The first month we ran live project margin on the studio, the picture was uglier than the year-end accounts had ever shown. A 6-week retainer that looked healthy on Xero was actually running at 4% gross by week three. We had been quietly subsidising it for two months without knowing. The accountant's year-end view said we were profitable. The project-level view said two of our top six clients were costing us money. Both were true.
Project cost management software is what closes that gap. It tracks every cost tied to a single project (loaded labour, freelancer hours, software allocations, travel) and compares it to revenue in real time. The shift it represents is from looking at the agency as one company to looking at it as a portfolio of projects, each with its own P&L.
What does project cost management software actually track?
Five cost categories belong on every project. Most accounting software collapses them into the agency's overall expense lines. Project cost management software keeps them separated and tagged to the specific job they belong to.
Loaded labour cost. The single biggest line, the one most often miscounted. Loaded cost is salary plus employer taxes plus benefits plus a share of overhead, expressed per hour. A designer on a $90,000 salary with full benefits costs the agency around $60-70 an hour to put on a project, not their billable rate. Confusing the two flatters margin by 30-40%.
Expenses. Travel, software paid through to the client, hardware, supplies. These run through the agency's accounting layer for tax reasons but need a project tag for margin purposes.
Freelancer pay. Pass-through cost when the agency invoices the work straight through. A real cost when the agency absorbs it. Project cost management software has to distinguish the two cases. Most agencies handle them as a single line in their accounting tool, which is why so many retainers quietly bleed.
Software allocations. Tools used on the project (a stock library, an analytics seat, a paid font) get billed at the agency level. Their project-level cost is rarely visible.
Pass-through costs. Travel, ad spend, third-party fees the agency invoices. These count as revenue and as cost, with the net being the agency's actual margin. Pass-through revenue at 100% of cost looks like growth in the accounts. It is not.
How is project cost different from project revenue?
This is where most agencies get confused. Revenue is what gets quoted, contracted and invoiced. Cost is what the agency actually pays out to do the work. The two are not symmetrical.
Take a 6-week retainer at $15,000. The revenue side is clean: $15,000 invoiced monthly or as agreed. The cost side has five lines. A senior developer at $70/hour loaded cost, scheduled 12 hours a week for 6 weeks: $5,040. A junior designer at $45/hour, 8 hours a week: $2,160. A project lead at $95/hour, 4 hours a week: $2,280. A freelance copywriter as an absorbed cost: $1,500. Software allocations and an overhead share: $700. Total: $11,680.
The gross margin is $3,320 or 22%. That is the version most agencies stop at. The full margin has to subtract the agency's share of unbilled time (sales meetings, internal admin) attributable to keeping the account, which often takes the figure to 12-15% net.
The mistake is to treat the billable rate as the cost. A senior developer billed at $150 an hour does not cost $150 an hour. The agency keeps the gap, which is the gross margin on that line. Confusing the two is why so many "20% margin" agencies discover at year-end that the real number is closer to 5%.
Why most agencies discover overruns too late
Three reasons. They are operational, not analytical.
One. The monthly close cycle. Costs land in the accounting tool when invoices and payroll run, which is once a month. By the time costs are visible on a P&L, the project may already be three weeks past where the overrun started. Year-end financials see the year as a whole, not the project pattern that produced it.
Two. Time entry lag. Developers and designers log hours at the end of the week. Or the end of the sprint. Or never. The Mavenlink "State of the Services Economy" study put the average agency time-entry lag at over a week. A project running 1.4x its budgeted hours in week two is invisible until week four, when the time finally enters the system.
Three. Reconciliation overhead. To answer "what is the margin on this project right now" requires combining time logs, scheduled work, expense receipts, the original budget and the billable rate, in real time. In most agency stacks that is a 90-minute job per project per check, which means it does not get done weekly. It gets done at month-end, on the projects that look obviously wrong.
The cumulative cost is large. The Agency Management Institute puts utilisation slippage among the top three drivers of margin loss in mid-size agencies, alongside scope creep and inefficient pricing. Each of those is a question you have to ask weekly to catch. None of them survives a monthly review cadence.
What "real-time margin" actually means for a services business
Real-time margin is the project's current cost (loaded labour plus expenses plus allocated overhead) measured against its contracted or invoiced revenue, refreshed every time something changes. Time entered. Expense logged. Scope amended. Invoice sent.
The technical requirement is one data model. Time has to share the database with the rate card. The rate card has to share the database with the contract. The contract has to share the database with the invoice. The moment any of these sit in separate tools, real time becomes "real time as of the last successful sync", which is structurally never quite now.
Real time matters because the lever it gives you is intervention. A project that is 12% over its labour budget at week two can be rescoped, restaffed or have a conversation with the client. A project that is 35% over at the year-end review can only be regretted. The two are the same data, observed at different points in the cycle. The earlier point is the one where you still have options.
What to evaluate when choosing a tool
Six things separate project cost management software that actually works in an agency from software that calls itself the same thing.
One. Native time tracking, not integration. If time has to sync from a separate tool, the margin number you see is at best yesterday's. Agencies that run on a cost management layer with native time tracking see margin update inside an hour of a developer logging the work.
Two. Loaded cost handling, not just billable rate. The platform needs to know what each person actually costs the agency per hour, with employer tax and overhead share built in. Tools that only handle billable rates can show revenue per project but not the margin underneath.
Three. Retainer model support. Retainers are not projects. They have a recurring commercial structure, a cap, a rollover policy and a margin target across the engagement, not within a single deliverable. Tools that treat them as recurring projects misread retainer health at every quarter.
Four. Forecasting against actuals. Knowing where margin is is half the value. Knowing where it is going (based on scheduled work, signed scope and current burn) is the other half. Without forecasting you have a rear-view mirror, which is useful but insufficient.
Five. Multi-currency, even if you do not need it yet. The agencies we have worked with mostly added their first overseas client within 18 months of switching. Adding multi-currency to a cost-tracking layer mid-flight is painful.
Six. Pricing that does not punish growth. Per-seat pricing on a cost tracking platform reaches absurd levels at 30-50 person agencies. Flat-fee structures (or team-based pricing tiers) keep the cost predictable. Our own model is $50/month for teams up to 25 people, $100/month above.
Turning around a losing retainer: a worked example
Real numbers from a real engagement, sanitised. A 12-week retainer at $24,000 ($8,000/month for 3 months). Original scope: 4 channels managed, weekly status, one monthly report. Original budget: 60 senior hours, 80 junior hours, 24 project lead hours. Forecast margin at the time of signing: 30%.
Week two state. The junior designer is logging 12 hours a week instead of the 6.7 the budget assumed. At $45/hour loaded that is $540 over the weekly budget. The project's live margin has dropped from a forecast 30% to 22% in two weeks. In a month-end-close system, this is invisible until week eight.
The flag. The system raises the drift on Friday morning of week two. The flag carries context: the overrun is in the junior designer line, the most likely cause is unplanned revisions, the projected end-state margin if the trend continues is 14%.
The conversation. The project lead has real options. Discuss the scope expansion with the client and bill for the additional rounds. Restaff (have a faster senior pick up some junior work). Rescope (drop one channel temporarily). Accept the reduced margin if the client warrants it. Each option is a concrete choice with a known cost.
The outcome. The project lead has a 20-minute conversation with the client. The fourth channel turned out to be the source of the overrun. The client agreed to either drop the channel or pay for the additional time. They chose to pay. The retainer ended at 28% margin, not 14%.
The counterfactual. In a month-end-close system, the project lead would have learned about the overrun in week eight, after the cost had compounded for six weeks. The conversation with the client would have been a difficult re-quote, not a clean renegotiation. The retainer would have ended in the 9-12% band, with the relationship strained. The lever real-time margin gives you is intervention. The earlier the intervention, the cleaner the conversation.
From accounting to operations
Most agencies have a project cost management problem, not an accounting problem. The accountant can tell you what the agency made last year. They cannot tell you which project made it, which retainer ate it or which client is silently below margin this week. Those are operational questions that need an operational tool. Project profitability at the project level is what the category is about. The deciding factor is whether the system updates in real time or at the close.
Key takeaways
- Project cost management software tracks loaded cost per project against revenue, in real time.
- Loaded cost is salary plus employer tax plus overhead allocation, not billable rate. Confusing the two flatters margin by 30-40%.
- Most agencies discover overruns at month-end close, when the money is already lost.
- Real-time margin is what lets a project lead pull the plug at week three instead of week eight.
- A spreadsheet works at 3 projects. By 15 it lies, because the data is never current.
Frequently asked questions
Is project cost management software the same as job costing?+
Closely related, with a different emphasis. Job costing comes from accounting and focuses on what each completed job cost, mostly for reporting. Project cost management software is operational. It tracks cost as it happens, against revenue, in real time, so the project lead can intervene mid-project. Both end up with similar numbers at the close, but only one of them helps you change the outcome before the close.
Do agencies really need project cost management software, or is QuickBooks enough?+
QuickBooks tells you what the agency made last quarter. It does not tell you which project made it, which retainer is below margin this week, or which client is silently unprofitable. Most agencies need both layers: an accounting tool for statutory reporting and an operational layer for real-time project margin. Trying to run project costing inside QuickBooks alone is possible at 5 projects, painful at 15 and impossible at 50.
How do I calculate loaded cost?+
Take the person's annual salary, add employer taxes and benefits (often 15-25% on top), then add a share of agency overhead (rent, software, admin staff) per billable hour. The result is the hourly cost of putting that person on a project. A $90,000 designer is usually $60-70 an hour loaded. Most agencies that try this exercise discover their loaded cost is higher than they had assumed, which means their real margin is lower.
What's a healthy gross margin per project?+
Agency advisors converge on a 50-60% gross margin band for project delivery work, with the higher end on retainers and the lower end on heavily-staffed campaigns. Below 40% gross is fragile. Above 65% gross usually means the agency is under-staffing the work or charging for outcomes it has not yet delivered. The net margin after overhead lands in the well-known 15-25% range.
Can a spreadsheet do project cost management?+
Up to about 3 concurrent projects, yes. Above that the spreadsheet has the same problem every other agency tool has: the data is never current. Time gets logged in one tool. The spreadsheet gets updated weekly at best. By then the project has drifted. Spreadsheets work for analysis after the fact. They do not work for the real-time intervention that prevents the overrun.
How is project cost management different from time tracking?+
Time tracking captures hours. Project cost management captures hours plus loaded cost plus expenses plus overhead, then compares the total to revenue. A time tracker tells you how long the work took. A project cost system tells you what the work cost and whether it made money. Many agencies have a time tracker and call it cost management. They are different layers.
What's the trigger to upgrade from a spreadsheet?+
Three reliable signals. First, you spend Sunday afternoons reconciling the cost spreadsheet against time logs. Second, project leads stop checking margin until month-end because it takes too long to assemble. Third, you discover an overrun at year-end on a project that "felt fine" in the moment. Any one of those is the trigger. All three together mean the spreadsheet has cost you more than the software replacing it would.
Sources
- What is a reasonable agency profit margin? - Agency Management Institute
- The Role of Profit in a Creative Enterprise - David C. Baker (2Bobs)
- State of the Services Economy - Kantata / Mavenlink
- Pulse of the Profession 2024: Project Success in Disruptive Times - Project Management Institute
