Governance is everything: Simo Ahava on making GTM a bridge, not a back door

I've been wanting to have this conversation with Simo Ahava for years. Not because I needed a hot take or a cool quote, but because Simo is one of the few people in this industry who made me better at my job without trying to sell me anything.

If you've worked with Google Tag Manager, server-side tagging, or any kind of client-side data collection in the last decade, you've probably landed on one of Simo's blog posts at 2am trying to figure out why your tracking is broken. The man has been writing technical guides, debugging walkthroughs, and JavaScript examples since before most martech vendors even existed. He's also the co-founder of TeamSimmer.com, a learning platform that's become the place to actually understand how this stuff works under the hood.

We covered a lot of ground in 45 minutes, but here's what I keep thinking about.

Tag management solved the wrong problem

Simo's origin story mirrors a lot of ours. Stat counters in the 90s. Google Analytics when it went free. Implementation work because, as he put it, "implementations suck." Google Tag Manager was supposed to fix that. It gave marketers the ability to deploy tags without waiting on developers. Problem solved, right?

Not exactly. What GTM really did was expose the next layer of dysfunction. Now you didn't need a developer to deploy bad tracking. You could do it yourself, faster, at scale. The deployment bottleneck disappeared, but the capability gap stayed exactly where it was. And that gap is what I see every time someone calls me in to "fix" a CDP that isn't working.

The same problems, just migrated

We spent a lot of time on Martech, and Simo was blunt. He's worked with a few, SaaS solutions and the pattern is always the same. The vendor shows up with a pitch deck full of promises. The implementation starts. Then reality hits. The data layer isn't clean. The identity stitching assumptions don't match actual user flows. The event taxonomy was designed by someone who left six months ago. And now the CDP is just another expensive pipe in a stack full of expensive pipes.

It's the exact same problem Google Analytics implementations had in 2010. The tools got better. The organizations didn't.

AI won't save you if you can't define an event

Of course we talked about AI. Simo's been teaching technical marketing for years, and his courses on BigQuery and server-side tagging are some of the most popular at Simmer. But when I pushed him on the "AI will solve martech" narrative, he was skeptical. Not because AI isn't useful, but because most companies still can't handle the basics.

If you can't design a clean data layer, AI isn't going to magically understand your business logic. If your event tracking is inconsistent, AI won't know which signals to trust. And if your org doesn't have the capability to maintain any of this, you're just adding another layer of complexity on top of an already fragile foundation.

This is something I've been circling around in my own work. The tooling isn't the bottleneck anymore. Capability is.

Building without data

One of the better tangents during our talk -> Team Simmer grew almost entirely without looking at their own analytics. No dashboards. No attribution models. Just a strong personal brand, a newsletter, and courses that actually solve problems people have. They tried Google Ads once this summer. It didn't work. Too niche, too much competition with generic terms, and their brand name conflicts with a food delivery service in the US.

I love this because it's a counterpoint to the "data-driven everything" religion. Sometimes the best strategy is just to build something useful and let people find it. Especially in a niche where the alternatives are either too shallow or too expensive.

Why this matters

I've been thinking a lot about capability lately. Not just skills, but organizational capability. The ability to actually execute on the tools you've bought. Simo's work, whether it's his blog, his courses, or this conversation, keeps pointing back to the same thing: the tools are fine. The vendors are fine. But if you don't have the internal capability to use them correctly, none of it matters.

CDPs are just the latest example. Tag management was the previous one. Before that, web analytics. The pattern doesn't change. We keep buying platforms that promise to solve problems we haven't actually diagnosed. And then we wonder why the implementation sucks.

Anyway, the full conversation is live now. Curious if this resonates with anyone else who's been in the repair business lately.