Executive summary
- Half of the advertising sector's output is space resale, not agency services: 50.9% of service output in 2024 [1].
- Agency services shrank in real terms in 2024 (−0.8%), while media resale grew 10.2% [1],[3].
- EBE margin of 11.0% with staff at 61.3% of value added: every unbilled hour comes straight out of margin [1].
- Portuguese advertisers are consolidating agencies: 3.9 per advertiser in 2025, against 4.4 in 2023; only 63% of creative agency relationships are under contract [4].
- In the US, 78% of agencies rarely or only sometimes bill out-of-scope work and 57% lose USD 1,000 to 5,000 a month in unbilled work [13].
- Only 24% of Portuguese business services firms pay on time, with a 23-day average delay; applied as an illustration to every client of a typical agency, that would be about €45,475 financed by the agency [8].
- Three simple systems add +1.8 pp of EBITDA margin in the base scenario, with €14,500 of investment and a payback of 13.6 months (Vektrum estimate).
01
The essentials in five minutes
Executive summary
This report is written for the people who run a marketing and communications agency in Portugal with a team of five to twenty-five. It is about where an agency's margin leaks every week, in a market where agency services came to €1.04bn in 2024 [1], and which of the systems available today recover part of that margin with modest investment and controlled risk.
The sector described by the statistics is not the sector your agency operates in. In 2024, advertising in Portugal turned over €2.30bn with 17,887 people and 7,405 firms, or 2.4 people per firm [1]. But 50.9% of service output was the resale of advertising space on behalf of third parties, not agency services [1]. In 2024 media resale overtook agency services.
Agency work shrank in real terms. Agency services grew 1.6% in nominal terms, below average inflation of 2.4%: a real change of about −0.8% [1],[3]. In the same year, space resale grew 10.2% [1]. Advertisers' spending is growing; the share that pays for agency work is not keeping pace.
Margins are thin and the cost is people. The sector's gross operating surplus was 11.0% of turnover, and staff costs represented 61.3% of value added [1]. In that structure, every hour that is not billed comes straight out of margin.
Portuguese clients are cutting suppliers, not just prices. The average number of agencies per advertiser fell from 4.4 to 3.9 between 2023 and 2025, and only 63% of relationships with creative agencies are under contract (66% in 2023), against 87% for media and 88% for communications and PR [4].
Margin leakage is the norm, even in the most mature markets. In a survey of 273 US agencies, 57% lose between USD 1,000 and 5,000 a month in unbilled work and 78% rarely, or only sometimes, charge for out-of-scope work [13]. Applied to a typical Portuguese agency of 12 people, a write-off of just 5% of client hours costs between €12,600 and €15,750 a year at cost, equivalent to 16% to 20% of its EBITDA (Vektrum estimate).
Three simple systems recover a meaningful part of that margin. Time capture with scope control, automated campaign reporting and pitch qualification add up, in the base scenario, to +1.8 pp of EBITDA margin (from 11.0% to 12.8%), with an investment of €14,500 and a payback of 13.6 months. In the conservative scenario, the effect drops to +0.4 pp and payback rises to 55.7 months. All assumptions are visible in chapter 6.
02
Who is who and how each one makes money
Sector profile and sub-segments
A sector of micro-firms running two different businesses
For Portugal's national statistics office (INE), "Advertising" is NACE division 73.1, which combines advertising agencies (73110) and media representation and space sales (73120). Public relations and communications have their own code (70210). In 2024, advertising counted 7,405 firms (up 11.3% on 2023), 17,887 people employed and €2.30bn in turnover [1]. The average of 2.4 people per firm confirms what any manager in the sector knows: the overwhelming majority of agencies are micro-firms.
The same sector contains two businesses with opposite economics. The 3,684 agency companies turned over an average of 334.4 k€, and the five largest account for only 16.2% of the total. The 216 media representation companies turned over an average of 4,222.7 k€, and the 50 largest hold 97.9% [1]. The creative side is fragmented; the media side is, in practice, an oligopoly. The merger of OMG and IPG Mediabrands into Omnicom Media Portugal in December 2025 reinforces that concentration at the top [7].
| Activity | Companies | Service output (€m) | Average per company (€k) | Share of top 5 | Share of top 50 |
|---|---|---|---|---|---|
| Advertising agencies (73110) | 3,684 | 1,231.8 | 334 | 16.2% | 49% |
| Media representation (73120) | 216 | 912.1 | 4,223 | 43.2% | 97.9% |
| Public relations and communications (70210) | 604 | 144.8 | 240 | 23.4% | 70.8% |
Average per company is a Vektrum calculation (service output / number of companies). INE's concentration tables cover companies only; the total of 7,405 firms includes sole proprietors. Sources: [1]
Four sub-segments, four ways of making money
INE does not split creative, branding, digital and content agencies within code 73110, and this edition did not collect data on design (NACE 74.1) or events (NACE 82.30). The sub-segment view therefore combines the official structure with the Scopen Agency Scope Portugal 2025/26 study, which surveyed 280 marketing leaders at Portuguese advertisers [4]. It is the client's view, not the agency's, and that is precisely the view that decides renewals.
| Sub-segment | How it makes money | Relationships under contract (2025) | Average relationship length | Main exposure |
|---|---|---|---|---|
| Creative and branding | Project fees, creative retainers, mark-up on production | 63% (66% in 2023) | 5.3 years | More one-off projects, more pitches, scope changes |
| Media and performance | Fee or commission on managed spend, campaign management retainer | 87% | 6.4 years | Manual optimisation and reporting; concentration at the top |
| Digital (social, content, web) | Monthly management and production retainers; web projects | 84% | 5.2 years | Volume of recurring deliverables and approvals |
| Communications and PR | Monthly retainers; launch and crisis projects | 88% | 5.1 years | Unmeasured hours inside the retainer; pitches |
| Events and activation | Per-event projects, mark-up on production and suppliers | n/a | n/a | Capacity peaks, subcontracting, collections |
Percentages and durations: Scopen (advertisers' view). Revenue models and exposure: Vektrum reading. Scopen does not publish event-specific data. Sources: [4]
| Creative | 63% |
|---|---|
| Digital | 84% |
| Media | 87% |
| PR | 88% |
| All agencies | 81% |
Creative agencies are the only type that lost contractual share compared with 2023 (66%). Sources: [4]
Client money is moving. In 2025, Portuguese advertisers' communication budgets split 37% to traditional media (ATL), 26% to events and point of sale (BTL) and 37% to digital, against 43%, 21% and 36% in 2021 [4]. BTL's share in Portugal is higher than in Spain (21%) and the UK (22%); among the 11 countries in the study, only India (33%) is higher. Within digital, social and influencers rose from 38% to 46% between 2023 and 2025, while search fell from 30% to 23%; 71% of marketing leaders ran influencer activity in 2024 or 2025 [4].
Portuguese advertisers invest 3.3% of sales in marketing and communications, above the global average of 2.6% and below Spain (3.7%) [4]. Digital advertising in Portugal reached €612m in 2025, up 10.2%, ranking the country 20th of 30 European markets [5]. IPG Mediabrands' forecast, made in December 2024, pointed to national advertising investment above €900m in 2025, with digital growing 10% and television 3% [6]. The market is growing; the question for an SME agency is whether that demand reaches its billed hours.
Structural pressure: fewer agencies per client, fewer paid hours
In Portugal, the average agency-advertiser relationship lasts 5.5 years (6.4 for media, 5.3 for creative, 5.1 for PR), and 81% is under contract or retainer [4]. But supplier rosters are shrinking: 3.9 agencies per advertiser in 2025, against 4.4 in 2023, and 58% of relationships already operate under an integrated model [4]. Vektrum reading: when a client consolidates, the agency that leaves tends to be the one whose service is easiest for another to absorb, and being a client's fourth agency is a fragile position.
Selection criteria show where to invest: portfolio and credentials (64%), campaigns featured in the media (55%), capabilities and skills (46%), recommendation (42%) and availability (42%) [4]. Availability is an operational criterion: an agency that responds faster because its operation is organised competes better, even without changing its creative work.
Outside Portugal, the signals are harsher. In a Gartner survey of 402 marketing leaders, agencies account for 20.7% of the marketing budget, a share that is eroding, and 39% of CMOs plan to cut that allocation [23]. In the US, 82% of ANA members have an in-house agency (42% in 2008), although 92% still work with external agencies [24]. In the UK, employment at IPA member agencies fell 6.8% in a year, with creative agencies down 14.3% and media agencies up 2.4%; the census does not attribute the fall to any cause [25]. The digital agencies studied by Promethean grew only 4.5% in 2023, but those specialised by service and sector grew 93% faster than generalists [15].
Even among the large groups, operations separate the results. WPP reported a 5.4% like-for-like fall in net revenue in 2025 and expects a further decline in the first half of 2026 [42]. Publicis grew 5.6% organically with an operating margin of 18.2%, and its chief executive describes AI as a driver of growth and margin expansion [43]. Vektrum reading: access to technology is much the same for everyone; the difference is more likely to lie in how each group connects it to its operation.
03
Margins, costs and revenue mix
Economics and finance
An 11% margin in a business that sells hours
- staff costs / gross value added[1]
- 61.3%
- value added per person employed[1]
- 37.7 k€
- staff costs per person (lower bound)[1]
- 23.1 k€
Gross operating surplus (EBE) is the official measure closest to EBITDA. In advertising it stood at 11.0% of turnover in 2024, practically the same as 10.9% in 2023 [1],[2]. Staff costs absorb 61.3% of value added. Value added per person is 37.7 k€, and staff costs per person employed are about 23.1 k€ a year; the latter understates the cost of an employee, because the denominator includes unpaid managing partners and owners.
The bridge below shows where €100 of sector turnover goes before it becomes EBE. The largest slice is not staff: it is everything the firm buys from third parties, including the advertising space it resells. That is why a creative agency should not compare itself with the sector average without first removing media resale.
| Turnover | 100 € |
|---|---|
| External purchases and services, including resold space (turnover − GVA) | −70.7 € |
| Staff costs | −18 € |
| Other adjustments (residual) | −0.3 € |
| EBE (EBITDA proxy) | 11 € |
Vektrum calculation on INE Table 2 (turnover, GVA, staff costs and EBE). This is not net margin and it includes media sales companies. Sources: [1]
Media resale versus agency services
| Product | 2023 | 2024 | Change |
|---|---|---|---|
| Services provided by advertising agencies | 1,025 | 1,041.5 | 1.6% |
| Sale of advertising space or time on behalf of third parties | 1,010.8 | 1,113.5 | 10.2% |
In 2023 agency services were still the largest component; in 2024 media resale overtook them. Sources: [1],[2]
INE highlights that advertising has the highest service output per worker of all business services: €122.4k [1]. For a creative agency, the figure is misleading. Removing space resale, service output falls to €1.08bn, or about 60.1 k€ per person employed, less than half the headline figure. On that base, staff costs represent 38.4% of revenue, and even that ratio is a floor, for the reasons given above. An agency that benchmarks itself against the sector aggregates is comparing itself with an intermediation business.
Clients want fixed prices; agencies still do not know their costs
On the side of multinational advertisers, hourly remuneration is disappearing. In a WFA study of 69 multinational advertisers, only 17% to 19% use labour-based fees as their main model, against 54% in 2011, and 33% to 35% already pay a fixed fee or by output; in Europe the figure is 32% [28],[29]. Only 20% have already changed their commercial model with agencies because of AI, but 61% intend to and 63% expect to increase the performance-based component [28].
On the agency side, the shift is slow. In Productive's global study of 93 leaders of agencies with USD 1–50m in revenue, 76% rely mainly on project fees and fewer than 5% use value- or performance-based models [10]. In the US, Ignition found a mix of 28% hourly, 25% project, 28% productised or subscription and only 10% retainer [13]. The risk is clear: selling fixed prices without knowing the real cost of each deliverable transfers all overrun risk to the agency.
The same Productive research shows that growth is not enough: 59% of agencies increased revenue, but only 31% improved margin; fewer than half operate above a 10% margin and almost one in five cannot report theirs [10]. 45% can only forecast revenue one to three months ahead and 52% allocate resources four weeks or less in advance [10]. As for AI, about a third of agencies have already received "AI discount" requests from clients (November 2025), but in the March 2026 survey those requests had not increased and only 3% reported significant staff reductions due to AI [11],[12]. In Australia, 93% of agencies raised prices in 2025, 41% by 5% to 10%, and they still keep losing margin to unbilled work [14].
Cash: Portuguese clients pay late
In Portugal, only 20.2% of companies meet the payment terms agreed with suppliers; 65.2% pay up to 30 days late, 9.0% between 30 and 90 days and 5.6% more than 90 days late [8]. In business services, 24% pay on time and the average delay is 23 days [8]. In the European CRIBIS study, with data from December 2025, Portugal ranked second to last among 37 countries [9]. For an agency, these numbers describe its own clients.
As an illustration, in the typical agency used in this report, with €721,664 of annual service revenue, if every client paid with the 23-day average delay measured in business services, the agency would be financing about €45,475, equivalent to about 30 weeks of its EBITDA (Vektrum illustration: the 23-day delay is that of business-services firms as payers, not that of any particular agency's clients). In the US, 63% of agencies describe cash flow as unpredictable and 82% have postponed hiring or investment because of it; 71% say at least one in four invoices is paid late and only 16% require full payment upfront [13].
04
Six common leaks in agencies
Where money and time are lost
There is no public Portuguese benchmark on utilisation, unbilled work or scope changes in agencies. The available data comes from international studies, many of them run by software vendors with self-selected samples, and should be read as orders of magnitude. Even so, they all point in the same direction, and they allow the problem to be sized for a Portuguese agency with explicit assumptions.
1. Unbilled work and scope creep
In Ignition's survey of 273 US marketing and advertising agencies, 57% lose between USD 1,000 and 5,000 a month in unbilled work, 30% say scope creep costs them more than USD 5,000 a month and 78% rarely, or only sometimes, charge for out-of-scope work [13]. The cause is rarely client bad faith. It is the absence of a formal moment when someone compares the request with what was contracted.
In a typical agency of 12 people, with 1,500 effective hours per person and 70% utilisation, there are about 12,600 h of client work a year. A 5% write-off (the low end of the range we use) is 630 h, between €12,600 and €15,750 at cost, and more if valued at selling price. Against EBITDA of about €79,462, that is a leak of 16% to 20%.
2. Utilisation and time sheets
Average billable utilisation in professional services fell to 66.4% in 2025, the lowest in the 19 years of SPI Research's study (509 organisations, not only agencies) [17]. Among the digital agencies studied by Promethean, the average target for production teams is 72% (68% in 2022) and two thirds of agencies have targets in place [15]. In Teamwork.com's study of more than 500 leaders, 52% of agencies struggle to exceed 50% billable utilisation and only 18% always hit their billable time targets [16].
The root is time capture. Only 33% of leaders say their teams submit time sheets on time about 90% of the time [16]. Without reliable time data, profitability per client is a guess. In the typical agency, each percentage point of utilisation is worth 180 h a year, or about €4,050 at cost.
3. Meetings, briefs and approvals
Agency leaders spend 69% of their time in meetings [16]. Much of what is decided in those meetings is written up again by hand: minutes, client recaps, internal briefs, tasks in the project tool. It is necessary work, almost never billed, and the first thing that gets dropped when the week gets tight.
The brief is where most waste happens. In the BetterBriefs study (more than 1,700 respondents in over 70 countries, 2021), 80% of marketers think they write good briefs, but only 10% of agencies agree; an estimated 33% of marketing budget is wasted on poor briefs, and 73% of agencies say rebriefs are too frequent [20]. After the brief comes approval: the typical creative process takes eight days and more than three versions, and about nine in ten account managers name waiting for feedback as the main blocker [19].
4. Pitches and proposals
The most cited benchmark comes from UK PR agencies: 27 pitches a year and 62 hours per pitch, about 1,674 h a year [21]. A typical agency of 12 people with 15 active clients will pitch less: at 12 to 15 pitches a year (Vektrum estimate) and the same 62 hours, that is 744 h to 930 h, or 4.1% to 5.2% of the team's available hours. In a MediaSense survey, 86% of agencies consider the pitch process excessively time- and cost-intensive, and 64% say it harms their teams' mental health [22]. 34% of agencies convert between 31% and 50% of pitches, and for the third year running client acquisition is the main operational pain for 34% [18].
Portugal has a particularity. In October 2020, the Competition Authority fined the advertising agencies' association €3.6m over a 2008–2009 good-practice guide that recommended tenders with at most three agencies (four including the incumbent) [40],[41]. The practical consequence is that agencies cannot coordinate tender rules as a sector: each agency has to decide, alone and with discipline, which pitches it answers.
5. Manual reporting and optimisation
In a DoubleVerify study of 1,970 executives worldwide, agencies spend 26% of their time optimising campaigns manually, about 10 hours a week per employee; in the US that equals more than USD 17,000 per worker per year in lost productivity [30]. The figure applies mainly to media and performance roles. At the same time, 70% of agency leaders consider client reporting extremely important for retention, and 81% name the relationship as the main retention factor [18].
We found no verified benchmark for reporting hours per client. We use a range of 3 to 6 hours per client per month, to be validated in each agency: pulling data from three to five platforms, formatting and writing the commentary. Assuming 8 of the 15 active clients have monthly campaign reporting (the media, performance and digital accounts; reporting weighs little in creative and events work), that is between 288 h and 576 h a year, most of it spent on collection and formatting, not analysis.
6. Collections
84% of US agencies spend between 3 and more than 10 hours a month chasing payments [13]. In Portugal, where only 24% of business services companies pay on time [8], it is reasonable to expect at least the same effort. It is low-value, repetitive, rule-based work: the ideal candidate for automation.
| Leak | Evidence | Scope of the evidence | Weight in the typical agency (estimate) |
|---|---|---|---|
| Unbilled work | 57% lose USD 1,000–5,000/month; 78% do not consistently bill out-of-scope work | US, 273 agencies | 630 h a year at 5% write-off |
| Utilisation | 66.4% average; 52% struggle to exceed 50% | Professional services (global); agencies (global) | 180 h per percentage point |
| Pitches | 27 pitches × 62 h a year | PR, United Kingdom | 4.1%–5.2% of available hours (12 to 15 pitches) |
| Reporting | 26% of time on manual optimisation; 70% say reporting is critical | Global (media and performance); agencies (global) | 288 h–576 h a year (8 clients) |
| Briefs and approvals | 80% vs 10% on brief quality; 8 days and more than 3 versions per piece | Global (2021); global (2023) | Not quantified |
| Collections | 24% of business services firms pay on time; 23-day average delay | Portugal | €45,475 financed by the agency (illustration) |
Weights computed with the typical-agency assumptions in chapter 6. They are Vektrum estimates, not measurements. Sources: [8],[13],[16],[17],[18],[21],[30]
05
Revenue, margin, cost and automation by sub-segment
Opportunities
AI is already in agencies, but mostly in production
In 2025, only 11.5% of Portuguese companies with 10 or more people used artificial intelligence (7.9% in 2023), against 20.0% in the European Union; the EU rate is about 1.7 times Portugal's [31],[32],[33]. Among small firms (10 to 49 people) the rate is 9.4%. Among companies using AI, 40.8% apply it to administrative and management processes and 36.9% to marketing and sales [31]. The barriers cited by non-users are lack of knowledge (74.4%), cost (60.4%), legal doubts (57.3%) and privacy (55.4%) [31].
In agencies, adoption is higher and concentrated in production. In the US in 2024, 61% of agencies already used generative AI and 30% were exploring it; 69% of creative agencies, 57% of media agencies, 78% of large agencies and 53% of small ones (fewer than 50 people) [34],[35]. Among those already using it, 74% give high priority to creative ideation, 59% to summarising audience insights and 49% to summarising marketing results [35]. In Spain, 86% of marketing and communications professionals use AI mainly to produce content, only 27.5% are at an advanced stage and 45% of agencies say AI has not yet significantly changed structures, processes or roles [36].
The results are uneven. 58% of agencies say AI has reduced content creation time and 42% have recovered 5 to 10 billable hours a week [18]. On the client side in the UK, 65% of advertisers use generative AI regularly, but only 28% report a meaningful impact on day-to-day work, and 63% have seen changes in their relationship with content production agencies in the last 12 months [26],[27]. When AI is used only to produce faster, the risk is that the client asks for that saving to be passed on: about a third of agencies have already received "AI discount" requests [11]. When it is used to organise the operation, the saving tends to stay with the agency.
Revenue: charging for what you already do
- Scope-change detection and change orders. Every request arriving by email, meeting or message is compared with the contracted scope; when it falls outside, a short estimate is generated for client approval before work starts. It directly addresses the 78% of agencies that do not consistently bill this work [13].
- Repricing based on real cost. With three to six months of hours per deliverable, the agency can offer productised retainers or fixed prices with a known margin, the model multinational advertisers are moving towards [28],[29].
- Retention through reporting. Consistent monthly reports, with human commentary and recommendations, protect the retainers that make up 81% of relationships in Portugal [4],[18].
Margin: knowing which clients pay for the operation
- Frictionless time capture and profitability per client. Time entries are prefilled from the calendar, completed tasks and transcribed meetings; staff confirm instead of filling in. Each month, every client appears with hours, cost and fee side by side.
- Capacity planning. With 52% of agencies allocating people four weeks or less in advance [10], a simple view of load per person per week avoids overtime and last-minute subcontracting.
Cost and time: automating the movement of information
- Automated campaign reporting. Data from ad, social and analytics platforms arrives every month in an already formatted report template; the team writes only the interpretation and recommendations.
- Meeting → recap, brief and tasks. The meeting is recorded and transcribed, and the system proposes the client recap, the internal brief and the tasks in the project tool; a person reviews before anything is sent. It addresses meeting time [16] and brief misalignment [20].
- Proposal generation from a structured brief. Credentials, case studies, team and pricing come from a library the agency maintains; the proposal is assembled from a form and reviewed, not written from scratch.
- Pitch qualification. A scoring grid (known budget, number of agencies, prior relationship, fit with the specialisation) decides go or no-go before the first hour is spent.
- Client knowledge base. Decisions, tone of voice, approvals and history become searchable in one place. In the UK, staff turnover at creative agencies reaches 27.6% [25]; without a record, client knowledge walks out with people.
- Retainer invoicing and collections. Recurring invoices issued on the right day and staged automatic reminders free up the monthly hours currently spent chasing payments [13].
| Opportunity | Creative and branding | Media and performance | Communications and PR | Events and activation |
|---|---|---|---|---|
| Time capture and client profitability | High | Medium | High | Medium |
| Scope-change detection | High | Medium | High | High |
| Automated campaign reporting | Low | High | Medium | Low |
| Meeting → recap, brief and tasks | High | Medium | High | High |
| Proposal generation | High | Medium | High | High |
| Pitch qualification | High | Medium | High | Medium |
| Client knowledge base | Medium | Medium | High | Medium |
Vektrum rating, based on each sub-segment's revenue model (chapter 2) and the leaks described in chapter 4.
06
Typical agency, three opportunities, three scenarios
Quantification
This chapter explicitly separates two kinds of number. Real data are values published by official sources or studies, with their numbered source. Vektrum estimates are our own assumptions, labelled as such, used to size the impact in a concrete agency; they should be replaced with your agency's values. Calculations use the cost of hours, not their selling price, which makes the results conservative.
The average firm in the sector has 2.4 people [1], but the leaders this report is written for run teams of five to twenty-five, with recurring clients and their own processes. We therefore define a typical agency of 12 people, 15 active clients and 60% of revenue from retainers, a share below the 81% of contractual relationships in Portugal, adjusted for the weight of creative agencies (63%) [4].
| Parameter | Value | Basis |
|---|---|---|
| People (FTE) | 12 | Vektrum estimate |
| Active clients | 15 | Vektrum estimate |
| Revenue from retainers | 60% | Vektrum estimate based on Scopen |
| Effective hours per person per year | 1,500 | Vektrum estimate |
| Billable utilisation | 70% | Vektrum estimate between SPI (66.4%) and Promethean (72%) |
| Client hours per year | 12,600 | Calculation |
| Fully loaded cost per hour | €20.00 to €25.00 (average €22.50) | Vektrum estimate based on WageIndicator and Randstad |
| Service revenue per person | €60,139 | INE 2024, excluding media resale (calculation) |
| Annual service revenue | €721,664 | Calculation |
| Starting EBITDA margin | 11.0% | INE 2024, sector EBE/turnover (Vektrum calculation) |
| Starting EBITDA | €79,462 | Calculation |
| Implied staff cost | €405,000 (56% of revenue) | Calculation (hours × average cost) |
Cost per hour: gross salary of €1,600 to €2,000 a month × 14 months, plus 23.75% employer social security and other charges, over about 1,500 effective hours. The implied staff cost falls within the 40% to 60% of service revenue range assumed in this report. Sources: [1],[4],[15],[17],[38],[39]
A. Time capture and scope control
Time entries prefilled from the calendar and tasks, a change-order rule for out-of-scope requests and a monthly profitability report per client. The gain is the share of write-off that is now billed or avoided. It is valued at cost; billed at selling price, it would be higher.
Assumptions
- Typical agency team12 FTEVektrum estimate
- Service revenue (excluding media resale)[1]€721,664Vektrum estimate
- Starting EBITDA margin (sector EBE/turnover)[1]11.0%Vektrum calculation
- Fully loaded cost per hour[38],[39]€20.00 / €22.50 / €25.00Vektrum estimate
- Write-off on client hours[13]5.0% / 7.5% / 10.0%Vektrum estimate
- Share of write-off recovered25% / 40% / 50%Vektrum estimate
- Annual licence cost (12 users × €10 a month)€1,440Vektrum estimate
| Result | Conservative | Base | Ambitious |
|---|---|---|---|
| Hours per year recovered or freed | 158 h | 378 h | 630 h |
| Gross annual gain (at cost) | €3,150 | €8,505 | €15,750 |
| Annual running cost | €1,440 | €1,440 | €1,440 |
| Effect on EBITDA margin | +0.2 pp | +1.0 pp | +2.0 pp |
| Initial investment | €6,000 | €6,000 | €6,000 |
| First-year ROI | −72% | 18% | 139% |
| Payback | 42.1 months | 10.2 months | 5 months |
Values per scenario in the order conservative / base / ambitious. In the conservative scenario the payback is long: it only pays to proceed after measuring the real write-off.
B. Automated campaign reporting
Connections to ad, social and analytics platforms, a report template per client and monthly delivery, with interpretation and recommendations written by the team. Most relevant for media, performance and digital agencies.
Assumptions
- Typical agency team12 FTEVektrum estimate
- Service revenue (excluding media resale)[1]€721,664Vektrum estimate
- Starting EBITDA margin (sector EBE/turnover)[1]11.0%Vektrum calculation
- Fully loaded cost per hour[38],[39]€20.00 / €22.50 / €25.00Vektrum estimate
- Clients with monthly campaign reporting (of 15 active)8Vektrum estimate
- Reporting hours per client per month3 / 4.5 / 6Vektrum estimate
- Time reduction from automation60% / 70% / 80%Vektrum estimate
- Share of freed hours reused50% / 65% / 80%Vektrum estimate
- Annual connector and hosting cost€1,200Vektrum estimate
| Result | Conservative | Base | Ambitious |
|---|---|---|---|
| Hours per year recovered or freed | 173 h | 302 h | 461 h |
| Gross annual gain (at cost) | €1,728 | €4,423 | €9,216 |
| Annual running cost | €1,200 | €1,200 | €1,200 |
| Effect on EBITDA margin | +0.1 pp | +0.4 pp | +1.1 pp |
| Initial investment | €4,500 | €4,500 | €4,500 |
| First-year ROI | −88% | −28% | 78% |
| Payback | 102.3 months | 16.8 months | 6.7 months |
Excludes the effect on client retention, which is the strongest argument for reporting [18] but which we cannot quantify with verified data.
C. Pitch qualification and library-based proposals
A go/no-go grid before every pitch and proposals assembled from credentials and case studies already written. We use the 62 hours per pitch of UK PR agencies [21] and 12 to 15 pitches a year (Vektrum estimate): the UK study's 27 would be too many for an agency with 15 active clients. The 20% to 30% reduction corresponds to declining one in four or five low-probability pitches and reusing material.
Assumptions
- Typical agency team12 FTEVektrum estimate
- Service revenue (excluding media resale)[1]€721,664Vektrum estimate
- Starting EBITDA margin (sector EBE/turnover)[1]11.0%Vektrum calculation
- Fully loaded cost per hour[38],[39]€20.00 / €22.50 / €25.00Vektrum estimate
- Pitches per year12 / 13.5 / 15Vektrum estimate
- Hours per pitch[21]62 hStudy
- Reduction in pitch hours20% / 25% / 30%Vektrum estimate
- Share of freed hours reused50% / 65% / 80%Vektrum estimate
- Annual library maintenance cost€600Vektrum estimate
| Result | Conservative | Base | Ambitious |
|---|---|---|---|
| Hours per year recovered or freed | 149 h | 209 h | 279 h |
| Gross annual gain (at cost) | €1,488 | €3,060 | €5,580 |
| Annual running cost | €600 | €600 | €600 |
| Effect on EBITDA margin | +0.1 pp | +0.3 pp | +0.7 pp |
| Initial investment | €4,000 | €4,000 | €4,000 |
| First-year ROI | −78% | −38% | 25% |
| Payback | 54.1 months | 19.5 months | 9.6 months |
Excludes the effect of a higher win rate, which may be larger than the hours saved.
The three systems together
Assumptions
- Typical agency team12 FTEVektrum estimate
- Service revenue (excluding media resale)[1]€721,664Vektrum estimate
- Starting EBITDA margin (sector EBE/turnover)[1]11.0%Vektrum calculation
- Fully loaded cost per hour[38],[39]€20.00 / €22.50 / €25.00Vektrum estimate
| Result | Conservative | Base | Ambitious |
|---|---|---|---|
| Hours per year recovered or freed | 479 h | 890 h | 1,370 h |
| Gross annual gain (at cost) | €6,366 | €15,988 | €30,546 |
| Annual running cost | €3,240 | €3,240 | €3,240 |
| Effect on EBITDA margin | +0.4 pp | +1.8 pp | +3.8 pp |
| Initial investment | €14,500 | €14,500 | €14,500 |
| First-year ROI | −78% | −12% | 88% |
| Payback | 55.7 months | 13.6 months | 6.4 months |
Simple sum of the three opportunities. In practice they share data (time capture feeds proposals and profitability), so the combined investment tends to be lower than the sum.
| Current EBITDA | 79.5 k€ |
|---|---|
| A. Write-off recovered | +8.5 k€ |
| B. Reporting hours freed | +4.4 k€ |
| C. Pitch hours freed | +3.1 k€ |
| Annual running costs | −3.2 k€ |
| EBITDA in the base scenario | 92.3 k€ |
Vektrum estimate, illustrative, built with the assumptions above. The initial investment is not part of annual EBITDA. Sources: [1],[13],[21]
In the base scenario, the three opportunities free or recover about 890 h a year and lift EBITDA margin from 11.0% to 12.8%, with a payback of 13.6 months. In the ambitious scenario, the margin effect reaches +3.8 pp; in the conservative scenario it is +0.4 pp and first-year ROI is −78%: the investment is only recovered after 55.7 months.
The gap between scenarios is explained almost entirely by a number most agencies do not know: how much work they give away to clients today. That is why the first action in the plan is to measure it.
07
Transferable capabilities, not agency case studies
What Vektrum has built and ideas for the sector
Vektrum has not yet delivered a project for a marketing or communications agency. What we have are systems built for our own operation and for clients in other sectors, which solve problems very close to those described in chapter 4: meetings that generate written work, proposals assembled by hand, scattered pipeline and tasks, meeting preparation done in a rush.
Vektrum case
Transferable capabilityMeeting recorder with transcription and CRM hand-off
Built for Vektrum's internal operation, so that no client meeting depends on manual notes.
What was built
- Simultaneous recording of microphone and system audio, from the menu bar
- Transcription with a provider processing in the European Union
- Export of the transcript and a structured summary, with decisions and tasks
- Sent to the CRM only after human review
What we learned
- In an agency, the same flow produces the client recap, the internal brief and the tasks in the project tool
- Human review before sending is what makes the system acceptable to clients and staff
Transferable capability, with no results measured in agencies.
Vektrum case
Transferable capabilityInternal CRM with pipeline, tasks and sales tooling
Built and used by the Vektrum team as the single source of clients, opportunities and tasks.
What was built
- Funnel and opportunity pipeline with stages and next steps
- Tasks linked to projects, clients and meetings
- Proposal support tools and sales preparation sheets (battlesheets)
What we learned
- Connecting meetings, tasks and clients in one place is the foundation for measuring hours and profitability per client
Transferable capability, with no results measured in agencies.
Vektrum case
Transferable capabilityProposal and deck generator from a structured brief
Built to produce Vektrum's proposals consistently, without assembling each document by hand.
What was built
- Structured brief in a simple file (YAML) with client, scope, deliverables and price
- Conversion of the brief into a validated intermediate specification
- Generation of the final document in HTML and PDF with the house template
What we learned
- In an agency, the library of credentials, case studies and team becomes the main asset; the proposal becomes a reviewed assembly
Transferable capability, with no results measured in agencies.
Vektrum case
Transferable capabilityAutomatic briefings before meetings
A capability developed in another sector, from each contact's history.
What was built
- Collection of the contact's history: previous interactions, notes and commercial status
- Generation of a short briefing before each meeting, with context and talking points
What we learned
- In an agency, the same principle applies to client status meetings and to pitch preparation
Transferable capability, developed in another sector, with no results measured in agencies.
Agency-specific ideas
- Scope alert in the inbox. Every client request is compared with the retainer scope; when it seems to fall outside, the account manager receives a ready-to-send change-order suggestion.
- Monthly profitability sheet per client. Hours, cost and fee for each account side by side, with an alert when a retainer exceeds an hours limit set by management.
- Brief checklist before starting. The incoming brief is checked against a list of required fields (objective, audience, budget, deadline, approver) and the gaps are sent back to the client the same day, before there is work to redo. Only 10% of agencies think clients write good briefs [20].
- Approval rounds with deadlines. Every version sent to the client has a response date and an automatic reminder; version history stays attached to the task.
- Go/no-go grid for tenders. A simple score before accepting a pitch, with the outcome recorded to refine the criteria over the year.
- Searchable client memory. Tone of voice, decisions, approvals and materials for each client in one place, so that one person leaving does not erase years of context. At UK creative agencies, annual staff turnover is 27.6% [25].
08
Twelve actions over 180 days
Action plan
The matrix ranks actions by expected impact, implementation effort and horizon. Actions in the first 30 days are mostly about measurement and rules: they cost little and decide where it is worth investing. Those from 30 to 90 days are the systems quantified in chapter 6. Those from 90 to 180 days depend on the data collected earlier.
Impact
EffortLow → High
0 to 30 days
- 1
Measure write-off for three months
Time recorded by client and request type, including what is not billed.
Impact 3/3 · Effort 1/3 · 0 to 30 days
- 2
Introduce the change-order rule
Every out-of-scope request gets a written estimate before work starts.
Impact 3/3 · Effort 1/3 · 0 to 30 days
- 3
Automate retainer invoicing and reminders
Invoices issued on the right day and staged reminders without manual work.
Impact 2/3 · Effort 1/3 · 0 to 30 days
- 4
Adopt a go/no-go grid for pitches
Five scored criteria before accepting any tender.
Impact 2/3 · Effort 1/3 · 0 to 30 days
30 to 90 days
- 5
Brief checklist with same-day return
Required fields checked and gaps returned to the client before production.
Impact 2/3 · Effort 1/3 · 30 to 90 days
- 6
Meeting transcription to recap, brief and tasks
Recording, transcription and proposed tasks, with human review before sending.
Impact 2/3 · Effort 2/3 · 30 to 90 days
- 7
Automated campaign reporting
Platform data in a monthly template; the team writes only the interpretation.
Impact 3/3 · Effort 2/3 · 30 to 90 days
- 8
Monthly profitability report per client
Hours, cost and fee per account, with an alert when a retainer overruns.
Impact 3/3 · Effort 2/3 · 30 to 90 days
- 9
Credentials library and proposal generator
Reusable case studies, team and pricing, assembled from a form.
Impact 2/3 · Effort 2/3 · 30 to 90 days
90 to 180 days
- 10
Review pricing based on real cost
Productised retainers or fixed price per deliverable, with known margin.
Impact 3/3 · Effort 3/3 · 90 to 180 days
- 11
Automatic scope-change detection
Requests compared with contracted scope and a change-order suggestion.
Impact 3/3 · Effort 3/3 · 90 to 180 days
- 12
Client knowledge base
Searchable decisions, tone and approvals, to withstand staff turnover.
Impact 2/3 · Effort 3/3 · 90 to 180 days
The order matters more than the list. Repricing without measuring real cost is a guess; automating reporting before knowing how much time it takes may solve the wrong problem. The first 30 days are for replacing estimates with the agency's own numbers, and those numbers decide whether the next actions go ahead, change order or leave the plan. In a sector where 45% of agencies can only forecast revenue one to three months ahead [10], a 180-day plan with decision points is an advantage in itself.
09
What is data, what is estimate and what is missing
Methodology, assumptions and limitations
Real data and estimates
- Official data: INE (Business Services Statistics 2023 and 2024, ICT surveys), Eurostat. Margin, staff and productivity ratios are Vektrum arithmetic on those values [1],[2],[32],[33].
- Studies: surveys by associations, consultancies and software vendors. Vendor studies (Productive, Ignition, AgencyAnalytics, Teamwork.com, Filestage) have self-selected samples and a commercial interest; we use them as orders of magnitude.
- Vektrum estimates: typical agency profile, cost per hour, effective hours, utilisation, write-off rate, reporting hours, hours reduction, recovery and reuse rates, investment and running costs. They are labelled in every table.
Assumptions introduced in this report
- Typical agency revenue: 12 people × INE service revenue per person excluding media resale (€60,139). Starting EBITDA margin equal to the sector's EBE/turnover (11.0%), which includes media companies.
- Write-off recovery rates (25%, 40%, 50%) and reuse rates for freed hours (50%, 65%, 80%) per scenario.
- Initial investment of €6,000, €4,500 and €4,000 and annual costs of €1,440, €1,200 and €600 for opportunities A, B and C respectively.
Limitations
- Banco de Portugal's sector tables (EBITDA margin, profitability and average collection periods by activity code) could not be extracted; the margins used derive from INE and include media sales companies.
- INE does not split creative, branding, digital and content agencies within code 73110. Design (NACE 74.1) and events (NACE 82.30) data were not collected; for events and activation, only the 26% BTL share of communication budgets is available [4].
- Benchmarks for utilisation, unbilled work and scope creep are international (US, UK, global). There is no equivalent public Portuguese benchmark.
- No 2026 salary guide specific to Portuguese agencies was obtained; cost per hour relies on WageIndicator and Randstad [38],[39]. Available talent data indicate that 40% of professionals would consider leaving for higher pay and 30% of companies postponed or reduced raises in 2025 [37].
- The Scopen study measures the view of 280 marketing leaders at medium and large advertisers, not that of SME agencies. Pitch data come from UK PR agencies, collected in 2021–2022.
- There is no verified benchmark for reporting hours per client or for time saved per transcribed meeting; for that reason, meeting transcription was not quantified.
Sources
Every figure in the report points to a numbered source. Vektrum calculations and estimates are labelled as such.
- [1]Business Services Statistics 2024INE · 2025 · Official source
- [2]Business Services Statistics 2023INE · 2024 · Official source
- [3]Inflation of 2.4% for 2024 as a whole (INE data)Diário de Notícias · 2025 · Press
- [4]Agency Scope Portugal 2025/26Scopen · 2026 · Study
- [5]Digital advertising in Europe exceeds €131bn in 2025 (IAB Europe AdEx Benchmark)ECO · 2026 · Press
- [6]IPG Mediabrands forecasts 6% growth in Portuguese advertising investment in 2025Meios & Publicidade · 2024 · Press
- [7]OMG and IPG Mediabrands merge operations into Omnicom Media PortugalMeios & Publicidade · 2025 · Press
- [8]Payment behaviour of companies in Portugal, 12th editionInforma D&B · 2026 · Study
- [9]Portugal second to last among 37 countries on payment punctuality (CRIBIS Payment Study 2026)Dinheiro Vivo · 2026 · Press
- [10]2025 Agency Industry ReportProductive · 2025 · Vendor
- [11]Agencies in the AI EraProductive · 2025 · Vendor
- [12]Agencies in the AI Era: Six-Month Pulse ReportProductive · 2026 · Vendor
- [13]2025 Agency Pricing and Cash Flow ReportIgnition · 2025 · Vendor
- [14]93% of agencies are raising prices, but why are so many still leaking profit?B&T · 2025 · Press
- [15]The State of Digital Services 2024Promethean Research · 2024 · Study
- [16]State of Agency OperationsTeamwork.com · 2023 · Vendor
- [17]2026 Professional Services Maturity BenchmarkSPI Research · 2026 · Study
- [18]2025 Marketing Agency BenchmarksAgencyAnalytics · 2025 · Vendor
- [19]The State of Creative Collaboration 2023Filestage · 2023 · Vendor
- [20]The Better Briefs ProjectIPA / BetterBriefs · 2021 · Study
- [21]Average PR agency spends £11k on pitching every yearPRCA · 2022 · Study
- [22]MediaSense Pitch SmartIt's Nice That · 2023 · Press
- [23]Gartner 2025 CMO Spend Survey: 39% of CMOs plan to cut agency allocationsChief Marketer · 2025 · Press
- [24]The Continued Rise of the In-House Agency 2023ANA · 2023 · Study
- [25]Agency Census 2025IPA · 2025 · Study
- [26]Generative AI Survey Results, July 2026ISBA · 2026 · Study
- [27]GenAI is reshaping agency relationships, finds UK advertiser surveyVideoWeek · 2026 · Press
- [28]Global brands are ditching labour-based agency fees, finds WFAVideoWeek · 2026 · Press
- [29]WFA: agency pay moves to output-based modelsContentGrip · 2026 · Press
- [30]Agencies spend 26% of their time optimising campaigns manually (DoubleVerify)ECO · 2025 · Press
- [31]AI adoption by Portuguese companies reaches 11.5% (INE 2025 data)ECO · 2025 · Press
- [32]20% of EU enterprises use AI technologiesEurostat · 2025 · Official source
- [33]Survey on ICT usage in enterprises 2023INE · 2023 · Official source
- [34]US agencies are currently leading generative AI adoptionForrester · 2024 · Study
- [35]Forrester / 4A's: generative AI in marketing agenciesMarketing Dive · 2024 · Press
- [36]First Barometer of AI in Marketing and Communication (AMKT and Foro-ia, Spain)Marketeer · 2026 · Press
- [37]Robert Walters Salary Survey 2026RH Magazine · 2026 · Press
- [38]Account manager: role and salaryRandstad Portugal · Study
- [39]Advertising and marketing professionals: salaries in PortugalWageIndicator · 2026 · Study
- [40]Advertising agencies association fined €3.6m by the Competition AuthorityJornal de Negócios · 2020 · Press
- [41]Competition Authority fines advertising association €3.6mRenascença · 2020 · Press
- [42]Strategy update and 2025 preliminary resultsWPP · 2026 · Study
- [43]Publicis Groupe 2025 full year resultsPublicis Groupe · 2026 · Study
Methodology
This report was prepared by Vektrum from a research dossier collected and verified on 16 November 2026. Every figure was checked against the cited source; values that could not be confirmed in the primary source were excluded, and those found to be incorrect were corrected before entering the text.
Sources are classified as official (INE, Eurostat), studies (surveys by associations and consultancies), vendors (studies published by software vendors) and press (news reporting studies or decisions). Whenever a figure comes from another country, its scope is stated in the text; no foreign data is presented as Portuguese.
Derived ratios (people per firm, EBE over turnover, staff over value added, per-person values, real growth, averages per company) are arithmetic calculations made exclusively on official values. Scenarios are computed in code from explicit assumptions, with the same functions across all Vektrum Insights reports, so that any value can be recalculated with a specific agency's numbers.
The systems described in chapter 7 are transferable capabilities built by Vektrum in other contexts. They are not agency case studies and we attach no numerical results to them.
