Benchmarking retention and expansion metrics against sector norms

A mid-size management consultancy's board reviews its annual scorecard: 85% client retention, flat for three years. The CMO calls it a strength. The CFO calls it stagnation. Both read the same number and reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → opposite conclusions, because neither has anything credible to read it against. Benchmarking settles that argument, and it is harder in professional services than in almost any other sector: the comparable data is thin, mostly vendor-published, and rarely measured the way your firm measures.

Two numbers, because one hides the other

In subscription software, retention is binary: the client renews or churns. In professional services a client can stay and shrink to nothing, or close one engagement and open another with a different partner in the same firm. Any benchmark worth reading therefore pairs two figures.

• Client retention rate (CRR): the share of clients from last period still active this period.
• Net revenue retention (NRR): whether revenue from the retained base grew, shrank or held flat.

CRR = (Clients at end of period − New clients acquired) / Clients at start of period × 100
NRR = (Starting revenue + Expansion − Contraction − Churned revenue) / Starting revenue × 100

85% CRR with 70% NRRNRRNet Revenue Retention measures the percentage of recurring revenue retained and grown from existing customers over a period, including upsell and expansion, net of downgrades and churn.View full definition → means retention looks healthy while revenue quietly erodes. 85% CRR with 115% NRR means a handful of retained clients are buying much more. Report one without the other and the board benchmarks the wrong thing.

Where the credible numbers actually come from

There is no authoritative global census of professional services retention. What exists is a short list of recurring studies, each with a known bias:

• The Hinge Research Institute's annual High Growth Study surveys several hundred firms across accounting and finance, architecture and engineering, consulting, legal and technology services, and splits them into high-growth (roughly 20% or more annual revenue growth) and no-growth cohorts. That cohort split is the useful part: comparing yourself to the high-growth quartile in your own sub-sector tells you far more than an all-firm average. Hinge sells marketing strategy and research to professional services firms, so its samples skew toward firms that already invest in marketing.
• Deltek's Clarity study covers architecture and engineering firms, drawing on both survey responses and the firm's own project data. It has for years put A&E win rates below half of proposals submitted, with repeat clients supplying the majority of fee income. Deltek sells project-based ERPERPA single integrated software backbone that runs core operations: finance, procurement, supply chain, HR and manufacturing on shared data.View full definition → software to the same firms, which is why its benchmarks lean operational (utilisation, backlog, win rate) rather than brand-side.
• Public annual reviews from the large accounting and advisory networks, Deloitte among them, break revenue down by service line and region. Deloitte reports global revenue above $65 billion split across audit and assurance, consulting, tax and legal, and risk and financial advisory. That structure is a usable reference for service-line mix. It says nothing about client retention, which none of the networks disclose.
• Software benchmarks such as OpenView SaaS Benchmarks (annual survey, self-reported) are widely quoted in advisory pitches: 90%+ NRR healthy, 120%+ best in class. Those numbers come from renewal-based economics and travel badly into project work, where "churn" often just means the project finished on schedule.

Two failure modes kill more benchmarking exercises than bad data does. The first is definitional drift: if your CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition → counts three divisions of one group as three clients and the benchmark firm counts them as one, your retention rate is structurally different from theirs and no adjustment will reconcile them. The second is the measurement window. Measure retention on a rolling 12 months in a practice where the typical client returns every 18 to 24 months and you will manufacture churn that does not exist. Cohort-based retention over a trailing 24 or 36 months is the honest version.

Sub-sector norms, and why mixing them is the usual error

Treat all of these as estimates and ranges, not published fact:

• Management consulting with recurring advisory relationships: annual client retention commonly quoted in the 75 to 90% band, with the top end driven by multi-year transformation programmes rather than by marketing.
• Corporate and M&A legal practices: often cited around 80 to 90%. Purely transactional or one-off matter work (single-deal mandates, some litigation defence) can sit at half that, and it is not a performance problem, it is the shape of the demand.
• Audit and accounting: frequently above 90%. Switching auditors carries real friction and, for US listed issuers, disclosure obligations to the SEC.
• Marketing and creative agencies: volatile, with average tenure often quoted at two to three years, implying annual retention nearer 70 to 80%.

So 85% for a consultancy sits mid-pack. Benchmark that same 85% against the audit norm and you will conclude the firm is failing; against the agency norm you will conclude it is excellent. Both conclusions are artefacts of the comparison set.

Cross-sell and wallet share

Retained accounts in this sector grow mainly through cross-selling a second service line, not through price. That is the local equivalent of SaaS expansion revenue, and it has a hard regulatory ceiling that software firms never encounter: under post-Sarbanes-Oxley independence rules, an audit firm cannot sell an audit client whatever it likes. A multi-service revenue sharerevenue shareThe percentage of total industry sales your company captures in a given period. It measures competitive position relative to rivals in a defined market.View full definition → that is normal across a Big Four network as a whole is unreachable inside its audit book. Benchmark cross-sell by service line, or you will set a target the compliance function is obliged to block.

Boutiques face the opposite constraint. A single-service consultancy with three adjacent offerings will post low cross-sell rates because it has almost nothing to cross-sell, and pushing that number up means launching a service line, which is a capacity decision, not a marketing one.

Worked example. A consultancy starts the year with 100 clients and $50 million in revenue ($500,000 average).

• 85 clients renew (85% CRR).
• 20 of those buy a second service line, adding $150,000 each: $3 million expansion.
• 15 contract by an average $100,000: $1.5 million lost.
• The 15 departing clients took $7.5 million with them.

Retained base revenue = 85 × $500,000 = $42,500,000
+ Expansion = $3,000,000
− Contraction = $1,500,000
= Net retained revenue = $44,000,000

NRR = $44,000,000 / $50,000,000 = 88%

85% CRR, 88% NRR. The clients who stayed are, on net, spending slightly less than last year. Against firms that hold NRR at or above 100% by covering contraction with expansion, that is a mild warning, not because clients leave but because the ones who stay do not buy more.

Acquisition costAcquisition costCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition → and lifetime valuelifetime valueLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition → connect here without needing restating: the loaded pursuit-cost figure the acquisition lesson builds, and the lumpy-spend valuation its own lesson models, both take retention and expansion as inputs. The second-order consequence is what boards miss. If a new client needs 18 months of fees to repay its pursuit cost, an NRR of 88% means each cohort repays more slowly than the last, so the payback window slides even while headline retention looks stable. Strong cross-sell can carry mediocre retention. Flat wallet share cannot outrun heavy pursuit costs indefinitely.

Reading the scorecard like an analyst

Back to the board. The CMO is right that 85% is defensible for a consultancy. The CFO is right that 88% NRR is a caution flag. Three checks before anyone acts on either:

• Strip out zombie retention. A client sitting on a framework or panel with zero fees billed in 12 months is retained on paper and worthless in cash. Set a minimum fee threshold for "active" and rerun the number; firms doing this for the first time typically lose several points of headline retention.
• Check who moved the average. If two accounts supplied most of the expansion, the NRR figure describes those two accounts, not the firm.
• Do not re-baseline every year against whichever study flatters you. Pick one source per sub-sector, note its sample bias, and hold it for three cycles so the trend means something.

The fix for an 88% NRR is rarely more logos. It is account planning that names the second service line per client, wallet share tracked per account rather than logo counts, and a cross-sell target owned jointly by marketing and the practice leaders who have to deliver the work.

🎬 [VIDEO: "Net Revenue Retention Explained" - https://www.youtube.com/results?search_query=net+revenue+retention+explained - a walkthrough of how NRR is calculated and why it matters more than logo retention alone, applicable directly to professional services accounts]

Key Takeaways

• Retention alone is not diagnostic here. Read CRR and NRR together, or the board benchmarks a number that hides the trend.
• Benchmark data in this sector is vendor-published and self-selected. Hinge Research Institute's cohort splits and Deltek's Clarity study are the most usable starting points, and both come from firms that sell to the audience they survey.
• Norms diverge by sub-sector far more than by firm quality: above 90% in audit, 75 to 90% in advisory consulting, closer to 70 to 80% in agencies, lower still for one-off transactional mandates.
• Measurement definitions break comparisons before the data does. Client counting rules, the fee threshold for "active", and a 12-month window over an 18-month buying rhythm each move retention by several points.
• Cross-sell targets need a service-line view. Independence rules cap what an audit client can be sold, and a boutique's low cross-sell rate reflects its service portfolio, not its account management.