Few brands, agencies or publishers are truly happy with their data platforms – if they have one at all, that is (disclaimer, I’m the MD at predictive data management platform 1plusX).
If they opted for a customer data platform (CDP), they wish it had the audience activation functions of a data management platform (DMP). If they opted for a DMP, they wish it had the insight – the unified customer view – of a CDP. Often, their board only approved budget for one of these systems. Yet in the rush to get a data system in place to remain competitive, plenty of businesses are regretting their decision – or at very least doubting it. Many wished they bought into a CDP instead of a DMP and vice versa. If they have invested in both a CDP and a DMP, they wish they didn’t have to spend so much money on two systems that do similar things, and would prefer to do all they want under a single license.
One of the problems for data platform buyers is the sheer amount of choice available right now. Chiefmartec’s recent MarTech landscape report lists more than 1000 data platforms. There are at least 10% more DMPs listed than the previous year. Many of these new entrants are often just a “me too” DMP. Hardly any stand out with platforms and services that truly offer something qualitatively different (even if you might expect me to say that). Hence buyers might be forgiven for making a decision they regret. The amount of choice is resulting in confusion.
CDPs are generally sold to buyers on their ability to enable humans to make decisions about their customers based on an analysis and presentation of all their first party data. Anonymous and personally identifiable information from numerous sources and channels can all be presented cogently on one screen.
Those buyers may truly want a deep, personal relationship with their customers, or they may simply recognise the commercial value of the first party data they are collecting. Either way, their problem is making the most effective use of the first party data their customers have entrusted to them: it’s not in one place, comes from various channels and sources, and is in multiple formats (cookies, social media IDs, given names, loyalty scheme numbers, email and postal addresses and more). They will benefit from CDPs’ tag management and data integration heritage.
Having access to all this data in one place should give businesses a holistic view of each of their individual customers. (If it doesn’t then what’s the point?) That should establish deep understanding of their customers. From that new level of knowledge they want to communicate personalised offers and information to each individual within their CDP. Such bespoke communication should result in more sales conversions, bigger purchase sizes and more loyal customers.
The thing is, it’s still humanly impossible to execute such hyper-personalised communications at scale. Some CDPs can present customers with targeted product recommendations (which may be for products the consumer has already bought) or provide suggestions for customer support departments, but that’s the closest they come to providing anything like this sort of personalisation.
That’s why the people that bought a CDP want a DMP – or at least they think they do.
DMPs are generally sold to buyers on their ability to reach masses of people with data targeted advertising. The buyers are content to compromise on insight into their customer base for the immediate mass scale of communications that DMPs enable. Or that’s what they tell themselves at first.
These buyers either haven’t got enough first party data for a targeted ad campaign, or haven’t been through the process of integrating that data into a data platform, often because they see it as so fragmented or difficult – more so now the GDPR is in force – that they don’t even want to try to do it. (Such a waste of a huge opportunity!) Luckily for them, many DMPs don’t have the technology to consolidate first party data into their system. They rely on data from external third party data suppliers instead.
Rather than targeting specific individuals with content and bespoke offers, DMPs enable media and marketing teams to target content and offers to relevant audience segments – defined groups of consumers identified as having similar characteristics and interests. These interests and traits have been gleaned from the previous activity of people within the third party data pool that the DMP uses.
Unfortunately, when DMP buyers see the actual results of these campaigns, they’re disappointed that the targeting is just not personalised enough. Well, of course it’s not. It’s based on an audience segment, and more often than not they’ve had very little say in how that audience segment has been defined.
Firstly, the vast majority of DMPs and third party data suppliers only sell audience segments with a fixed definition, and minimal if any customisation. Secondly, it’s based on third party data: not only is third party data renowned to be frequently out-of-date and poor quality, the consumers in the data pool may have had absolutely no relationship with the client before.
What’s more, even though DMPs are attempting to target more consumers more accurately by deploying automated data processes, such as lookalike modelling, the outcomes and recommendations are far too inaccurate far too often. That’s because their lookalike models are based on third party data. The accuracy of all targeting technology is reliant on the accuracy of data fed into it. The results would be far more accurate if they were using current first party data.
To avoid data regret, buyers must really look for a data platform with the following fundamental functions:
- Easy consolidation of data from numerous sources and in various formats into a single platform;
- A customisable dashboard that analyses and cogently presents this customer and campaign data to users, ideally with a holistic view of their entire audience as well as the ability to drill down to individual consumer data;
- Custom audience building and segmentation functions through data processes including lookalike modelling, machine learning and predictive profile generation – preferably automated.
- Audience activation functions: so content and offers can be hyper-targeted to consumers at scale;
Is that a CDP? Or is it a DMP?
Or something else? A Unified Data Platform?
Who really cares?
Just use the criteria above to whittle down your choices from the 1000+ data platforms so you don’t regret your data decision.