The Dutch Consultancy Squeeze and the 7 Metrics That Matter Now

The 9.8% number and what it hides

SPI Research publishes a benchmark every year. For Dutch IT professional services firms, the 2025 number came in at 9.8% average EBITDA. Lowest in five years.

That figure covers the period before Wet DBA enforcement escalated. Before EU AI Act high-risk obligations went live. Before most firms started absorbing the real cost of converting their ZZP base to FTE. In other words, 9.8% is what things looked like at the start of the compression cycle. Not the middle of it.

Underneath that average, the spread is wide. Top-quartile firms are running above 15% EBITDA with revenue per FTE north of EUR 205,000. The bottom quartile sits below EUR 130,000 per FTE and burns cash during bench cycles. Same market, same country, radically different commercial positions.

Billable utilization tells the story faster. The average dropped from 73.2% to 68.9% between 2023 and 2025. Firm-wide productivity (which includes non-billable staff) fell to 57%, three percentage points below 2023. Every percentage point of utilization lost at an average bill rate of EUR 95/hour costs roughly EUR 3,600 per consultant per week in forgone revenue. At 250 FTE, that adds up fast.

The difference between struggling and existential

A struggling consultancy has margin pressure. An existential one has a structural defect in its commercial model that market recovery alone will not fix.

The distinction matters because the response is different. Margin pressure responds to pipeline discipline and rate management. A structural defect, like ZZP dependency above 40% in a Wet DBA enforcement cycle — Wet DBA is now a GTM problem, does not resolve through better sales execution. It requires a model change. Model changes take 12 to 18 months to show up in the P&L, which means the window for starting is already narrower than most boards think.

The seven metrics below follow from that logic. They are not performance indicators. They are structural diagnostics. A firm can have strong revenue growth and still carry three of these at danger levels.

The 7 existential data points

1. Billable utilization below 65%

The sector average is 68.9%. Top performers run above 75%. Below 65%, the firm cannot cover its fixed cost base during normal operating conditions. A bench week at EUR 95/hour bill rate loses approximately EUR 3,600 in gross revenue per consultant. At a 200-person firm with 65% utilization, that's roughly 70 consultants on the bench at any given time.

The number becomes existential when it persists for two or more consecutive quarters. Short dips happen. Structural underutilization is a different animal. It usually signals a sales capacity bottleneck, a skills mismatch between the bench and market demand, or client concentration risk where losing one account creates a utilization crater.

2. ZZP share above 40%

Before Wet DBA enforcement, a high ZZP share was a margin advantage. Flexible capacity, no bench cost, variable cost base. The math flipped in 2025.

Converting ZZP to FTE raises the fully loaded cost by 30 to 40 percent. At an average salary of EUR 85,000, that is not rounding error. It shows up as a gross margin hit in the first quarter after conversion and takes 6 to 9 months to absorb through rate adjustments, if the market absorbs them at all.

The ZZP Nederland survey of 3,100+ respondents found that 59% of ICT ZZP'ers report negative consequences from Wet DBA enforcement. 14% have already lost an assignment. 39% don't know if they'll still be working as ZZP in 2026. HeadFirst reports approximately 25% of high-skilled ZZP'ers have already lost assignments due to Wet DBA anticipation.

Above 40% ZZP share, the conversion cost is large enough to push EBITDA negative for one to two quarters during the transition. Above 60%, it can trigger covenant breaches on PE-backed firms' credit facilities.

3. Top-3 client concentration above 50%

Concentration always looks safe until it is not. A consultancy with 55% of revenue from three banking clients, say ING, ABN, Rabobank, has a diversified-sounding book right up until one of those banks consolidates its vendor panel. In Dutch financial services that happens every 2 to 3 years.

Above 50% concentration in the top three, a single client loss creates a utilization gap that cannot be filled through normal pipeline velocity. The recovery period is typically 6 to 9 months, during which the firm burns cash and often loses its best people to competitors who have the project flow to keep them busy.

4. Partner-tier downgrade

Microsoft Solutions Partner designations, Databricks Elite, AWS Premier. These are not marketing badges anymore. They are de facto RFP qualifiers at ING, ABN, Ahold, and Dutch public sector.

Legacy Microsoft Gold/Silver ended 22 January 2025. The 70-point Partner Capability Score now governs everything. Losing a Solutions Partner designation reduces win rates by an estimated 20 to 40 percent on competitive RFPs. Recovery takes 12 to 18 months. A year of reduced pipeline velocity from a single tier change.

The competitive position is stark: only Valcon holds Databricks Elite in the Netherlands. AWS Premier in NL is limited to Xebia, Cloudar, Devoteam, and Capgemini. When one of these firms loses a Champion or MVP, it's visible on LinkedIn weeks before the directory updates.

5. Annual attrition above 20%

At a 250-FTE firm with an attrition rate above 20%, the hidden churn cost runs to approximately EUR 7.4 million per year. That includes recruitment, onboarding, lost billable hours during ramp-up, and the knowledge drain that reduces delivery quality on long-running engagements.

Attrition above 20% in a tight labor market (71,100 open technical vacancies in Q3 2025) means the firm cannot replace fast enough to maintain delivery capacity. It becomes self-reinforcing: remaining staff absorb the overload, burnout rises, more people leave.

6. EU AI Act readiness gap

The provider/deployer distinction in Article 25 of the EU AI Act is not a legal technicality. A consultancy fine-tuning models, doing RAG, or writing meta-prompts on a third-party model for a client is a provider under the regulation. Not an integrator. Not a deployer. Penalties reach EUR 35 million or 7% of global turnover.

Article 4 AI literacy obligations are already enforceable since February 2025. Annex III high-risk duties take effect August 2026. The Digital Omnibus target of 28 April 2026 may defer some requirements to December 2027, but cannot be treated as deferred until the Official Journal publishes on 2 August 2026.

Detection logic: no AI Act content on the firm's website, no ISO 42001 certification, no dedicated AI governance lead, absence from Consultancy.nl AI rankings. These are negative signals that correlate strongly with unpreparedness.

7. KvK filing delinquency

Dutch companies must file annual accounts with the KvK (Kamer van Koophandel) within specific deadlines. For BVs, the publication deadline is typically 12 months after the fiscal year-end. Filing late is not just a penalty. It is a leading indicator of operational distress.

When a firm that has always filed on time starts filing late, something has changed. Usually it's a combination of: the CFO is managing a cash crunch rather than closing the books, the auditor has flagged issues that require resolution before signing, or the firm is in pre-transaction due diligence and the accounts are being restructured.

Combined with any two of the other six EDPs, a KvK filing delay is a strong negative signal.

The threshold table

Each metric has three zones. The thresholds are not arbitrary. They come from the structural dynamics of the Dutch IT professional services market at its current compression point.

Three or more metrics in the danger zone simultaneously is the pattern we see in firms that end up in distressed M&A within 18 months. Two in danger plus two in warning is the leading edge of the same trajectory.

The valley of death between 50 and 300 FTE

Dutch consultancies tend to fail at two transition points. The first is around 20 to 50 FTE, where the founder can no longer personally sell every deal and the first real bench problem appears. The second is between 150 and 250 FTE, where the boutique identity dissolves but scale economics haven't arrived — valley of death between 50 and 300 FTE.

The firms stuck in the 50-to-300-FTE range in 2026 face compression from three directions at once. GenAI is eating the junior delivery base: the routine implementation, migration, and testing work that used to justify junior billing rates. Hyperscalers are crowding out generalists on enterprise accounts through co-sell motions. Specialized boutiques are outcompeting on domain depth in AI governance, cloud-native architecture, data mesh.

Revenue per FTE is the clearest diagnostic. Below EUR 130,000 is the danger zone. The average across the sector sits at EUR 140,000 to 147,000. Top quartile runs at EUR 205,000. That gap between average and top quartile tells you size alone is not the problem. Commercial positioning is.

The failure patterns are visible in recent history. Carlyle exited Dept to a new PE sponsor after 20+ acquisitions, a textbook case of acquisition indigestion. Centric bled clients between 2022 and 2024. Be Informed, Nolan Norton, BoerCroon all showed the same concentration-then-collapse pattern.

What you can see from the outside

Every one of these seven metrics is detectable without inside access to the company. That's what makes them useful for GTM targeting, competitive intelligence, and investment screening.

• Billable utilization: LinkedIn headcount trajectory vs. reported revenue growth. A firm adding consultants without corresponding revenue growth has a utilization problem.
• ZZP share: Sudden burst of permanent job postings in a 90-day window, especially if the firm's LinkedIn headcount was previously flat. Job ads shifting from "ZZP/freelancer" to "medewerker" language.
• Client concentration: Case studies on the website, named client logos, conference speaking topics. A firm whose last 5 case studies all name the same 2 clients has a concentration problem.
• Partner tier: Microsoft Partner Finder (public), partners.amazonaws.com, databricks.com/partners. Named MVPs/Champions on LinkedIn, departures in clusters precede tier downgrades.
• Attrition: Glassdoor 90-day rating deltas, LinkedIn "open to work" clusters from the same employer, Glassdoor review velocity increasing.
• AI Act readiness: Website content audit: does the firm mention AI Act, Article 25, conformity assessment, AI governance anywhere? No? That's a signal.
• KvK filing: Publicly available. Check 10 months after fiscal year-end. Late filing is visible to anyone.

Weekly diffs on partner directories, monthly checks on KvK filing deadlines, quarterly Glassdoor reviews. That's the monitoring cadence.

What this means commercially

If you sell into Dutch IT consultancies (transformation services, managed services, tooling, advisory), these seven metrics change your targeting logic entirely.

A firm with ZZP share above 40% and a Wet DBA enforcement letter on the desk is not evaluating your product on features. They are evaluating whether you can help them survive a structural transition. That is a CFO-level conversation, not a CIO-level one. The pain is financial and HR-legal, not technical.

A firm that just lost its Databricks Select badge and has two MVPs showing "open to work" on LinkedIn will not respond to a generic outbound sequence about data platform modernization. But they will respond to someone who understands what that badge loss means for their ING pipeline this quarter.

These are not predictions. They are observable now. The firms that read these metrics correctly and act on them, whether as operators fixing their own position or as commercial partners arriving with relevant solutions, have a window. The firms that do not will learn the same lessons through their P&L statements, 6 to 12 months later.

RPT

Flagship Research

The Dutch IT/Data/AI Consultancy Squeeze: Why 2026 is Different

→