Which Project Cost You Money? Most Service Businesses Only Find Out in December

A 20-person consulting firm in Stuttgart reviews its 11 projects from the past year in December. The most important engagement — five months, €320,000 revenue — shows an 18% contribution margin in the bookkeeping. When the project team reconstructs the actual hours and third-party costs, the real margin is 4%. The gap: 160 unbilled hours of change requests, two freelancers deployed outside the original bid framework, one additional revision round that nobody charged for.

The client was satisfied. The project was delivered on time. And the firm subsidized the client with roughly €45,000 from its own EBITDA.

This discovery in December is worthless. The project was finished in August.

This is the structural root problem of project-based service businesses in the mid-market: project profitability only becomes visible when it's too late to change anything. Not because the data doesn't exist — time tracking, third-party costs, and bid calculations exist somewhere in almost every firm. But because they're never brought together into a shared picture.

The problem has two sources that are each individually reasonable — but together create a blind spot.

**Bookkeeping aggregates by account, not by project.** DATEV and every BWA show: personnel costs in the quarter €X. Third-party services €Y. Total revenue €Z. What they don't show: which part of these personnel costs belongs to which project? Which freelancer invoices are attributable to which engagement? The logic of bookkeeping is chart-of-accounts-based — not project-based. That's not wrong for tax purposes. It's blind to project profitability.

**The project management tool doesn't know costs.** Jira, Asana, Monday, or the self-built spreadsheet know which tasks were completed. Some know booked hours. But they don't know internal fully-loaded rates, third-party costs, or deviation from the original bid. They're progress tools, not margin measurement tools.

**Time tracking is an island.** In many consultancies and agencies there's a time tracking system — MOCO, Harvest, TimeMaster, or plain Excel. But this data rarely flows automatically into a project cost accounting system. It's used for billing (when at all) or internal capacity comparison. Not for running margin calculations.

Three data islands. None of them alone answers the only question that really matters: how much margin does this project have left?

In practice, project margin losses in mid-market service businesses arise repeatedly from the same three patterns:

**1. Scope Creep Without Billing** The client asks: 'Can you just quickly…' The team says yes — reasonable in isolation, fatal in aggregate. A typical consulting or agency project has 15–30% additional effort from scope creep that is never billed. At a realized hourly rate of €140 and a 500-hour project, 20% untracked additional effort is roughly €14,000 in destroyed margin — per project, systematically and unnoticed.

The real cause isn't a lack of courage to bill additionally. It's lack of visibility: when nobody knows that 80% of the project budget is already consumed after 60% of the project timeline, there's no trigger for a decision. The decision then comes in December — as a finding, not as an option to act.

**2. Resource Mix Variance** The bid assumed 70% junior hours (internal cost €85) and 30% senior hours (€165). In execution it was reversed — because seniors didn't delegate, juniors weren't available, or because the client implicitly demanded senior contact. The cost difference between a junior day and a senior day is €640. Over 50 days of variance, that's €32,000 in unexpected cost overrun — invisible in the bookkeeping, because everything is summed under 'personnel costs.'

**3. WIP Financing Costs** Project-based companies deliver work before they send invoices. Costs run from day one. Revenue only arrives with billing — sometimes only after final project sign-off. On a 4-month project with €50,000/month in personnel costs, the firm pre-finances up to €200,000 in work in progress.

According to the Deloitte Working Capital Report 2025, the average cash-to-cash cycle in Germany in 2024 was 65 days. At overdraft interest rates of 8–10%, pre-financing €200,000 over 65 days costs roughly €2,900–3,600 in financing costs — which nobody factors into project calculations and which therefore directly eat margin.

Real-time here doesn't mean: a dashboard that refreshes every five minutes. It means: you know the current margin status of every active project at any point — not only at project end.

This requires connecting three data sources that exist in isolation in most companies:

**Data Source 1 — Time Tracking:** Every booked hour is assigned to a person, a project, and an activity category. The person has an internal fully-loaded cost rate (not the billing rate, but the real cost). Hours × fully-loaded rate = running internal personnel costs per project.

**Data Source 2 — Third-Party Costs:** Freelancer invoices, tool costs, travel expenses, external service providers — all are assigned directly to the respective project at the time of receipt, not just booked to a general expense account.

**Data Source 3 — Project Budget from the Bid:** The proposal contains: planned hours by seniority level, planned rates, planned third-party costs. This budget is the reference point for every variance analysis.

From these three sources, a deterministic real-time margin formula emerges:

**Projected Final Margin** = Agreed Price − (Actual Costs to Date + Remaining Budgeted Costs)

When this projected final margin falls more than 15–20% below the bid margin, a decision is triggered: bill the additional effort, reduce scope, or consciously accept the margin reduction.

The key word is 'consciously' — not 'accidentally discovered in December.'

This is not an AI system. It's not a machine learning forecast. It's deterministic arithmetic based on data the company already has.

Implementing real-time project margin controlling is not a year-long project. It's a structured 5–7 week initiative built on existing data.

A 20-person service business with €3M annual revenue typically runs 15–25 projects per year. If each project has an average of 20% untracked scope creep effort (a conservative estimate based on industry experience), and the average project revenue is €120,000, then roughly €360,000–600,000 in additional effort is not billed annually. Even if half of that is legitimately written off as client goodwill: €180,000–300,000 in EBITDA potential is left on the table every year.

Add resource mix variances: in a consultancy with €3M revenue delivering 30–40 consultant months per year, an average 15% deviation on the junior/senior mix can mean €40,000–80,000 in additional cost burden — completely invisible to all involved.

And WIP financing costs: with an average working capital cycle of 65 days (Deloitte 2025) and revenue of €3M, annual WIP financing costs amount to roughly €43,000 at 8% overdraft interest — costs that nobody budgets for and that directly eat EBITDA.

In total: a typical 20-person service business with €3M revenue leaves €200,000–400,000 in EBITDA potential on the table annually through missing project transparency. Not through poor work. Not through wrong prices. Through missing visibility into what is already happening — and late enough that nothing can be done about it.

This is not an accounting problem. It's a steering problem. And steering problems are solvable — if you stop waiting for December.