The summary

When the client in today’s case study came to us, they had a strong product, an established customer base in their home region, and credibility to expand in regulated and high-complexity industries.

What they did not have was a commercial growth path into new markets they had never sold into before. Our job was to design that strategy & path. The strategy we delivered was a 12-month international expansion plan built around one principle: go narrower with intent, sequence the tactics, and get proof before scale.

This case study walks you through how we developed the plan, what we delivered, and the early evidence the first 3.5 months of execution generated.

About the client

The client is an enterprise AI platform headquartered in the Netherlands, working with large organzations in regulated and high-complexity sectors.

Their product and services around the complex-tech SaaS they have is built for the operational reality of large enterprises: workflow integration, observability, governance, local deployment support, and the safe transition from AI experimentation to governed production. They had an established customer base across their beachhead market when they engaged us, with strong case study evidence, and they were ready to expand internationally.

The ambition was clear. Enter Western European markets within a 12-month window without burning the year on a broad pan-European push.

IMPORTANT NOTE: We are bound by NDA, so we will describe the work we did without naming the company.

The DMU: Who was in the room on the client side as a decision-making unit?

The decision-making unit was small, senior, and well-aligned. We worked across four people throughout the engagement.

The CEO was the starter and the actual decision maker. He brought us in, owned the strategic ambition, and signed off on every major choice the strategy required. His clarity on what he was willing to give up to make the strategy work was the single biggest reason the project worked.

The Head of Growth was our primary strategic counterpart and the secondary decision maker. He owned the operational translation of the strategy, ran the cadence with us week to week, and made the resource allocation calls inside marketing and demand. Every framework, every funnel design, every measurement decision passed through him.

The Marketing Manager owned day-to-day execution. She turned the plan into running campaigns, briefed creative work, managed paid media, and produced the operating reports we used in the bi-weekly reviews.

The Sales Director held the veto and acted as the strategy’s loudest internal supporter, that could have transformed into detractor easy enough. He challenged every choice touching commercial qualification, the workshop offer, and the marketing-to-sales handoff. Once those choices were made, he carried them into the sales team. A strategy failing to win the Sales Director’s confidence, usually does not survive Q2 in any enterprise environment, and we knew this going in.

Client situation analysis

To put it as simple and short as possible, the client had an internationalization growth & commercial translation problem.

Their product story read well on the website, the architecture made sense to highly technical buyers, and existing case studies in regulated industries were credible in their home market. What was missing was the bridge between the product story and a buying decision in a market where they had no commercial history, no local references, not much awareness, and definitely no relationships / network inside target accounts.

The targeted enterprise buyers were drowning in vendor outreach, fatigued by AI vocabulary, and pretty much skeptical of cold pitches failing to address the realities of their market.

The category context made things even harder. Enterprise AI buyers in financial services, telecom, and other regulated verticals were watching their internal pilots stall. They were blocked by workflow integration, governance, controls, and the operational discipline required to move an experiment into production. A vendor arriving with broad platform language was going to lose to a vendor arriving with a sharper operational story.

We saw four marketing & lead generation execution traps the client risked falling into if we built the strategy badly.

  • A pan-European push diluting focus across markets with different buyer behaviour
  • A vertical-first push splitting the message across multiple sectors, before any one wedge had landed
  • A content engine launched before the growth language and proof architecture were mature
  • A paid media build-out before the conversion path and trust assets were ready

The strategy needed to refuse all four traps without losing the team’s appetite for movement.

The challenge

The client team was explicit about what they feared most. They did not want a just a long and boring deck they would forget inside a folder on the server. They had been through a previous engagement getting a fancy and shiny document and no actionable strategy, and they were uninterested in repeating the experience.

At the end of our project they wanted a working system the team would run after we left, a measurement layer the CEO would read at a glance, and stage gates protecting them from their own enthusiasm if the early evidence said something they did not want to hear.

The market added its own constraints. Enterprise buying cycles in the targeted verticals run six to twelve months. The buying group on the customer side typically includes an executive sponsor, an operations lead, a technical evaluator, and a risk and compliance stakeholder, each with different fears and different proof requirements.

The deal does not close until all four are aligned and most of them stay invisible to the sales team for the first half of the cycle.

Hidden buyers shape the deal before sales gets access to them to address their fears and worries, which means trust transfer should operate as part of the marketing & sales strategy directly, rather than as separate activations.

Let’s talk objectives

We agreed three objectives at the start of the engagement, in this order.

  1. Validate one lead market with credible pipeline evidence inside the first two quarters. The UK was one of the strongest working hypothesis based on enterprise density, English-language execution, and active national AI policy support, but the choice had to be earned by interview evidence and early commercial signal.
  2. Establish one repeatable consultative growth path from interview to pilot to implementation. The first marketing tactics deployed would be insight-led and qualification-heavy, with demos coming later in the conversation.
  3. Leave the business with a working lead gen, demand gen and conversion framework rather than a collection of campaigns. The deliverable had to be designed more like an operating system the in-house team would run with some help from us, with the agency in a strategic governance role over time.

The buying group we designed the B2B digital marketing strategy for the enterprise AI platform

The first ICP was a bit narrower than the client wanted, but the benefits became clearer after the customer interviews. So, we started with large enterprises in financial services where service complexity, customer experience pressure, regulated workflow volume, and governance requirements made the product’s value clear without translation. The wedge was customer support, service operations, and internal support workflows needing orchestration, observability, and integration rather than another FAQ assistant.

Inside each target account, we designed against four roles in the buying group.

  • The executive sponsor cared about business case and strategic upside, and feared pilot theatre with no operating impact
  • The operations lead cared about service quality, throughput, resolution, and handoff design, and feared workflow disruption or poor adoption
  • The technical evaluator cared about architecture, integrations, and model flexibility, and feared security risk, brittleness, and maintenance burden
  • The risk and compliance stakeholder cared about governance, controls, and auditability, and feared data exposure or policy failure.

Every asset we designed had to serve at least one of these roles directly without alienating the other three.

Our work process

We ran the engagement in four phases plus the first 3.5 months of execution, with each phase ending in a stage gate decision.

Phase 1: Diagnosis and market validation

The first six weeks went into assessment. We audited the client’s historical inbound activities, demand by geography, mapped the existing customers for portability into the UK market, ran competitive research across the broader enterprise AI category, and conducted the first wave of buyer interviews to validate the UK choice and the financial services wedge. Public sources we leaned on included the UK government AI Opportunities Action Plan and 2026 progress materials, and the UK Department for Science, Innovation and Technology AI Adoption Research from January 2026. On top of this we used data kindly provided by the amazing team at Veridion to understand the market density, how they present themselves, where they are, and the firmographics. The analysis confirmed UK-first as the rational choice and financial services as the right beachhead vertical. Still, a lot of assumptions were used, but analyzing real data helped us remove some of the assumptions and focus on validating or invalidating the remaining assumptions faster and cheaper.

Phase 2: Strategic choice and positioning

Once the first step cleared the first stage gate, we moved into strategy design. The strategic choice had four parts: UK as the lead market, complex service and operations workflows as the wedge, an interview-to-workshop-to-pilot funnel marketing as the primary motion, and operational and transformation leaders inside large enterprises as the primary audience.

The positioning spine moved away from selling AI agents as a category headline and toward an operational frame: helping large enterprises move complex workflows from fragmented AI experimentation into governed production.

The message architecture was built around four testable claims:

  • production over pilots,
  • governance and observability over generic automation
  • workflow transformation over vanity AI phrases framing
  • time-to-value with deployment support over internal build complexity

Phase 3: System design

With the message and the audience locked, we designed the system the team would run. This included:

  1. the conversion layer of the website (a UK landing page, a financial-services page, a trust page covering governance and security, a form page, a success page and a benchmark page to support the interview funnel)
  2. the funnel architecture (three funnels designed to launch in sequence rather than in parallel, to support the high frequency low TAM impressions paid ads approach)
  3. the measurement foundation (stage-based key events tied to GA4 and connected to Google Ads conversions only for the actions guiding bidding, with enhanced conversions used where first-party data made it appropriate)
  4. the thought leadership activation model (one founder voice plus one technical voice with a defined cadence)
  5. the operating cadence (weekly buyer ICP review, bi-weekly funnel decisions, monthly proof extraction, quarterly stage gate reviews).

Phase 4: Sequenced rollout

The first funnel went live before the others. Funnel one carried the interview and benchmark offer, designed to learn buyer language and surface the early shape of pipeline.

Funnel two (the workshop and executive briefing path) was held back until the message had been validated through the first wave of interviews.

Funnel three (the proof-led ABM motion) was scheduled for the back half of the year, after the first case studies from the early commercial wins were available.

The paid media layer ran throughout, sized deliberately small at €7,600 over 3.5 months across Meta and Google Search, with Meta optimized for reach against ICP-adjacent audiences and Google Search picking up high-intent buyer keywords, with a small TAM (around 5000 accounts) and high frequency supported by 50+ creatives extracted from the landing pages, thought leadership pieces, interviews, and internal conversations.

Deliverables

The core deliverable was a 16-section GTM strategy document the in-house team would operate from, each with clear SOPs, checklists, and control keys.

It covered the strategic ambition, the analysis results, the strategic choice, the guiding policy, the market entry logic, the ICP and buying group design, the positioning and message architecture, the customer journey, the year-one roadmap by quarter, the funnel system, the website and measurement foundation, the thought leadership activation model, the operating model and governance cadence, the stage gates and KPI tree, and a final section on risks, trade-offs, and what to refuse in year one.

Around the strategy document, we delivered the conversion-layer wireframes and copy for the five priority pages, the message matrix for the first two funnels, the workshop and benchmark offer designs, the measurement specification for GA4 and Google Ads, and the operating cadence template the team would run weekly and bi-weekly. The document set was designed to be picked up and run by the Head of Growth and the Marketing Manager with us in the room as supporting characters.

Our solution

The strategic spine of the work was a series of choices about what to refuse. Six principles carried the plan from page one to page sixteen.

One market before many.

The UK was the only paid-media and commercial-motion market in the first 2 quarters, with Germany and the Baltics held as the secondary markets for second half of the year. Year-one optimization was for signal quality and execution speed.

One wedge before category expansion.

Financial services, complex service and operations workflows were the only attention-receiving combination for the first two quarters. Telecom and travel sat in the plan as secondary verticals that were ready for deployment once the first ICP and vertical gained enough traction.

One growth marketing path before channel expansion.

Interview to workshop to pilot was the first consultative and repeatable motion. Other paths stayed available but held back until the first one produced stable conversions.

Proof before paid scale. Paid media played an air-cover role at a deliberately small budget. The engine of the motion lived in the consultative work.

Operational readiness before SEO volume. The measurement layer, the trust page, and the conversion-layer pages were built before any meaningful content production began.

Leadership visibility with editorial discipline. Two voices with a simple cadence, instead of an AI content factory.

The point of these principles was to give the team a decision rule for every request landing in their inbox over the next 12 months. Anything strengthening market understanding, proof, conversion, pipeline, or internal repeatability belonged in the year-one plan. Anything else stayed out.

Overcoming the challenges

The hardest internal work in the engagement was saying no to ICP expansion without losing the energy of the team. Enterprise leadership teams instinctively want to do more, faster, in more markets, with more verticals, against more buyer types.

A bit unexpected, but the CEO was the strongest internal advocate for the discipline, and the Head of Growth carried the operational version of it through the weekly cadence. The Sales Director’s early support for the narrow ICP was decisive.

Once he carried the financial-services-first choice into the sales team, the rest of the organization aligned easier.

The second hardest piece was sequencing the funnels.

The Marketing Manager had the budget and the appetite to put two or three campaigns into the market in week five, and the discipline required to hold funnel two and funnel three back until the evidence from funnel one was in took explicit governance. We built the stage gates partly to make this hold defensible. Launching the next funnel was a leadership decision tied to evidence, with marketing in the operator role.

The third piece was the measurement layer.

We resisted the temptation to instrument everything and instead defined six stage-based key events: benchmark download, trust-pack request, workshop request, contact form submit, qualified meeting booked, and solution-design request. Only the events guiding bidding were promoted into Google Ads conversions, in line with current Google Analytics and Google Ads guidance on conversions versus key events. This kept the dashboard readable and stopped the team from optimizing against vanity signals.

Results

The right way to read the first 3.5 months of execution is by KPI tier and by what each number was designed to validate. The strategy was built around stage gates, and the early evidence cleared three of the five gates the plan defined.

Frequency over reach

The paid media layer delivered approximately 2,000,000 impressions across Meta and Google Search on a budget of €7,600 over 3.5 months. The blended CPM works out to around €3.80, with an average frequency of roughly 400 impressions per ICP account against the 5,000-account beachhead universe we designed.

The frequency math is one of the most important strategic points.

A broader TAM at the same impression total would have produced reach numbers a marketing report would be proud of and a per-account frequency invisible to enterprise buying groups. We picked the inverse.

The caveat is worth saying out loud:

Meta cannot guarantee LinkedIn-grade account-level delivery, so the per-account figure represents the strategic intent rather than a guaranteed delivery profile. With job title, industry, and company size filters on Meta plus high-intent keywords on Google Search, the targeting was tight enough for the saturation logic to hold, topped with a higher number of creatives.

We treat the impressions number as a tertiary signal.

Its job in the report is to confirm two things: ads were being delivered against the right audience filters, and frequency was high enough for the message to have a chance of landing inside the buying group and create top of the mind.

The middle of the funnel

The 2,000,000 impressions produced approximately 10,000 website visits across the campaign window, an effective click-through rate of 0.5 percent. The figure is inside the band of B2B Finance benchmarks on Meta .

The 10,000 visits produced 80 funnel entries: form fills for the entry offer requests, and trust-pack downloads. For cold paid social traffic from Meta to a high-intent B2B offer, this is a healthy figure. The conversion-layer pages did the work they were designed to do, after a session of adjustments and improvements in week 3 after we got the first data available.

The bottom of the funnel

Of the 80 funnel entries, 40 converted into completed interviews. A 50 percent interview rate from funnel entry tells us the message and the offer were calibrated correctly, because the people entering the funnel were qualifying themselves into the conversation rather than dropping off after a content download.

Of the 40 interviews, 25% were disqualified after the conversation, leaving 30 ICP-validated accounts. We treat disqualification as methodology calibration, rather than a loss. Disqualification protects sales capacity downstream and sharpens the ICP for funnel two and funnel three. The 30 ICP-validated accounts gave us the buyer language, the objection map, and the proof requirements feeding the message refinement, the workshop design, and the trust pack.

Of those 30 accounts, 6 became qualified leads handed to the Sales Director’s team inside the 3.5-month window. Of those 6 qualified leads, 2 converted:

  • one paying client (5 months after the start of the project)
  • one POC project (inside the 3.5 months for the initial stage), both inside the financial services vertical.

Key takeaways

This project reinforced principles applying across enterprise SaaS and deep-tech western expansion work.

Strategy is a series of choices about what to refuse.

Year one of a international expansion tests the discipline of the leadership team before it tests anything else. The plans winning year one are the plans refusing breadth.

Sequencing beats parallelism.

When three funnels launch simultaneously, they jam team capacity and dilute the data signal coming back from each one. Sequencing them produces a portfolio with compounding learning between stages.

Hidden buyers shape the deal before sales gets access.

Thought leadership in enterprise sales operates as trust transfer rather than volume publishing. Two voices with editorial discipline outperform a content factory across a 12-month window.

Stage gates protect the business from mistaking activity for traction.

Without gates, the team will keep producing motion and call it progress. With gates, the leadership team gets a clean go or hold decision at every quarter.

Report results by tier, with the headline number serving the strategy rather than the other way around.

Two million impressions on a saturated 5,000-account ICP is air cover for a consultative high-ticket motion. The same impression total spread thin across an unfocused audience would produce noise. The meaning of an impression number depends entirely on the audience underneath it.

Narrow ICPs beat broad ones at the close rate.

The reason our 33 percent close rate sat above benchmark was the upstream qualification work, at the ICP design and the interview filter, before the sales team invested cycles.

Paid media plays an air-cover role in enterprise ABM.

A €7,600 budget over 3.5 months is enough to give a consultative motion the saturation it needs to be heard. The engine of the commercial motion lives in the interviews, the workshops, and the sales conversations the air cover supports.

Work with Milk and Cookies Studio

We work with B2B SaaS, complex-tech, deep-tech companies bringing a strong product, an ambitious growth horizon, and the discipline to choose narrow before going broad. Our work spans go-to-market strategy, full-funnel marketing, demand and lead generation projects where the leadership team needs strategic governance rather than another agency execution complexity layer.

If you are planning your international expansion, we would rather have one direct conversation about the choices in front of you than send you another deck.

Reach us at [email protected] or through the contact form on our Contact page, here.