Project | Fee | Hours delivered | Effective rate |
|---|---|---|---|
| Retainer A | $60,000 | 400 | $150 |
| Website C | $45,000 | 300 | $150 |
| Launch E | $30,000 | 200 | $150 |
| Rebrand B | $18,000 | 164 | $110 |
| Campaign D | $12,000 | 190 | $63 |
| Microsite F | $9,000 | 146 | $62 |
$174,000 over 1,400 hours comes to just over $124 an hour. Against a $120 floor, the quarter passes. Nobody has a reason to open the file.
Two of those six jobs returned less than half the rate card. They burned 336 hours doing it, which is 24% of the quarter's delivered hours spent earning 12% of the fee income. The blend absorbed them because three clean projects carried enough hours to cover the damage.
Now run the same quarter forward with one change. Retainer A renews at the same fee but takes 500 hours instead of 400, which nobody would call a crisis. Blended rate drops to $116. You have crossed your floor without a single new problem project. No report will name the job that did it.
A blended figure answers a question nobody urgently needs answered. The spread answers the one that pays.
Bennett Financials makes the same point about client margin. Your agency-wide gross margin is a weighted mean of every client margin in the book, so one bad account drags the whole figure down. Two or three, and your best clients are quietly subsidising your worst ones. That same arithmetic applies project by project, which is why a single number cannot tell you where to intervene.
So sort the list by effective rate, lowest first, and ask three questions of the bottom of it:
That first one is what you track quarter on quarter. A blended rate can hold perfectly steady while the share of hours below your floor doubles underneath it, because your strong projects improve at the same time your weak ones decay. Two trends. One figure, reporting neither. Our other posts on margin and profitability come back to that pattern repeatedly.
Question two is what changes your pricing. Underwater jobs are not random. They cluster by size, by service, or by the one client whose approvals take four rounds. When Campaign D and Microsite F turn out to be the two smallest fees in the quarter, you have not found two bad projects. You have found a minimum viable fee you are pricing under.
Set the bucket boundaries once and reuse them. Above rate card, between rate card and floor, below floor. Counting projects in each band takes a pivot table, and it will tell you more than any blended figure ever does, because the shape of that distribution starts moving long before the average notices. A quarter that moves two jobs from the middle band to the bottom is a quarter that needs a conversation, whatever the headline rate says.
Start with the third question, because the answer decides who owns the fix.
A job that was mispriced at proposal is an estimating problem. The hours were always going to land where they landed, and the fee was set weeks before anyone opened a file. Fix that upstream. In the scoping conversation, and in the fee itself.
A job that started at rate card and finished at $62 an hour is a delivery problem. It was visible while it happened, in the timesheets, weeks before the final invoice. The leading indicators that a fixed-price job is losing money before it finishes are the checks that catch that version. They only work mid-flight, while someone is still looking.
Most agencies have both and treat them as one problem. That is why the fix never sticks. You cannot re-scope your way out of a delivery drift. No amount of weekly cost review rescues a fee that was 30% short on the day it was signed.
Pull last quarter's closed projects into a spreadsheet with two columns: fee and hours booked. Sort by fee divided by hours, ascending. The bottom three rows are your next pricing conversation, and you will recognise at least one of them before you finish typing the list.
COMMON QUESTIONS:
RELATED READING:
The biggest invoice of your year is rarely your best project. Here is how to adapt client profitability analysis to a licence-heavy IT services cost base.
Status meetings track progress. Margin reviews track money. Five numbers per live project, a threshold against each, and a named owner before anyone leaves.
The 15% uplift on your proposal has no derivation, so it cannot be defended or released. Price the risks you logged instead, and let the percentage fall out.
Korrel turns briefs into structured proposals, then tracks what each job actually costs.
START FREE TRIAL