Rid your digital marketing stack of projects

As a marketing leader, you've likely experienced the cycle: kicking off a new website relaunch project, aiming for faster performance, higher conversion rates, and a more appealing design, all while ensuring you're on the latest CMS version. You partner with a top agency and secure internal resources, and the initial phase is exhilarating. Working towards a clear goal like a website launch is engaging, with everyone focused on delivery. Months later, the site goes live, bringing well-deserved internal recognition and celebration.

The challenge arises post-launch. Most team members return to their daily tasks, leaving lingering bugs and unfinished features. Creating new landing pages becomes cumbersome due to complex signup forms, and email automations are difficult to configure for each event. Crucially, the maintenance budget is often significantly smaller than the initial project budget, with fewer dedicated resources. The cycle continues as new frontend frameworks emerge and CMS capabilities require "upgrading."

How we got here

Many marketing and IT leaders now recognise that a website, or any system, is never truly "done." It must continuously adapt to evolving work methods, business model changes, and new technologies, with AI being just one example. A project, by its very nature, is finite and static once concluded. It struggles to accommodate changing requirements, which often derail project budgets. This inherent rigidity is the core issue.

The famous principle states that you cannot fix scope, cost, and timeline simultaneously; something must yield. While skilled project managers strive to fix scope and cost, they often find themselves adjusting timelines. Having worked in the software industry for a long time, I believe vendors share some responsibility. It's simpler to sell a product with a defined implementation project – a tangible, fixed endeavour – rather than an ongoing "product" approach.

IT leaders and consultancies also play a part. Securing budget and resources is easier for initiatives promising clear, tangible results by the next quarter.

What defines a product

Most people perceive a product as something purchased, whether physical or intangible, like an email marketing system. Here, it's helpful to think of a product akin to a SaaS solution. While physical products have new models, SaaS solutions excel at providing continuous updates and the most recent versions. (Modern cars with over-the-air updates are blurring this distinction.)

Consider your marketing stack as a SaaS product. While it comprises various software products, view the entire solution as "a product." Instead of defining all capabilities upfront, actively seek opportunities for constant evolution. Can conversions be improved? Can a marketer's workflow be simplified? Is there a data gap hindering customer experience?

Instead of a project manager, we have a product manager. This role involves managing and prioritising work with the team, but critically, it also focuses on how the product can enhance business outcomes. We must "sell" our product – even an internal one – by ensuring it addresses the business problems we aim to solve.

How this affects how we work

Six years ago, at Optimizely, I was responsible for our website and partner portal. We had managed our web experience in various ways, but I felt it needed to be treated as a product. We assembled a team comprising architects and developers from Epinova, our creative agency, and internal content strategists and specialists. Our goal was not just another website, but a platform designed for longevity.

The team operated like a modern product team. With marketing expertise embedded (uncommon in many product teams), the cycle from defining a task to developing a solution and empirically testing it was remarkably short. This fostered a deep understanding of our specific needs, reducing overhead and handoffs. Crucially, the website extended beyond the CMS, integrating deeply with our marketing automation system, data and measurement systems (tag management and analytics), and authentication infrastructure. This complexity was managed without cost overruns or siloed thinking, as everything was viewed as a requirement for our evolving product.

How to apply product thinking with your agency

Many agencies, such as Framna, who developed well-known apps like Swish and SJ, market their ability to develop digital products collaboratively with clients. This is also true for some web agencies.

I believe most companies are becoming software companies, necessitating in-house product leadership experience, whether within their martech or operations teams. Epinova, along with other agencies like Valtech or KnowIT, can and do form part of such in-house product teams. Sometimes, we even provide complete product teams for companies early in their project-to-product journey.

What sets Epinova apart is its internal product approach. Optimizely CMS is famously a "toolbox," requiring customers to develop their own foundation. This means significant work before creating unique aspects of their "product" – their digital customer experience. Epinova recognised this early, and instead of repetitive efforts, built a "product on top of a product" – a foundational layer benefiting every customer. This product is dynamic, constantly evolving, ensuring all clients access improvements and largely avoiding costly "infrastructure upgrades."

Before encountering Epinova, I was surprised by how most agencies operate: they invest very little in building the "institutional knowledge" that a product embodies. Even with excellent agencies, success often depended on the specific team assigned and their prior experience. With such an approach, I honestly questioned why one wouldn't simply hire in-house.

A layered approach to your martech product stack

It's clear I advocate for a product-centric approach, and for good reason. I also believe there's no such thing as an "off-the-shelf" software product; it always requires configuration, training, and re-evaluation as it evolves.

My advice is to adopt a layered approach. First, view your martech stack as a product. Gartner offers resources to help your IT team make this transition. Second, if needed, hire an agency that has itself built products fulfilling your needs. These agencies, in turn, should leverage products, whether commercial (like Optimizely's platform) or open-source. Increasingly, a foundational LLM model will also be part of this stack.

This also means favouring "use and extend" over "building from scratch." A website or app is a well-understood entity, as are integrations with marketing automation systems. Neither you nor your agency should be building these from the ground up; the "skateboard" should already exist. (This references Marty Cagan's famous skateboard analogy.)

Opening up for data-driven development and experimentation

Finally, one of the most crucial aspects of treating your marketing stack as a product is empirical evidence. Often, we don't truly know what works; we have hypotheses that need validation. We must observe customer reactions and learn from their behaviour. This is where analysis and experimentation become vital.

Data-driven work is virtually impossible within a project framework. Before launch, there are no users, limiting data collection to user tests (which are highly recommended!). Older software development approaches also fell short here, leading to the rise of agile methodologies three decades ago.

Whether we use a dedicated experimentation tool or rely on analytics, we always want data and customer input to drive our product backlog. Marketing teams increasingly share this desire. Philosophically, a product approach is dynamic and ever-changing – precisely what marketers and businesses need to be.