Agency utilization and realization: the two metrics that explain margin.

01 Utilization, defined cleanly

Utilization is billable hours divided by available billable hours. The numerator is the hours that get coded to client work in the PSA. The denominator is the contracted work-hours net of PTO, statutory holidays, and where the agency chooses, internal time on RFPs and new-business pitches. The number is the same arithmetic at the individual, team, and agency level. What varies is the denominator definition and the band considered healthy.

The benchmark bands have converged across the published agency-ops literature over the last two years. Swydo and Parakeeto (2025) put the role-tier targets at 80–85% for junior and execution-focused FTEs, 70–80% for mid-level and lead roles, 50–60% for senior and director-level, and 40–50% for new-business and relationship-led roles. Ravetree (2026) sits in the same band — 75–80% for junior team members, 60–80% for production staff broadly, and 30–40% for directors. Harvest's 2024 marketing-agency calculator splits delivery roles at 75–90%, administrative at 50–75%, and agency-wide at 50–70%. Productive.io flags 70–80% as ideal at the individual level and >90% as a burnout signal. The shape is the same across all four: producers in the 75–85% band, mid-level in the 70–80% band, senior and director materially lower at 50–65%, and executives below that.

The agency-wide annual rollup — the number that goes into a board pack — sits much lower than any of the individual targets because the denominator includes holidays, sick days, internal time, and the role-mix drag from senior and executive headcount. Swydo/Parakeeto put this at 50–60% typical. Ravetree puts it at 50–60% annual net. Productive.io quotes HubSpot at 60% for the average advertising agency. The 2019 historical average for marketing agencies, per Harvest, ran at 53%. The takeaway: the headline agency-wide number that looks alarmingly low — 55%, 58%, 62% — is almost always benchmark-normal. What matters is the individual-tier distribution underneath it.

Most agency operating reviews I walk into are tracking utilization correctly. The PSA computes it automatically, the COO or Head of Delivery owns the metric, and the weekly cadence is in place. The instrumentation is there. What is missing is the second half of the picture.

02 Realization, the metric most agencies miss

Realization is billed rate divided by standard rate, computed on collected revenue. The numerator is what the client actually paid against the hours delivered. The denominator is the rate card the agency would have invoiced at if every hour cleared at the published rate, against the agreed scope, with no write-downs and no scope creep absorbed into engagement margin. Agencies almost never track this with the same discipline they bring to utilization, and the reasons are structural rather than ideological.

Three reasons surface in every engagement. First, the rate card is itself fluid — many mid-market agencies do not maintain a formal published rate card and instead operate with implicit blended rates negotiated on a per-engagement basis, which makes the denominator unstable. Second, the billed rate varies by client tier, engagement scope, and discount governance, and the variance is not always logged in the PSA in a way that aggregates cleanly. Third, realization is only computable on collected revenue, which is 30–90 days behind hours logged, so the metric lags utilization in a way that frustrates real-time visibility. The result is that realization sits in the CFO's domain rather than the COO's, gets reviewed monthly at best, and almost never appears in the same dashboard as utilization.

When realization is rebuilt against actuals across the engagements in our sample, the cohort lands in a tight and recognisable band. The median runs at 78%. The interquartile range sits between 72% and 86%. The bottom-decile clients — the chronic over-service accounts where scope creep is absorbed rather than billed — drop to 60–70%. The published 2024–2026 benchmarks are consistent with this picture: Monograph's 2026 A&E benchmarks put baseline firm realization at 96%, median at 95%, bottom quartile at 83–84%, and top quartile at 107% (where value-based pricing yields effective rates above the rate card). Per agency type, the benchmark-aligned target bands published by Parakeeto and Swydo run 92–98% for digital agencies, 90–97% for creative shops, 88–95% for PR firms, and 90–97% for integrated marketing agencies. The gap between our cohort median of 78% and the benchmark target band of 90–95% is the leakage, and it converts directly into operating-margin compression line by line.

The 12-to-17-point gap between the cohort median and the benchmark target is not a measurement artefact. It is the cumulative effect of seven mechanisms that compress realization over time, none of which any single PSA dashboard surfaces cleanly on its own. The most consequential of the seven is scope creep handled as informal client accommodation rather than as a structured change-order — the mechanism is named below — but rate-card discipline, estimation rigor, time-tracking lag, billing write-offs, role-mix mismatch, and unstructured client communication all compound. Each is recoverable; the combined recovery is the operating-margin lift the leadership team is leaving on the table.

03 The product of the two — effective revenue rate

The product of utilization and realization is the effective revenue rate. It is the percent of theoretical maximum revenue the agency actually captures from its FTE base, and it is the single most informative operating number for an agency CFO or operating partner. Parakeeto names this the effective billable rate or effective hourly rate. AMI walks it through the 55:25:20 framework — 55% people, 25% overhead, 20% profit — and the effective rate is what determines whether the people line covers itself. Productive.io builds its margin dashboards against it. The metric collapses two independent operating questions into one number that the leadership team can manage to.

The math is straightforward. If utilization runs at 70% and realization runs at 78% — roughly the median position in our 11-engagement cohort — the effective rate is 54.6% of theoretical maximum. If utilization runs at 80% and realization at 95% — the top-quartile published benchmark — the effective rate is 76.0%. The gap between the two positions is 21.4 percentage points on the same FTE base, the same rate card, and the same overhead footprint. That gap is the operating margin gap. On a $100/hr loaded FTE cost and a $200/hr standard rate, the 54.6% effective rate yields contribution of roughly $9 per available FTE hour — below the operating-overhead absorption line for most mid-market agencies. The 76.0% effective rate yields contribution of roughly $52 per available FTE hour, which clears the overhead line and lands the agency in the 20–25% operating-margin band that Parakeeto and Swydo call "high-performing."

The single most useful observation in our cohort work is what happens when both metrics move by five points each. A 75% utilization paired with 88% realization yields a 66.0% effective rate. Moving each by five points to 80% and 93% — both achievable inside three quarters — yields 74.4%. That is an 8.4-point lift on the same FTE base, translating to roughly 8–12 points of contribution-margin lift. The eight-to-twelve point band shows up consistently in published case studies: Productive.io puts GM improvement at 5–12 points from integrated PSA visibility; Kantata at 5–8 points from structured resource planning; Parakeeto's typical engagement delivers 8–15 points of realization recovery. The product of the two is the operating reality.

04 The scope-creep mechanism — where realization actually breaks

Realization compression is rarely a pricing problem. It is almost always a scope-discipline problem. The pattern, named cleanly in Parakeeto's agency-margin work and in the Monograph 2026 A&E benchmarks, runs the same way across every engagement we have rebuilt: client requests an addition mid-engagement (an extra revision round, a new channel, a stakeholder review added late), the request is handled informally in Slack or email rather than logged as a Potential Change Order, the production team treats client satisfaction as the goal and absorbs the work into the engagement, and at billing time the over-budget hours are written off rather than billed because the conversation is now too uncomfortable to have. The hours show up on the timesheet, which keeps utilization healthy. The fees never make it onto the invoice, which is where realization drops.

The published quantification of the mechanism is consistent across sources. Parakeeto's agency-margin analyses (2019–2024) show uncompensated scope creep eroding 5–12 percentage points of realization on fixed-fee projects. Monograph 2026 reports a 4–10 percentage-point realization gap between baseline projects and those run with disciplined change-order management. Construction analogs — Deltek, Trimble Viewpoint, ConstructionCoverage — consistently show unmanaged change orders consuming 30–50% of project margin. Legal benchmarks from LeanLaw and ClearPoint show 3–7% of potential revenue lost in write-downs alone, and bill-time write-offs causing 5–15 points of realization erosion on the worst 20–30% of projects. The pattern transfers cleanly from one project-based professional-services sub-sector to the next.

The remedy is structural, not behavioural

Telling project managers to "be better about scope" does not work. The remedy is a Potential Change Order workflow with clear triggers and clear authority. Trigger thresholds — adopted from the construction discipline and adapted to agency context — fire whenever a scope change adds more than 8–16 hours of work or 5–10% of the budget, whenever the request introduces a new channel or net-new deliverable, or whenever it exceeds the contractual revision count. A PCO ticket logs in the PSA within 24–48 hours of the request with a preliminary hours-and-timeline impact. A formal Change Order proposal lands with revised fee, schedule impact, and explicit assumptions and exclusions. The "no change order, no work" rule applies, with a defined emergency lane for urgent client needs that proceed at risk pending sign-off.

Agencies that institute the discipline see realization recover 5–15 points on projects with high historical scope creep and 3–7 points portfolio-wide inside six to twelve months. The Monograph A&E cohort numbers — 96% baseline realization moving to roughly 100% on comparable projects after change-order discipline — sit at the upper end of what we see in agency rebuilds because A&E firms operate with more contractual rigour to begin with. The mid-market agency starting point is lower and the gain is correspondingly larger. Across our 11-engagement cohort, the median realization lift from change-order discipline alone, taken twelve months after implementation, has been 7 percentage points. That alone is approximately 0.6 turns of operating-margin recovery at typical agency cost structures.

Rate-card discipline, estimation rigor, and write-off controls compound on top of the change-order work. Rate-card review tied to comp and inflation — 5–10% annual adjustments per Parakeeto — recovers 3–6 points of realization in twelve months when paired with discount governance. Historical-actuals-based estimating templates per service line lift project-level realization from 70–80% to 85–95% on recurring project types. Write-off controls with maximum unilateral thresholds plus quarterly client profitability review add another 3–7 points across the portfolio. The full programme lands 8–15 percentage points of realization recovery in the first twelve months for a typical agency starting in the 75–85% range.

05 Tracking cadence and the team-and-engagement view

Both metrics need to be tracked weekly at the team-and-engagement level, not monthly at the agency level. The monthly agency-level version is fine for board reporting and for trend lines, but it is too aggregated and too lagging for management. The drift that matters happens at the engagement level, week by week, and the lag between drift onset and detection determines whether the leadership team can intervene before the engagement closes underwater. Real-time daily entry with in-flight budget burn dashboards exposes drift within 1–3 days. Weekly engagement review exposes drift within 7–14 days. Monthly review exposes it at 30–60 days, which is typically too late to recover the engagement — the work is delivered, the over-service is absorbed, the write-off is already locked in.

Three layers of granularity matter and each serves a different decision. The agency-wide monthly rollup serves board and investor reporting and tracks the 12-month rolling trend on effective rate, margin walk, and rate-card positioning. The team or department weekly view serves manager-level intervention — individual utilization vs role-tier target, team-aggregate effective rate, engagement red flags rolled up, BD and internal time as a percentage of available capacity. The engagement weekly view serves PM and AM-level intervention — budget burn in hours and fees against scope completion, effective rate for the specific engagement, role-mix variance against plan, outstanding PCOs and CO signature status, write-off accrual against threshold. Healthy agencies report all three. Most agencies in the 20-to-200 FTE band report the first and a partial version of the third, with no team view bridging them.

Weekly leadership operating review should anchor on the effective rate trend (rolling 4-week and 12-week), the top five engagements at risk by budget burn vs scope completion, the bottom-decile clients by realization, individual capacity flags above 85% utilization, and outstanding change orders aged more than seven days. Monthly review extends to the full margin walk, realization distribution by service line and AM, and rate-card discount-governance review. Quarterly review covers client profitability with explicit repricing or offboarding decisions, rate-card adjustment against comp and inflation, and capacity plan. The cadence is identical to what mature law and architecture firms run — both sectors have more contractual rigour around realization and both publish benchmarks above where mid-market agencies typically sit.

The gating factor on all of this is time-tracking culture. None of the cadence works without a daily time-entry SLA, enforced. ClearPoint and LeanLaw both quantify the cost of delayed entry at 10–20% under-recording of time, which compresses both utilization and realization simultaneously and undermines every downstream metric. Best practice — adopted from the legal and accounting professional-services discipline — is daily entry with manager weekly review of missing or implausible entries, easy capture tools (mobile, calendar integration, browser extensions), and leadership logging time visibly to set the cultural norm. If leadership does not log time, no one will. This is the single highest-leverage cultural change available to the operating partner, and it costs nothing.

06 Sequencing the rebuild — what to do in what order

The remediation work sequences across roughly twelve months in four phases. The phases are not strict — early wins from later phases will surface in the first sixty days — but the sequence matters because each phase establishes the instrumentation or discipline that the next phase relies on. The total realistic recovery for a typical mid-market agency starting at 75–85% realization and 50–60% agency-wide utilization is 8–15 percentage points of realization recovery and 5–10 points of utilization recovery, translating to 8–15 points of contribution-margin lift over the twelve months.

1. 01
   Phase 1 — Instrumentation (days 0–60): Enforce daily time entry with manager weekly review. Implement in-flight budget burn and margin dashboards in Productive, Kantata, BigTime, or whichever PSA the agency already operates. Baseline current-state realization by client, project type, and AM. Build the agency-wide, team-level, and engagement-level views. Stand up the weekly leadership operating review with the agenda described above. Expect to surface 10–20% of additional billable time that was previously under-recorded — utilization lifts mechanically before any new scope discipline is in place.
2. 02
   Phase 2 — Scope and change-order discipline (days 60–180): Build standard SOW templates mapped to PSA project templates with enumerated in-scope deliverables, explicit out-of-scope examples, and stated assumptions. Implement the PCO-to-CO workflow with trigger thresholds, 24–48 hour PCO logging, and the "no CO, no work" rule with emergency lanes. Train all client-facing staff on identifying scope changes and escalating. Add scope-status to every client meeting agenda. Expect 5–10 points of realization recovery in this phase, concentrated on engagements with historical scope creep.
3. 03
   Phase 3 — Pricing and resourcing (days 90–240): Conduct the rate-card review against compensation and inflation; adjust 5–10% where justified. Implement discount governance with PM authority up to 5%, leadership approval above 10–15%. Build historical-actuals-based estimating templates per service line, replacing gut-feel estimates. Stand up skills-based resource planning in Float, Forecast, or the PSA. Implement role ladders mapped to the rate card with escalation rules when senior steps in. Expect another 3–6 points of realization recovery and 2–4 points of utilization recovery from better role-mix discipline.
4. 04
   Phase 4 — Write-off control and portfolio restructuring (days 180–360): Implement write-off thresholds with root-cause tagging — under-estimate, scope creep without CO, quality/rework, or client-relationship concession. Standardise invoice narratives. Move to mid-project milestone billing where appropriate. Run the quarterly client profitability review and make explicit decisions on chronic low-realization accounts: reprice, scope-reset, or offboard. Expect 3–5 points of additional realization recovery in this phase as the bottom-decile clients clear the portfolio.

The sequencing applies whether the agency is digital, creative, PR, or integrated marketing. Benchmark target bands shift modestly by type — digital and marketing 92–97%, creative 90–97%, PR 88–95% — but the mechanisms, playbook, and cadence are identical. The framework transfers cleanly to consulting firms, law firms with project-based practice areas, and accounting firms with advisory practices, with sub-sector-specific rate-card adjustments.

07 Where this fits in the operating-rebuild pattern

Agency operating-margin rebuilds are one expression of a broader pattern in our practice. Three other recent reads sit alongside this one. The Q2 2026 DTC and CPG M&A multiples read at /blog/dtc-cpg-mid-market-ma-multiples-q2-2026/ frames how operating discipline maps to multiple ranges in consumer-brand transactions. The Q2 2026 manufacturing multiples read at /blog/manufacturing-ma-multiples-q2-2026/ is the parallel piece for industrial sellers, where working-capital discipline plays the role realization plays in agencies. The commercial diligence read on customer concentration and churn at /blog/commercial-dd-customer-concentration-churn/ is the buyer-side companion — customer-level realization patterns are exactly what the buy-side commercial diligence team probes when mapping revenue durability.

For an agency operator three to five years out from a sale, the realization rebuild is the single highest-ROI operating discipline available. The 8–15 point lift in effective rate translates roughly one-for-one into operating-margin lift, and at typical mid-market agency multiples — 5–8x EBITDA for $5M–$30M EBITDA professional-services targets — every point of operating margin on a $20M revenue base is worth $1.0M–$1.6M of enterprise value. The twelve-month rebuild horizon fits comfortably inside a three-year exit-prep window. The work is structural, the playbook is published, the metric is measurable.

08 Three questions for the next operating review

1. Are both utilization and realization tracked weekly at the engagement level, and does the same dashboard surface both numbers side by side?
2. What is the current realization gap from the published 90–95% target band, and is scope creep being managed as formal change orders or absorbed into engagement margin at billing time?
3. What is the combined effective revenue rate (utilization × realization) — and is that the number the leadership team is managing to, or is it utilization-only with a quarterly realization review at the CFO level?