An SEO account can look successful from several angles and still be commercially weak for the agency delivering it.
The client may be pleased. Organic traffic may be rising. The monthly retainer may look substantial in the sales report. Yet excessive senior input, untracked revisions, dedicated software and recurring out-of-scope work can quietly remove most of the margin.
The opposite also happens. A calm, well-systemised account may receive less internal attention because it creates fewer problems, even though it produces healthy margin and has strong retention potential.
Measuring SEO client profitability corrects that imbalance. It gives agency leaders a client-level view of revenue, delivery cost, scope behaviour and commercial risk. That view supports better pricing, staffing and renewal decisions without reducing account management to a crude hours-versus-fees exercise.
The objective is not perfect accounting precision. It is a sufficiently reliable operating model that helps you decide where to intervene.
Define profitability before building the model
Agencies often use revenue, gross margin and profit interchangeably. That creates avoidable confusion.
For client-level management, I recommend starting with contribution margin:
Client contribution = net client revenue − direct and attributable delivery costs
Client contribution margin = client contribution ÷ net client revenue
Net client revenue is the amount the agency earns after discounts, credits and pass-through costs that should not be treated as service revenue. Direct delivery costs normally include employee time, freelancers, client-specific software, production expenses and an appropriate share of pooled tools.
Contribution is not the agency’s final net profit. Office costs, general management, sales, finance, insurance and other overheads still exist. You can allocate these later, but contribution margin is usually the more useful first layer because account teams can influence it directly.
Keep two views if necessary:
- Contribution view: shows whether the client funds its own delivery and contributes towards agency overhead.
- Fully allocated view: adds a consistent share of central overhead to support portfolio and financial planning.
Do not switch allocation methods from one client to another. A consistent imperfect rule is more useful than a theoretically precise rule applied selectively.
Step 1: Choose the reporting unit and period
Use the client account as the primary unit. If one client buys SEO, paid media and conversion work, decide whether you need profitability by client, service line or both.
A monthly calculation is practical for ongoing management. It aligns with retainers, payroll and most software billing. Add a rolling three-month view to reduce overreaction to a single technical project, content sprint or holiday period.
For fixed projects, measure both project-to-date performance and forecast-at-completion performance. A project can look profitable halfway through simply because the expensive implementation work has not started.
Assign every account a stable client ID. That ID should appear in time tracking, invoicing, project management, software allocation and the CRM where possible. Client names change and abbreviations multiply; identifiers are safer join keys.
Step 2: Calculate net client revenue
Begin with recognised service revenue for the period, not the headline contract value.
Include:
- Monthly SEO retainers earned during the period.
- Project or milestone revenue recognised under your chosen accounting policy.
- Approved change requests and additional work.
- Performance fees only when the relevant conditions have been met.
Subtract discounts, service credits and refunds. Separate advertising spend, link placement budgets or other pass-through amounts unless the agency genuinely earns and manages a margin on them.
Cash collection should be tracked alongside profitability, but it is not the same measure. An account may be profitable on paper while creating cash-flow pressure through late payment. Add debtor status as a risk signal rather than rewriting revenue whenever an invoice becomes overdue.
Step 3: Convert delivery hours into loaded labour cost
Hours alone do not reveal cost. Ten hours from a senior strategist and ten hours from a junior analyst have different economic effects.
Create a loaded hourly cost for each employee or role:
Loaded hourly cost = annual employment cost ÷ realistic annual productive hours
Annual employment cost can include salary, employer taxes, pension contributions, benefits and other employment costs. Productive hours should reflect leave, training, internal meetings and non-client responsibilities. Dividing salary by every theoretical working hour will understate delivery cost.
Then calculate:
Labour cost per client = sum of hours by role × loaded hourly cost by role
Use actual time where the data is reasonably complete. If time tracking is poor, start with scheduled capacity and run a short calibration exercise. Do not pretend incomplete records are precise.
Time categories should be specific enough to diagnose problems:
- Strategy and analysis.
- Technical SEO.
- Content planning, briefing and editing.
- Digital PR or outreach.
- Reporting and meetings.
- Account management.
- Rework and unplanned requests.
The last category matters. If revisions and reactive requests are hidden inside normal delivery, scope creep becomes difficult to quantify.
Step 4: Attribute software, contractors and production costs
Client-specific costs are straightforward. If an account requires a dedicated crawling instance, reporting environment, freelance writer or specialist developer, assign the cost directly.
Shared software needs a documented allocation rule. Common options include:
- Usage-based: allocate according to crawl volume, tracked keywords, reports or API consumption.
- Seat-based: allocate according to the team members serving each client.
- Portfolio-based: divide the cost equally across active clients when usage differences are immaterial.
- Revenue-based: allocate according to each client’s share of agency revenue.
Usage-based allocation is usually the fairest, but it can create more administration than insight. I would not build a complex allocation engine for a low-cost tool. Focus first on software and external costs large enough to change an account decision.
Automation can reduce collection work. For example, internal systems can use approved API connections, including services documented through OpenAI’s developer resources, to classify time notes or flag probable out-of-scope tasks. Human review should remain in place because an automated label is not a contractual judgment.
Step 5: Measure scope variance, not just total hours
High delivery cost tells you that margin has fallen. Scope variance helps explain why.
Record an expected monthly delivery plan by role and workstream. Compare it with actual activity:
Scope-hour variance = actual delivery hours − planned delivery hours
Scope-cost variance = actual labour cost − planned labour cost
Then classify the cause. Useful categories include agency inefficiency, underestimated recurring work, client-requested additions, approval delays, rework, technical incidents and intentional investment.
That distinction prevents unfair conclusions. Extra work caused by an agency error should not automatically become a client price increase. Conversely, a recurring stream of additional landing pages, meetings or stakeholder revisions is a commercial pattern that should be addressed.
Some overdelivery is deliberate. An agency might invest additional senior time during onboarding or before a renewal. Record it as intentional rather than hiding it. The cost still affects profitability, but the decision can be evaluated on its merits.
Step 6: Add retention risk without corrupting the accounting
A profitable account at serious risk of cancellation should not be valued like a stable account. However, retention estimates are planning inputs, not recognised revenue or accounting profit.
Create a simple risk assessment using observable signals:
- Payment behaviour.
- Executive sponsor strength.
- Client engagement and approval speed.
- Evidence of commercial outcomes.
- Contract end date and termination terms.
- Unresolved delivery or relationship issues.
- Dependence on one agency or client-side contact.
Use low, medium and high risk if your evidence does not support numeric probabilities. If the leadership team can calibrate probabilities against historical renewals, calculate a planning metric:
Risk-adjusted contribution = forecast contribution × estimated retention probability
Keep the unadjusted and risk-adjusted values side by side. Otherwise a subjective retention estimate can obscure the actual economics of delivery.
A concrete sourced-pipeline example
Suppose an agency wants to show whether SEO is producing commercially relevant demand while reviewing account retention risk.
GA4 events and fields
At minimum, capture session_start, relevant engagement events and generate_lead. The lead event contains these non-personal fields:
lead_id: a random identifier created when the form loads or submits.client_id: the agency’s account identifier, not a person’s identity.event_timestamp.page_locationandlanding_page.session_source,session_mediumand campaign fields.user_pseudo_idandga_session_idwhere available.
CRM fields
The corresponding CRM record contains lead_id, lead_created_at, account_id, opportunity_id, opportunity_stage, opportunity_amount, currency, close_date, won_revenue and lost_reason.
The join key is lead_id. The website writes the same random value to the analytics event and a hidden CRM form field. Email is not used as the key.
Calculation
Define sourced organic pipeline as the sum of opportunity value where the first eligible recorded session is organic search, the generate_lead event occurred within the reporting window, and the opportunity has reached an agreed qualified stage.
If two qualifying opportunities are worth £18,000 and £12,000, sourced pipeline is:
£18,000 + £12,000 = £30,000
A separate £25,000 opportunity that first came from a partner referral but later returned through organic search belongs in assisted pipeline, not sourced pipeline. This preserves the distinction in the reporting rules.
The figure can inform the account’s outcome and retention assessment, but it does not change the agency’s monthly contribution calculation. For more granular qualification, use a structured SEO lead-quality tracking process.
Worked client profitability calculation
Consider an illustrative £8,000 monthly SEO retainer. After reviewing actual delivery, the agency records the following:
| Item | Calculation | Monthly cost |
|---|---|---|
| Strategy | 18 hours × £65 | £1,170 |
| Technical SEO | 14 hours × £55 | £770 |
| Content delivery | 28 hours × £45 | £1,260 |
| Account management | 10 hours × £50 | £500 |
| Freelance production | Direct cost | £700 |
| Dedicated software | Direct cost | £450 |
| Shared software allocation | Documented allocation | £180 |
| Onboarding cost amortisation | Allocated over agreed period | £300 |
| Total delivery cost | £5,330 |
Monthly contribution is £8,000 − £5,330 = £2,670. Contribution margin is £2,670 ÷ £8,000 = 33.4%.
The account was planned for 58 hours but used 70. The 12-hour variance has already affected actual labour cost, so it should not be deducted again. It should be shown separately as a diagnostic signal.
If review shows that eight of those hours came from recurring stakeholder revisions and four from an agency implementation error, the response should differ. The agency should fix its error internally. The revision pattern may require a clearer approval process, reduced deliverables or a price change.
If the account’s estimated near-term retention probability is 70%, risk-adjusted planning contribution would be £2,670 × 70% = £1,869. That number is useful for capacity planning, but £2,670 remains the actual monthly contribution.
Turn the model into account decisions
A profitability report is useful only when it changes behaviour. Review accounts across four dimensions: contribution margin, scope variance, outcome evidence and retention risk.
| Pattern | Likely response |
|---|---|
| Healthy margin, low risk | Protect the operating model and avoid unnecessary customisation. |
| Healthy margin, high risk | Address outcomes, communication or sponsor alignment before cutting delivery. |
| Weak margin, low scope variance | Revisit original pricing, seniority mix, software requirements or service design. |
| Weak margin, high scope variance | Clarify scope, introduce change control and correct inefficient workflows. |
| Weak margin, weak outcomes | Build a recovery plan with a deadline; consider an orderly exit if the model is not repairable. |
Do not apply one universal margin threshold without considering the agency’s overhead and positioning. A specialist consultancy with senior delivery has a different cost base from a production-led agency. Set internal target, warning and intervention bands from your own financial requirements.
Build the monthly operating cadence
The model should not become a finance exercise that arrives too late to influence delivery.
By the first working week after month-end, combine invoicing, time, contractor and software data. Account leads then review scope causes and retention signals. Leadership should make explicit decisions: monitor, redesign, reprice, recover or exit.
Automate data movement, not judgment. A system can import time records, apply loaded cost rates, allocate tools and flag variances. It cannot reliably decide whether additional work was strategically justified or whether a client relationship is recoverable. This guide to automating agency reporting without losing judgment explains that boundary in more detail.
Limit manual overrides. When one is necessary, record the owner, reason and date. Silent spreadsheet adjustments quickly destroy trust in the model.
Use profitability data for pricing and resourcing
Client-level data improves new-business estimates. Instead of pricing from an ideal delivery plan, agencies can use actual hours from comparable accounts, including realistic account management, revision and reporting effort.
It also reveals where senior people are being used as a substitute for process. If strategists repeatedly perform routine reporting or production checks, the answer may be workflow design rather than a blanket fee increase.
For renewals, show the operational cause of any proposed change. A specific explanation—such as expanded regional reporting, additional technical environments or repeated content revisions—is more defensible than saying that costs have risen.
Portfolio planning then becomes clearer. Agencies can identify which service combinations are scalable, which client profiles create avoidable complexity and where standardisation would improve both margin and delivery quality. A broader SEO operating system for a small team can help turn these findings into repeatable workflows.
Frequently asked questions
Should agencies show profitability data to clients?
Usually not at the detailed cost level. Clients need clear scope, performance and commercial terms. Internal salary costs and margin targets are management information. Scope evidence may be shared when discussing a change request or renewal.
How accurate does time tracking need to be?
Accurate enough to identify material patterns. Use consistent categories, simple entry rules and periodic audits. False precision is not the goal, but systematic missing time will distort every account.
Should sales commission be assigned to a client?
It can be included in a fully allocated or customer-acquisition view. Keep it separate from recurring delivery contribution so leaders can distinguish acquisition cost from service economics.
How should onboarding costs be handled?
Show them directly during onboarding and, if useful for planning, amortise them over a documented expected period. Keep both views visible so amortisation does not conceal a costly implementation.
Can a low-margin client still be worth retaining?
Yes, temporarily or strategically. The account may provide capability development, portfolio balance or expansion potential. State the rationale, investment limit and review date rather than calling the account profitable when it is not.
Conclusion: make every exception an explicit decision
A useful SEO client profitability model does five things: records net revenue, converts delivery time into loaded cost, allocates material external expenses, exposes scope variance and adds retention risk as a separate planning layer.
Start with a monthly contribution calculation and a rolling three-month trend. Add outcome and attribution data without mixing client pipeline into agency revenue. Then require an action whenever an account falls outside your chosen operating range.
The biggest improvement is not the spreadsheet itself. It is the discipline of making trade-offs visible. Agencies can intentionally overdeliver, invest in a relationship or accept a lower margin—but those should be conscious, time-bound choices rather than surprises discovered after growth has already strained the team.