Our Insights

As a full-service media agency, Lewis Media’s tech stack isn't a side consideration. It's core to how we deliver for clients. We've spent years refining how we evaluate and adopt new platforms, and the process comes down to four questions.

Unfortunately, there is a pattern that plays out at many marketing agencies more often than people would like to admit. Someone comes back from a conference excited about a new platform. A competitor mentions they switched to something. A vendor gets a foot in the door. And suddenly there’s momentum to onboard a new tool before anyone at the agency has fully thought it through.

That’s how most marketing tech stack decisions get made: backwards. At Lewis Media, our data team has sat on both sides of this decision, building systems from scratch and inheriting ones that didn't fit, and the pattern holds up every time.

The result is an expensive lesson. Not because the tool was bad, but because it was the wrong fit for where the agency was at the time. The process gets driven by exciting demos and pricing tiers instead of an honest look at real requirements.

Here's a better starting point: four questions to consider before getting started.

Four Questions to Ask Before Choosing a Martech Stack

Question 1: Who Is Actually Going to Use This?

Who on the team will be using this tool every day?

A tech stack for marketing agencies usually serves two very different users: the data team who builds and maintains the system, and the internal media team or client who reads the output. What works for one rarely works for the other without deliberate planning.

A platform that a data engineer loves might be overly complex for an account manager trying to pull a quick number before a client call. A self-serve dashboard that clients can navigate easily might be limited in ways that frustrate the team building it. When you evaluate marketing technology tools, consider the teams using them and evaluate its capabilities from the perspective of each one. This is the same principle behind how we approach business intelligence and data visualization for our own clients; the dashboard has to work for the person reading it, not just the person who built it.

Question 2: What Does Your Real Data Look Like Right Now?

It's tempting to buy for where you hope to be in three years. The problem is that a tech stack for data analytics sized for a 40-client agency with 15 data sources is very different from what a 10-client shop needs today.

Buying for scale has a real cost. You end up paying for complexity your team isn't ready to use, and the gap between what the system can do and what the team does with it becomes a source of friction rather than value. Part of tech stack optimization is being honest about your current needs, not overshooting based on aspirations.

Ask: what data do we move, from where, and how often? What does our client mix look like and is it likely to change significantly in the next 12 to 18 months? The answers narrow the field faster than any feature comparison will.

Question 3: What Happens When the Person Who Built It Leaves?

This is the question agencies skip most often and regret the most.

Employee knowledge management is a real infrastructure problem. When a platform lives inside one person's head and that person exits, you don't just lose a resource: you lose the institutional knowledge of why things are built the way they are, where the edge cases are, and how to fix something when it breaks at 4pm on a Friday before a reporting deadline. 

One of the real advantages of moving to industry-standard tools is that more people know them. Documentation is better. Hiring into the stack is easier. But that only matters if the work itself is documented and transferable, which requires intention regardless of what tools you're using.

The best marketing technology integration decisions account for documentation, training, and handoff from the start. If onboarding a replacement would require rebuilding from scratch, that's not a tech problem. That's a risk that belongs in your evaluation criteria.

Question 4: What Is the Real Cost?

License fees are easy numbers to compare. They're also incomplete. 

The true cost of platform implementation includes migration time (moving existing data, rebuilding pipelines, validating that outputs still match), retraining time (every person who had muscle memory with the old system now needs to build new habits), and the slow period where both systems run in parallel because you can't fully cut over until the new one is validated.

For agencies mid-transition, that parallel period can stretch months. It's normal. It's also expensive in staff time and mental bandwidth in ways that don't show up on the pricing page.

When organizations evaluate tech stack effectiveness, the ones who build in realistic transition costs upfront make better decisions than the ones who discover those costs as surprises on the back end.

How to Build a Martech Stack That Actually Holds Up

A good martech stack isn't built around the most impressive demo or the platform with the longest feature list. It's built around fit: fit with the team's capabilities, fit with the current client portfolio, fit with how the agency produces and delivers work. 

How to choose the right tech stack comes down to being honest about these four questions before you start evaluating tools. Most agencies that struggle with their stack didn't pick a bad tool. They picked a tool based on where they hoped to be or what a vendor convinced them they needed. 

The good news is that martech stack optimization doesn't always require a full replacement. Sometimes the right move is rationalizing what you already have: cutting tools with overlapping functions, improving how the ones you use connect to each other, or investing in the training that closes the gap between what your stack can do and how your team uses it. 

Business intelligence reporting and digital transformation don't require the most sophisticated platform on the market. They require the right platform for the team using it, maintained well, with a plan for what happens when things change. 

Start there. The vendor list comes after.