Most NZ brands don't need a CDP yet. Fix these four things first.

Short answer: if your tracking, consent and customer identifiers aren't in order, a customer data platform will just make an expensive mess faster. Fix the foundations first, then buy the platform.

I've spent a big part of the last decade selecting and implementing customer data platforms. Our Tealium work with Mr Apple helped grow its customer database by 150% and launch its first e-commerce campaign in India, so I'm not a sceptic. But the brands that get real value from a CDP have one thing in common: they did the unglamorous work before the software arrived.

What a CDP actually does

A customer data platform pulls first-party data from your website, app, CRM, point of sale and other systems, works out which records belong to the same person, and builds a persistent profile that other tools can use in real time. Done well, it means your email, ads, website and service team are all working from the same picture of the customer.

The catch is in that first step. A CDP inherits whatever you feed it. Identity resolution can't match people on identifiers you never captured. Audience activation can't fix consent you never recorded. And the AI personalisation everyone is excited about is only as good as the data underneath it.

The four things to fix first

1. Your tracking

Browser-only tracking leaks. When I wrote about signal loss for the Marketing Association in 2022, complete signal loss in key markets was already exceeding one in five users, driven by Apple's tracking changes, short-lived cookies, ad blockers and privacy-first browsers. It hasn't got better.

The fix is a hybrid approach: client-side tracking for rich behavioural detail, server-side tracking for the events that matter commercially, like sign-ups, purchases and renewals. Add retries and logging so you don't lose data when something breaks, and make sure you can backfill history when you need to.

2. Your event taxonomy

Most messy data starts with naming. One team tracks "Sign Up", another tracks "signup_complete", and a third doesn't track it at all. Write a tracking plan: a single document listing every event, its properties, where it fires and who owns it. It's dull, and it's the highest-return hour your team will spend this year.

3. Consent and identity

Decide what your primary customer identifier is, whether that's an email address, a customer number or a hashed phone number, and make sure every system stores it in the same format. Capture consent properly at the point of collection and store it alongside the profile. Privacy law in New Zealand and Australia is moving one way, and retrofitting consent is far harder than designing it in.

4. A use case with a number attached

"Single customer view" is not a use case. "Lift repeat purchase among first-time buyers by targeting them in their second month" is. If you can't name the first three audiences you'd activate and the metric each one should move, you're not ready to choose a platform, because you don't yet know what you need it to do.

Signs you're ready for a CDP

• You're stitching together more than a handful of data sources to understand one customer.
• You need to act on behaviour in real time, across several channels, not in a weekly batch.
• Different teams keep building their own lists, and they never quite match.
• Your people spend more time joining spreadsheets than acting on what they find.
• You need central control over consent and data governance.

If you're not there yet

There are cheaper first steps. Server-side tagging fixes a lot of signal loss. A data warehouse with reverse ETL can push clean audiences into your channels. Many email and marketing automation platforms already hold a usable customer profile. Start there, prove the value, and the business case for a CDP will write itself.

If you'd like a second opinion on whether you're ready, get in touch.