Star Schema & Relational Tables Videos

"OK, here's how I explain it. Most CDPs treat your customers like a phone book. Name, number, maybe a note. One row. Done. But your customers aren't that simple. They've got household accounts, multiple products, years of history… And when you try to cram all of that into a single row? You lose the connections that actually matter.

It's the difference between a phone book and a family tree. One tells you a name. The other tells you how everything fits together. And if your personalisation doesn't understand those connections..., it's just guessing."

"So you've got all this beautifully structured relational data. Great. But here's the catch – every time you want to build a segment, the system has to join five, six, or more tables together. And as your data grows? That gets slow. Really slow. We're talking hours for a segment that should take seconds.

That's what star schema solves. It re-organizes your data so it's ready to answer the questions you actually ask… in real time. One central fact table, a few dimensions, and suddenly a ten million users segment in seconds. That's not a nice-to-have. That's the difference between planning campaigns and reacting in the moment."

"And I see this all the time. A customer does something important — upgrades their plan, abandons a cart, whatever — and the system captures it. But then it has to sync, batch-process, match it against data stored somewhere else.

By the time your follow-up message arrives it's hours later. Completely disconnected from what actually happened. Or worse. You send a promo code for something they've already bought.

The campaign logic isn't the problem… It's the gap between data and action. When your relational model and star schema live in the same platform, that gap disappears. You react while the moment still matters."

"Here's what I think is the real breakthrough. For years, the promise has been 'put your marketers closer to the data.' But the reality is, every time you need a complex audience, you're raising a ticket with engineering. You're waiting days. You're losing momentum.

Now imagine this: a marketer simply types 'users who added to cart but didn't buy in the last three days'... and the audience just builds itself. No SQL, manual joins or dependency on a data team. That's not just efficiency. That's a fundamental shift in who gets to make decisions. The architecture does the heavy lifting. The marketer just describes what they need to run the marketing."

"And this is the trap I see so many teams fall into. They think they need to model everything perfectly before they can do anything. So they spend months planning. Mapping. Restructuring. And the whole time, no campaigns are running. But what if you didn't have to wait?

Start with what you have. One use case. One audience. Prove the value, and then evolve the model as you go. The teams that win at personalization aren't the ones with perfect data. They're the ones who started before they felt ready. Two weeks. That's all it takes."

"That’s a question I am asked often. Look, if your AI can only see one customer in isolation — their clicks, their opens — how much is it really understanding? Because churn risk isn't just about what one person does. It's about what's happening around them. Changes in household behaviour. Declining engagement across linked accounts.

A signal tells you what happened. A pattern tells you what's about to happen. And when those predictions run on a star schema, like with D·engage, they're available right now. Not after an overnight batch job. That's the difference between reacting to churn and preventing it."