Adtech industry expert Volker Ballueder reflects on a year of mergers and acquisitions (M&A) in the martech space, breaks down the accompanying industry jargon, and then explains what this maelstrom of acronyms could offer marketers in the 12 months to come.
'Data is the new oil' – that’s what I've been hearing for a few years now, but despite us talking about programmatic and using data for so many years, it's now becoming real.
I had a discussion with an industry friend the other day where we asked about what's fueling the latest hype around programmatic advertising and the use of data? After all, this technology or concept is not exactly new.
Emerging technology means we can now make sense of data
The difference is the sophistication of applying the technology to the data already out there, the scalability, and the artificial intelligence (AI) that sits on top that allows us to analyse more data sets than ever.
It also allows us to make more sense of the data too. Plus, we are seeing the raise of what I call ‘analytics workforce’, smart people within the industry that work solely on interpreting and analysing data, making it ‘understandable’ for the marketer and chief executive officer.
So the industry is maturing and as we see more and more M&A, a new path has been opened up for other players to come into the market, and forge new dynamics.
Out of the large three data management platforms (DMP) in the industry – that's BlueKai, Krux and Lotame – the former two have been acquired by Salesforce and Oracle respectively. However, now there are new players coming to market, most demand-side platforms (DSPs) have integrated DMPs, or have acquired one along the way.
So while the need for a DMP is greater than ever before, there are changes on the horizon in 2018; ie the rise of the customer data platform (CDP) or single customer view (SCV) – just to throw more acronyms in.
The difference between a DMP and CDP
Gartner describes a CDP as 'an integrated customer database that unifies a company's online and offline customer data to enable modelling and drive customer experience.'
What this means is that it stores information beyond cookie data, compared to a DMP. For instance, a CDP works with both anonymous and personally identifiable individual (PII) data, eg name, postal addresses, emails, phone number etc.
Meanwhile, DMPs work with anonymous entities such as cookies or increasingly device IDs (as mobile becomes more of a priority for advertisers). Hence you find DMPs in the programmatic world a lot more, eg building online audiences, through using cookies for lookalike modelling, or re-targeting etc.
To complete the acronym journey and main differentiation, a customer relationship management (CRM) system mainly looks at existing customers while a CDP can identify new visitors to your website, hence bridging the gap between an anonymous cookie and a known customer.
How all these acronyms work together
So when I say information around the user, we mean the single view of the customer. Let’s look at an example. Let’s say you are a car brand that has a website. With a DMP you can identify cookies A and B, plus know whether the user has seen the sports car or the SUV. Based on that you retarget that user or find lookalikes via second and third party data providers. All facilitated via your DMP, fed back into your DSP for targeting.
What you don’t know, particularly in the car case – and we can discuss this for online transactions differently – is whether the user bought a car or not.
The CDP however can identify a user in a CRM database, so the auto manufacturer then knows that 'driver X' has bought the SUV. Then based on the purchase they know the annual household income, and how the car was purchased (cash vs. finance).
So through the use of clever analytics (as above) the brand can explore how much the driver might still be willing to spend, or how quickly they might upgrade to another model.
Bringing AI to the equation
Meanwhile, AI, which has somehow became a buzzword often met with cynicism (so let’s just call it a clever machine executed algorithm) can analyse all the data points of all the buyers across the world, and an analyst will from there be able to predict new target audiences.
Let’s stay with the example of the car company. If we think of countries where winter tyres are required – think central and eastern Europe – a relevant manufacturer can target users in that region with ads for winter tyres. Not only that, they can target those users based on weather and location data as well as a user's purchase history, along with a personalised offering.
Someone who bought the car in cash might get an offer to buy the tyres with a rooftop box in a one-off payment, while someone who did a finance deal gets an offer which is in monthly installments and for tyres only. And once sold, there are more accessories offered... that sophistication only comes out of a CDP.
Why CDPs will work well in social environments
The complexity of the user life cycle is ever increasing. Several touch points and different vendors and technology vendors make it complicated to track an individual user. Essentially for brands it only makes sense to get a CDP rather than DMP as CDP can match, eg emails with Facebook and target users according to behavioural Facebook data on top.
That in my opinion will be a more powerful way of using customer data than the sole use of unidentified users via cookies in a DMP. The latter still has its own rightful uses, and the combination is useful – when you can mix and match the PII with cookie data – but in the wake of privacy regulations such as GDPR, we should keep that discussion for another time.