What Changes Inside a Digital Agency When Growth Strategies Deliver

TL;DR

• Agencies in our Fast Growth tier averaged 34% growth in 2025. Agencies in the Slow Growth tier declined 15% on average.
• Positioning begins to provide its promised benefits when it changes which work the agency accepts, the proof it builds, the channels it uses, and the team it needs.
• Growing current accounts requires clear ownership, client intelligence, and separation between relationship development and delivery pressure.
• Retention and expansion create a valuable revenue base. Agencies still need to calculate how much new-client revenue is required to replace normal losses and reach their growth targets.
• AI is changing the work clients value and the way they discover agencies. Leaders need to revisit their positioning and channel assumptions as those changes appear in the market.

In 2025, digital agencies in our Fast Growth tier grew by an average of 34%. Agencies in the Average Growth tier grew 8%. Agencies in the Slow Growth tier declined 15% on average.

That is a 49-point spread inside a single year.

The data in our 2026 Digital Agency Growth Guide helps explain that difference. Fast-growing shops tend to make deliberate choices about positioning, elevate account management, systematize referrals, and align the functions responsible for revenue generation. Slower-growing firms tend to live with some combination of an undifferentiated position, an owner-dependent pipeline, passive referrals, and disconnected revenue functions.

To get a closer look at those decisions, we joined Carl Smith from The Bureau for a conversation with four agency leaders:

• Andi Graham, CEO and Managing Partner at Big Sea
• Ernesto Tagwerker, Founder and CTO at Ombu Labs
• Nora Lahl, COO and Partner at Lightburn
• Sebastian Chedal, Founder and CEO at Fountain City

This interview panel adds deeper operating detail that enhances our other survey data: what it feels like to stop accepting work that once paid the bills, split a role that has existed for years, or discover that the specialization that drove growth is losing value.

The conversation reinforced a basic point: a growth strategy becomes useful when it changes the agency. It changes what the firm sells and refuses. It changes the proof the market sees. It changes who owns the client relationship. It changes the mix of revenue the agency needs. Eventually, it changes the organization itself.

Positioning becomes real when it limits the agency

Specialization is now standard across digital agencies. Ten years ago, roughly 30% of agencies identified as specialists. Today, 86% use that label.

That makes specialization a weak differentiator on its own. A market full of specialists quickly learns to distinguish between agencies that describe a focus and agencies that have built around one.

Andi Graham described Big Sea's first phase as “positioning light.” Beginning in late 2022, the agency aimed its marketing and public presence toward the niches it wanted to serve. It continued taking attractive work outside those niches.

That was a rational decision. Big Sea had around 30 people to support, a long history as a generalist, and a valuable referral network of CMOs who could send the firm work. Refusing those opportunities before the preferred market could replace them would have created immediate financial pressure.

The agency committed more fully in 2024. Some team members left because they did not want to follow the new direction. That reduced the agency's cost base and made it easier to reject poor-fit work. The leadership team defined the projects and clients Big Sea would accept and began saying no to the rest.

It took time for the market evidence to catch up with the strategy. Big Sea needed enough right-fit engagements to replace the generalist work in its portfolio. It dedicated a full year of content to one niche. Its public proof gradually began to reflect the position it wanted to own. Andi now sees prospects arriving with a closer match on industry, project type, and engagement size.

The financial cost begins when the qualification rules change. Restricting the work entering the agency, rebuilding the portfolio, concentrating marketing investment, and accepting the talent implications require far more conviction than updating website language.

A usable position shapes which opportunities qualify for the pipeline and which ones the agency refuses. It determines the proof the team prioritizes, where marketing and business-development time goes, which capabilities the firm builds or stops supporting, and how leaders explain the choice internally.

When those decisions remain general, the market receives mixed evidence. The website can claim expertise while the portfolio, team, and sales process continue to describe a generalist.

A specialization still needs to remain valuable

Ernesto Tagwerker's experience adds another important dimension.

FastRuby.io built a strong position around Ruby on Rails upgrades and technical-debt remediation. The focus simplified marketing, helped the firm rank for valuable technical content, and drove growth for years.

AI has begun changing the economics of that work. Coding agents can now perform portions of the remediation that once required more human effort. Ernesto sees the bottleneck moving toward quality assurance, code review, and the judgment required to manage a much larger volume of AI-generated work.

Ombu Labs had already begun a separate move into AI and machine-learning work. Maintaining distinct brands allowed the team to concentrate technical-debt content at FastRuby.io and AI content at Ombu Labs. The firm is also experimenting with engineering as marketing, including an open-source Claude Code skill that puts FastRuby's expertise inside the workflow developers are adopting.

The important issue here is the value of the client problem. A clear specialization can still weaken as technology, competition, or client priorities change. The agency needs enough commitment to learn from the position and enough market awareness to update it.

This is why we recommend treating positioning as a living stance. Review it against won and lost work, pricing pressure, project economics, changes in buyer language, and the problems emerging next to the work clients currently purchase.

The commercial model should determine the organization

Lightburn's shift toward recurring revenue created a different problem.

The agency's project managers were being asked to occupy two distinct seats. They had to manage scope, schedule, technical coordination, and delivery risk. They were also expected to protect the relationship, understand the client's broader goals, find future opportunities, and help the account grow.

Those responsibilities can pull in opposite directions. A project manager who hears about additional work may immediately see a capacity problem. An account manager should be able to explore the opportunity without carrying the same delivery pressure.

Lightburn separated the responsibilities into account managers and producers. Account managers became responsible for the ongoing relationships supporting recurring revenue. Producers took responsibility for technically complex, longer-term projects with defined scopes.

Nora described the sequence clearly:

“The structure of the agreements came first. We structured toward what our clients needed, and then built the seats to match that service.”

That sequence matters. Lightburn did not begin with a generic account-manager job description and search for someone to fill it. The agency first designed its commercial agreements around client needs. It then created the roles required to deliver and grow those relationships.

The people were largely already there. The seats were shaped incorrectly.

Agencies often reverse this sequence. They add an account-management title, move an available PM into it, and attach a growth expectation. The underlying agreements, decision rights, compensation, delivery model, and access to client intelligence stay the same. The agency has changed the org chart without changing the system.

There is no universal requirement to separate account management and project management. A smaller shop with short, low-complexity projects may operate well with a combined role. The conflict becomes harder to manage as relationships become longer, work becomes more consultative, and expansion becomes a meaningful part of the revenue plan.

Our research consistently places account management among the most effective revenue-generation activities. We recommend elevating account management to peer status with sales, marketing, and business development. That means giving the function a voice in ideal-client decisions, service design, account strategy, and the broader revenue plan.

Administrative ownership can maintain an account; strategic ownership creates the conditions to develop it.

Existing-client growth has to be designed

Relationship-based growth continues to outperform most alternatives in the agency market. In our research, growing current accounts and client referrals were the only two tactics that scored above the effectiveness midpoint at every level of sales complexity.

Their reliability often hides the operating discipline behind them.

At Lightburn, the new account-management structure created room for account planning, roadmapping, broader stakeholder relationships, and earlier participation in client decisions. Account managers can look beyond the current scope and ask what the client will need next. They can introduce the agency to another business unit, make more of the agency's capabilities visible, and surface a problem before it becomes a formal project request.

The agency also created a client whitespace report. For every client, the team maps the services currently being delivered and the areas where it sees a credible opportunity to add value. That lets the agency prioritize outreach instead of waiting for clients to discover adjacent capabilities on their own.

Ernesto described a complementary system at Ombu Labs. The team follows up with current and former clients using timely technical developments as a reason to be useful. A new Ruby on Rails release or compliance deadline creates a legitimate reason to reconnect. The conversation begins with relevant information rather than a thinly disguised sales check-in.

The firm also moved away from rigid account-call questionnaires. More open-ended conversations helped the team understand why a client was requesting something and what larger need sat behind it. Repeated questions from clients eventually revealed that the market did not understand the full range of work the firm could provide. That contributed to a clearer staff-augmentation offer and opened conversations about custom AI work.

These examples point to a practical account-growth system:

1. Tier the accounts. Concentrate strategic account-management time where client fit, relationship strength, unmet need, and revenue potential justify it.
2. Assign a clear owner. Someone needs explicit responsibility for relationship health, value communication, and future opportunity.
3. Track account health. Combine delivery results with relationship depth, stakeholder engagement, client sentiment, and commercial risk.
4. Map the whitespace. Record which services the client uses, which problems remain unsolved, and which stakeholders could benefit from additional help.
5. Create a useful cadence. Give the team real reasons to reconnect, review progress, discuss the client's priorities, and communicate delivered value.
6. Bring delivery intelligence into the process. The specialists doing the work often see the next problem first. Train them to surface it without turning every interaction into a pitch.

This system supports retention, expansion, and referrals. It also reduces the chance that a client hires another agency for a capability your team already provides.

We explored the retention side of this system in more detail in Limiting Client Churn. The recurring activities are straightforward: track account health, maintain a consistent communication cadence, and regularly communicate the value the team delivers. The management discipline required to execute them is where agencies tend to struggle.

Account expansion cannot replace new-client replenishment

Strong account growth creates stability. Overreliance on it creates concentration and replacement risk.

Sebastian Chedal described a period when Fountain City's existing clients were expanding and producing meaningful growth. New-client acquisition received less attention. When a large, six-year client completed the digital-transformation work the agency had been hired to perform, that account ended and left a substantial hole. Fountain City rebuilt around its move into AI engineering.

The experience changed how Sebastian evaluates growth. He wants existing clients to stay and expand. He also wants a consistent flow of new clients entering the portfolio. Every account has a natural boundary, and every relationship eventually ends.

Andi described a more explicitly planned revenue mix at Big Sea. The agency's average client tenure is approximately six years. It begins the year with roughly 35% to 40% of revenue booked through retainers. It expects another 35% to 40% from additional work within existing relationships. The remaining 20% to 30% needs to come from entirely new accounts.

Those percentages describe Big Sea's model and should remain attributed to it. The right mix for another agency will depend on client tenure, account concentration, growth goals, engagement size, capacity, and the volatility of its market.

The useful practice is working backward from the revenue requirement.

Required new-client revenue
= Target revenue
− Expected retained revenue after modeled churn
− Expected expansion revenue from current accounts

Then translate the revenue requirement into a client requirement:

Required new clients
= Required new-client revenue
÷ Average first-year revenue per new client

Finally, use the agency's close rate to estimate the qualified opportunities required to produce those wins.

This calculation turns “we need more leads” into an actual commercial requirement. It also shows whether the agency's current marketing, sales capacity, and average deal size can plausibly support the growth plan.

Account management and acquisition should be planned together. Expected retention establishes the base. Expansion raises the value of that base. New-client acquisition replaces normal losses, reduces concentration, and supplies the next generation of accounts that can grow.

AI is changing service value and agency discovery

AI appeared on each side of the panel's growth stories.

For Ernesto, AI is absorbing parts of the technical-debt remediation work that supported a successful specialization. That creates pressure to move toward the judgment, review, and implementation problems surrounding AI-generated code.

For Sebastian, AI systems have become an important discovery surface. Fountain City is working to understand how different models perceive the firm, which underlying queries they use, and which sources influence recommendations. Sebastian's view is that recommendation position matters more than being cited somewhere in the answer. That is an informed practitioner observation, not a settled industry finding.

Andi reported that Big Sea recently closed a sizable engagement that originated from an AI recommendation. She still placed her highest confidence in targeted in-person conferences. Big Sea's tighter position makes it easier to identify the specific events where its buyers gather, decide where to speak or exhibit, and spread that work across several members of the team.

These cases fit a broader shift we describe in the >Growth Guide. Buyers are using AI to research and shortlist agencies. Traditional reach-based channels have become noisier. Trust-based channels such as referrals, account relationships, and live events have held up more effectively.

The practical requirement is making the agency legible to AI systems and credible to the humans making the decision. Clear positioning, relevant proof, authoritative content, third-party validation, and real relationships contribute to that outcome.

Agency leaders should also watch for AI changing the problem underneath their position. The risk extends beyond using AI to deliver the same work faster. AI can reduce the value of an existing service, move the delivery bottleneck, create a new buyer expectation, or change how the shortlist gets built.

Structural change needs sequencing and reinforcement

Designing the new model is only part of the work.

Lightburn spent months thinking through its account-management change before announcing it. Leadership understood the reasons, the tradeoffs, and the intended end state. The rest of the team received that context much later and continued pulling toward the roles and behaviors they already knew.

That created predictable friction. Production leadership had to release responsibilities that moved with the account managers. Digital-marketing analysts needed to make more of their work visible so account managers could understand and manage the full relationship. Team members who had asked for more integration still had to adjust when integration changed their meetings and routines.

Nora's team used employee feedback to connect the change to problems people had already identified. Managers asked pointed questions during one-on-ones, surfaced concerns to leadership, documented the new responsibilities, and repeated the reasons behind the structure. The agency eventually reduced some of the meetings it had added during the transition.

The delivery model changed after the people were sitting in the new seats. Lightburn redesigned its sprint model in the second quarter, once the account managers and producers had clearer ownership. Nora expected significant resistance. The delivery change was easier than expected because the role structure already supported it.

The sequence was:

1. Design the commercial model around client needs.
2. Define the roles and decision rights required by that model.
3. Place people in the new seats.
4. Redesign delivery around the clarified responsibilities.
5. Reinforce the change through feedback, management cadence, and documentation.

Skipping ahead creates unnecessary resistance. A new process layered on top of unclear roles forces people to negotiate ownership while trying to learn the process. Clear roles remove much of that argument before it begins.

Questions for your leadership team

The four agencies in this conversation are pursuing different positions, clients, services, and growth models. Their choices are inputs for sharper questions about your own agency:

• What work would we refuse today, even if we had capacity to take it?
• Does our portfolio prove the position we claim?
• Has the client problem underlying our specialization become more or less valuable?
• Do our commercial agreements, roles, and delivery model reinforce one another?
• Who owns client momentum independently of day-to-day delivery throughput?
• Which accounts deserve deliberate expansion plans, and what evidence supports each opportunity?
• How much revenue do we expect from retention, expansion, and entirely new clients next year?
• How many new clients and qualified opportunities are required to reach that number?
• Where is AI changing the work clients value or the way they build a shortlist?

The 49-point growth divide reflects the cumulative effect of all of those decisions that reinforce or fight with one another.

The full 2026 Digital Agency Growth Guide maps the system, benchmarks the major decisions, and lays out the sequence for improving agency growth systems. If you already know parts of your revenue engine are underperforming but need help identifying where the important gaps sit, our Digital Agency Growth Review is designed for exactly that problem.