The road to succesful marketing analytics

Are you a marketer, manager, or something in between?

Then chances are you're on the hunt for the perfect marketing mix for your company. The right balance that helps you grow without blowing the entire budget on marketing. Sounds ideal, right? But in reality, this is one of the trickiest things to get right. Because… how do you really know if your marketing is working?

Spoiler alert:
Google Analytics won’t save you. It's based on clicks, so it ends up overvaluing click-heavy channels. That’s like judging a book by how fast someone scrolls past the cover.

And here comes the real spoiler:
You’ll never know the perfect marketing mix. But—good news—you can get very close by using data, smart models, and solid experimentation. I’ve had the privilege (and the pain) of going down this road a few times now.

Curious how to tackle this?

I’ve tried to explain a pretty complex topic without making it too dry, let’s call it:

The Climb to Marketing Effectiveness

It all starts in the quicksand of last-click attribution (1).
A place where many marketing teams get stuck. Decisions are made based on clicks, or insights handed to you by the marketing channels themselves. Unsurprisingly, this leads to most of your budget going to direct-response marketing. It’s hard to escape this cycle, because spending less feels very scary. That’s the quicksand.

Then something shifts.
You bring a first attribution model (2). Datadriven, maybe. Marketing mix modeling. Maybe even some AI or external tools. Slowly but surely, you pull yourself out of the quicksand. The effects of your marketing become clearer, and you can finally make long-term decisions. The model gives you stability, but there is a trap. 

You’ve made it to the mirage (3):
“We know exactly what every marketing dollar brings in!”
Mission accomplished! Right? Well… not quite. It feels like you’ve reached the peak, but you’re actually still down in a valley. Be careful not to settle there. Because the more you dig into your models, the more you realize: they’re not perfect. Some are flawed, others misleading. You're back to asking: “What is really working, and is there anything really working?”

Then you spot a path: the ladder of incrementality experiments (4).
You start running deliberate, carefully planned tests and compare those results to your models. you start to climb again. Sometimes that means turning off a campaign to see what happens to revenue. It’s uncomfortable. It’s hard. But it's progress.

At the top of those ladders? A steep cliff (5), and a bridge (6).

On the other side lies a new level of clarity, but to cross the bridge, something big has to happen: analysts and marketers must work together.
Together they:
a) get enough variance in your data to trust the results.
b) give each other feedback on what insights are actually useful for decision-making.

That bridge crosses a dangerous gap:
“We have insights, but we’re not doing anything with them.”
If marketers and data scientists don’t collaborate, you risk falling in. Models don’t help marketers. Marketers don’t help improve the models. All the effort? Wasted.

But once you cross the bridge, you reach solid ground (7).
Here’s where the road gets easier. You keep experimenting. There’s constant feedback between model outputs, marketing strategies, and data teams. You begin to value individual channels, sub-channels, and campaigns more accurately. Long-term marketing planning finally gets the spotlight it deserves.

Look back at your journey! And don’t forget to wave at your competitors, still stuck in the quicksand.But don’t pause for too long.There’s still more mountain to climb. Who knows how high you can go?