Before committing to reading any further, you might well be wondering why this collection of letters is worth understanding. After all, you may just be getting your head around integrated CRM, the SCV or DMP.
In this post, I’ll try to explain – and I’ve been challenged to avoid jargon unless absolutely necessary. Wish me luck…
Firstly, what is a CDP when it’s at home?
The Customer Data Platform is a clever bit of tech finally merging those irksome channel silos (the focus for most marketers) to provide ‘true’ cross-channel communications. This fits alongside your existing CRM infrastructure (including SCV or a DMP) and in my view, will significantly unify the Martech space.
The aim of the CDP is to bring all available customer data, understanding, and ‘decisioning’ into one place, irrespective of the channel in which it happens. The team here has been developing our ‘Customer Engagement Programme’ over the last few years and the CDP works on a strikingly similar basis, providing an empirical understanding of the relationships between and relative contributions of channels. In short (and I know so many before have claimed it), it’s much more than an events stream: we are being presented with the first fully rounded view of every individual customer.
The aim of the CDP is to bring all available customer data, understanding, and ‘decisioning’ into one place, irrespective of the channel from which it originates.
This reduces the risk in assumption fuelling our business strategies – and the likelihood of John Wanamaker’s famous predicament becoming a reality*.
Its adoption is super-fast
With such critical abilities, it may be little surprise that the technology has gained significant traction in the last twelve months. In fact, the Customer Data Platform Institute [catchy title, guys!] claims that CDP vendors are forecasting an increaseon their 2016 revenue ($300 million dollars) by three-fold over the next two years – to $1 billion dollars by 2019. So this is most certainly something of which we should at least be aware – and quite possibly, consider.
I get it: it’s important. So what exactly can it do?
The CDP deserves to sit at the top table of marketing technology for four key reasons:
1: It provides an instant, truly unified view of the customer [Also: Why a CDP is not an SCV or DMP]
Q: Isn’t this exactly what a Single Customer View or CRM system already promises?
Well, not quite. While your SCV is positioned as the ‘Golden Record’ for every customer, it tends, in reality, to fixate on your ‘Known’ customers (in other words, those with personally identifiable data) and their associated transactions. A CDP goes much further by providing a platform to ingest both your Known customers AND ‘Unknowns’: customers about whom we may know very little. A CDP connects all of these customers’ ‘events’ which you may only currently see as a cookie, a username or a social media handle.
- The technology can consume all this data and form a ‘pool’ of Unknown, Partially Known and Fully Known customers – all with associated behaviours and brand interactions. In short, all of your addressable audience reside in a single place.
- ID management is a key component of the CDP: Deterministic and Probabilistic** algorithms are used to pull together identities from different sources. The methodology is complex, but for the purposes of this blog, the CDP is a master of matching: it can connect everything from IDs across sessions, known devices or personal data (such as emails or name address) through to scenarios without customer details (so-called ‘fuzzy matching’).
- As individual customer interactions increase, the CDP learns more about your customers, enabling us to engage in more sophisticated messaging by directing us on the most appropriate communications channels.
Q: Right. So the CDP is just a DMP then?
Again, not quite. But this a very common misunderstanding in our acronym-riddled world. The Customer Data Platform is not the same as a Data Management Platform (DMP) either:
- DMPs (generally) focus on ‘Unknown Cookie pools’, acquisition and bulk marketing – predominately through display advertising. They provide excellent capabilities for audience reach and acquisition through ‘look-a-likes’ and media spend optimisation.
- The individuals’ customer journey is not a priority; the DMP does not provide the open platform nature of a CDP (which caters for rapid ingestion of new data and provides an analytical and measurement workbench).
Ultimately, the combination of a CDP beside a DMP reduces the familiar risk of chasing our own customers online to promote a product they have either purchased – or simply do not want.
It’s more useful to consider a CDP as sitting upstream from a DMP. It’s hugely valuable, consistent view of a customer can then be harnessed by the DMP to bolster its understanding of the customer base, to drive improved targeting. Ultimately, the combination of a CDP beside a DMP reduces the familiar risk of chasing our own customers online to promote a product they have either purchased – or simply do not want.
2: Data can be uploaded and analysed at lightning speed
*Health Warning: Jargon about to be used*
Most CDPs are built on a non-SQL platform such as Hadoop or MongoDB. For the uninitiated [non-data nerds] among you, this allows for multiple new feeds and frequent structural changes – so ‘unstructured’ and digital data of any description can be handled easily.
This means an end to the need to requesting quotes for new data feeds to your SCV from your development department – and the associated time taken up in the process. With the CDP, data can now be ingested and mobilised for analysis and decisioning in a matter of hours or days. Data heaven!
This level of speed and access, coupled with the richer insights that come with connected data, will bring joy to your Data Scientists and Analysts. After all, they can now mine customer and behavioural insights to drive more value for your business in a fraction of the time.
3: It’s open for business (literally)
Being an ‘open system’, the CDP can connect securely with all of your channels and tools, enabling a genuinely unified view of the data. So whether you use a DMP, email provider, call centre or a physical in-store terminal such as EPOS, can all be easily integrated by your customer data platform.
To do this, the tool/technology associated with each channel provides their native version of an identifier; this is matched to the CDP by the ID management using all available information across unlimited channels. At this point, a profile is returned. As a result, there’s less risk of ‘restarting’ conversations as your customer moves across channels (or as ordinary people say, “go about their daily lives…”).
This ‘open’ approach has another major benefit: attribution. Historically, this was a manual and siloed affair, generally undertaken across a partial mix of channels. However, using a CDP presents very best standard depicting accurate attribution of marketing spend and associated impact across all known channels.
4: Power to the Marketer!
Finally – and possibly the most important aspect to note: the CDP is constructed from its core to service the Marketing team.
With its ability to pull in and mobilise unrelated data a speed, there is far less reliance on the capabilities of your IT infrastructure to develop an instant view of the customer.
Pulling in every form of data has the added benefit of allowing us to test our more creative and lateral hypotheses within the business. For example, you can establish the impact of weather conditions, local events or even transport strikes on product sales.
In short, the Customer Data Platform is providing a level of power and control to the marketer to understand and communicate their customers – and crucially, report impact back to the business.
We should be aware that with all the benefits also come structural considerations. Matt Heffer, our Head of Marketing Technology, had this to say:
“CDP’s will enable organisations to organise around their customers, radically changing the way they approach ‘customer experience’. And while we should definitely be very enthused by this development, as you’d expect, this new technology will also deliver fresh challenges. Businesses must consider the demise of their siloed channel team structures – it’s time to stop debating which business unit (or channel) “owns” the customer.”
So to sum up, it looks like the CDP is here to stay. And while I do feel a sense of guilt for introducing you to another bland set of letters, I think it’s worth memorising these – they are set to make quite an impact on our world.
*The US marketing pioneer is often sighted for his quote: “Half the money I spend on advertising is wasted; the trouble is I don’t know which half”
**Deterministic data requires direct matching of two or more data sets (the first usually being PII/user-inputted) based on unique identifiers (e.g. Cookies or IDs). Conversely, Probabilistic data isn’t ID based (offering scale and overcoming privacy concerns) and requires interpreting with statistical analysis. Through ‘fuzzy’ and assumptive matching, we are therefore able to make connections using further non-identifiable data.