What is a Customer Data Platform?
Customer Data Platforms are sophisticated repositories and sorting-houses for multiple data sources. A CDP takes data in, organises it, and prepares it for analysis and actioning.
The result is a comprehensive, accessible and actionable record of your customers’ online and offline interactions at an individual level, that can then be used to enhance your outbound campaigns. In other words, the highly-prized Single Customer View.
Although the actual term “Customer Data Platform” was first coined in 2013, the concept of unifying disparate strands of data within an organisation is much older. Initially, these were relatively unsophisticated databases that existed within a piece of marketing automation or personalisation software.
The basic concept of the SCV, however, actually goes all the way back to the 90s, with databases that organised customer information into a simple hierarchical structure; postcodes->households->individuals. So you effectively have individuals living at a house, and houses residing in postcodes. Alchemetrics was the first company to launch this to the UK market in 1996.
The usefulness of these databases led to standalone CDPs, which have evolved to become capable of exchanging data sets between the modern marketer’s ever-growing range of sources, engagement channels and analytics systems.
This image by Gartner does well to illustrate the core common features of the modern CDP products available.
Taken individually, these features all exist in some shape or another within other tools, or in standalone products. It is the combination of them, in real-time, that makes CDPs so valuable to both marketers and CX teams alike. Let’s break down each of them in more detail, as per Gartner’s definition.
1. Data Collection
This relates to real-time ingestion of multiple data types from a variety of sources, including behavioural, transactional and biographical. It should be possible to set strict limits on how long data is held and made available for processing.
2. Profile Unification
One area where CDPs really excel in comparison to other examples of data management software is cross-device identification (XDID) in the instances of users who interact with brands on multiple devices. By using Matching Elements (MELs), such as an email address, the CDP must be able to identify that it is the same individual user in order to avoid profile duplication. This is the basis of the Single Customer View.
Matching data is, in fact, the complicated aspect of the entire process, one that sets a true SCV apart from what many multichannel marketing hubs offer. A true CDP is one which allows multiple MELs to be used either as a combined string or via a cycled matching loop where you match on X first, then Y next, then Z at the end. MELs can be anything from third party systems IDs, personal contact info and even in some cases dates of birth, table numbers, booking times etc.
However, with the changes that came in under GDPR the use of date of birth, table numbers etc. needs to be handled very carefully and not used for the purposes of directly referencing a recent dining experience in a restaurant for example in any form of messaging. Such data should really only be used as indicators for decisioning and insight.
This is by no means a feature unique to CDPs. Most Mobile Marketing Platforms and Multichannel Marketing Hubs will offer some level of segmentation, albeit to varying degrees. Still, a CDP should enrich your segmentation capabilities, combining various data types to create “micro-segments” of customers for advanced targeting.
Of course, all of this rich information is useless in a vacuum. CDPs must also be seamlessly integrated with your “activation arm”. This could be a single channel point solution (e.g. your email service) or it might be your mobile marketing platform. However, it could also be your actual website, eCommerce solution or anything “customer-facing”.
5. Prediction & Decisioning (optional)
One element that Gartner doesn’t view as essential is some sort of decisioning component. This is likely because the algorithms powering this capability are well beyond the scope of the average CDP.
At the very least, however, you should be looking for a CDP that is capable of layering in techniques such as propensity models or next-best-action recommendations as they will become increasingly commonplace. Additionally, it is often possible for a CDP to deliver information to a call-centre or online chat agents.
When researching a CDP that does not have a decisioning component, be sure to investigate the openness of its integration capabilities. This will allow you to plug in your existing A.I./decisioning engine, or allow you to do so in time. As technology continues to advance and the cost of these engines falls, they will become more and more commonplace. Make sure you are future-proof!
Commonly used terms related to Customer Data Platforms
- Data silos
Unintentionally isolated sources of data. They lead to organisational inefficiencies, sub-optimal engagement campaigns and wasted resources.
- Cross-Device Identification (XDID)
This is very important. It involves identifying a customer across all of the channels and devices they use to engage with a brand in order to paint a holistic view of their activity. XDID is essential to the Single Customer View. Also referred to simply as Profile Unification.
- Keys / Matching Elements (MELs)/ Matching Keys
In the context of profile unification and customer data, a ‘key” is the specific piece of information that we use to identify an individual. An email address, for example, is a very useful key because a customer will use it to log into their account across different channels (website and mobile app). Any PII (as below) could be used as a key.
- Personally Identifiable Information (PII)
Any piece of data that can be used to identify a specific individual. Under GDPR, this covers a wide range of data points. Home address, email address and phone numbers are the most common examples of PII. It can also be an IP address, login ID, geolocation etc.
- Customer attribute
Any piece of information known about a customer. e.g. location, name, app version used, date of birth, content preferences etc.
- Customer profile
A collection of all known attributes about an individual customer.
- Preference / Permission Centre
These are on-site and in-app pages where a customer can self-curate the types of content/offers they want to receive. They can also opt in or out of marketing campaigns. This important for GDPR, as covered later in this handbook.
What types of data does a CDP process?
So far, we’ve talked a lot about data in a general sense. Let’s look at the specific data types that a CDP handles.
This is internal data collected directly from your own properties (website, mobile app, PoS). It includes transactions, behaviour (views, clicks), biographical information and anything that would typically be part of a detailed customer profile. It’s the essential raw material for personalisation and one to one marketing.
First-party data is used to offer highly accurate “next-best-action” recommendations for segments of your customer database. It’s the most valuable data you can have, whether it’s past purchases or specific product interests, and helps you to increase wallet-share and retain customers for longer.
This may well be a new concept to some readers, but in time everyone will be talking about Zero-party data!
Essentially, this is data that is proactively shared by the customer and relates to their preferences and future purchase intentions. With so many marketers and CRM professionals attempting to predict the desires and goals of their customers, this type of data can be transformative for an engagement strategy.
So how do you collect zero-party data? You just need to ask for it! We talk all the time about the “value-exchange”, usually in the context of increasing your marketing subscriptions. If you want a customer to share something of such personal importance and obvious value, then you must offer them something of value in return.
In the pre-digital age, surveys were a fantastic, though laborious, method of collecting data. In return for conceding personal insights, they might be entered into a competition or given some sort of discount.
We often work with clients to build a preference centre (either on-site or in-app), where customers can self-serve and choose the kinds of communications they wish to receive. The end result is happier, more engaged customers and reduced churn and unsubscribes.
With so many seamless and convenient digital channels available, there are ample opportunities to open a real dialogue with your consumers and gather data quid pro quo. And a CDP is the best place to store this data, which can be matched up with that customer’s wider profile for a well-rounded picture of the historical behaviour and future intentions. A very powerful combination!
This is externally-sourced data, primarily on prospects but certainly customers too. One way of looking at it is that it’s another brand’s first-party data.
There are actually several second-party data marketplaces in operation, and their popularity is growing. With such a prominent focus on privacy laws (and rightly so) your alarm bells are probably ringing already!
Be very wary of where this data is coming from. We recommend that you only source data from trusted partners who you know have permission to share it. It’s really not worth taking a risk on this!
What about third-party data?
Third-party data is increasingly seen as a relic of a bygone digital age.
This is the kind of anonymous data used to feed acquisition at the top of your funnel through ad targeting campaigns. It’s anonymous, behavioural information, usually purchased through one of the massive aggregators out there or collected via tracking cookies. It’s typically used to find “look-alike” prospects whose broad demographic profile matches your current customers.
We strongly advise that you never ingest third-party data into a Customer Data Platform. Given the importance of GDPR compliance (and the growing digital privacy regulations elsewhere in the world), the value of 3rd party data is not worth the risk. In any case, many organisations and consumers now frown upon its usage.
In a few years, this may not even be an issue, with the likelihood that all third-party data will be phased out of collection and banned across social media platforms (where it is now most commonly used).
Why are CDPs so popular right now?
There has been a lot spoken about CDPs in the last couple of years, with Gartner asserting that inquiries from their clients on the topic doubled between 2017 and 2018 (Source requires Gartner subscription). The trend continued throughout 2019 and into 2020. They are not new, however!
When multichannel marketing began to really take shape in the mid-2000s, brands suddenly found themselves with a plethora of data sources and campaign analytics from the various customer engagement channels to contend with. And so the CDP was born to unify them.
Why the sudden surge in interest then? Good question. Three reasons.
1) Firstly, there is an increased focus amongst C-level executives on customer engagement and a drive to create “brand champions”.
Look at the graph below. Positive advocacy leads to more sales.
With more choices available to the consumer than ever, loyalty is the golden egg for revenue officers. Customer experience is the goose that lays it. And unified customer data is the corn the goose feeds on.
Alright, enough with this metaphor, but you get the point, I hope!
2) Secondly, as we mentioned earlier, the volume of customer data that brands now have is overwhelming. There are multiple platforms, more and more touchpoints, online and offline transactions to record and an endless flow of behavioural analytics at every stage of the funnel.
All of this has created a network of data silos. That’s a major problem. And a CDP is the best way of solving it.
3) And finally, GDPR. With many organisations still not prepared for the full obligations of data stewardship, CDPs offer a straightforward route to compliance. By consolidating all of the data currently stored about an individual in one place, as well as a comprehensive record of their marketing permissions, CDPs allow you to quickly and easily service Subject Access Requests. This amounts to modifying, deletion or export of data as per the customer’s request.
Is a CDP all that different from DMP and CRM software?
There’s a lot of acronyms out there and you could easily be forgiven for confusing them!
A CDP is not the same as a Data Management Platform or Customer Relationship Management software. But there are certainly similarities and, in a way, both DMPs and CRMs are fore-runners to CDPs.
CDPs vs CRMs
CRM software has been around since the 90s, primarily used by sales teams to coordinate their efforts and draw basic insights to help drive revenue. It really only stores biographical information on existing customers and prospects (name, location, purchase history and so on).
Typically, CRMs require a lot of manual input in order to get the best out of them. They are perfect for making notes on an individual customer’s relationship and for organising your sales pipeline in a place where all team members can see it.
Essentially, the primary differences between the two technologies boil down to this; CRMs are for sales teams and CDPs are for marketers.
CDPs vs DMPs
With the rise of online advertising in the 2000s, there was suddenly a need to capture anonymous-user data as well. DMPs were created instead to collect and segment “cookies” and IP addresses in order to optimise the targeting of digital ad campaigns. They only serve these kinds of anonymous acquisition campaigns, however, and won’t be of any use to you in terms of Customer Lifecycle Management.
Traditionally, DMPs don’t store Personal Identifiable Information (PII), the kind of rich first-party data that enables you to personalise campaigns and track customers journeys at a granular level. In other words, a DMP will not give you a true Single Customer View. This is a significant drawback! It should be noted that some DMP vendors are now expanding their capabilities to increase the range of data they can handle.
What’s the difference between a CDP and a Data Warehouse?
At this point, you may be wondering how a CDP differs from a data warehouse. If your organisation has already invested in a warehouse solution, you’re probably hoping that it can do a similar job. Unfortunately, that’s not the case.
A data warehouse does act as a repository, yes, and most can even handle the processing and cleaning side of things as well. However, they are not built to facilitate customer engagement. They do not have the ability to execute cross-channel customer identity resolution (the basis of a single customer view). Fundamentally, they are better suited to more general data analysis and high-level reporting.
CDPs are purposefully designed to enrich marketing and customer engagement campaigns. They allow you to leverage the data you have in order to maximise the ROI from your campaigns.
One of the biggest drawbacks of a warehouse in this regard is that the data is “batch-processed” (uploaded daily or weekly), whereas a CDP does this in real-time. This is vital for event-triggered engagements and mobile-first campaigning.
Finally, within the context of a brand’s technical architecture, CDPs are typically “marketing-owned”. Although some IT support will inevitably be necessary, particularly in the early days, the day-to-day users of a CDP are marketers and the best ones are designed with this in mind.
This chart is useful for a quick comparison between the various data management solutions available.
Which of these data management systems do I need?
As you have seen, CDPs really are superior in terms of the scope of what they can do, but that doesn’t mean that another system couldn’t work for your organisation. It just depends on the use cases you have in mind and what you hope to achieve long-term.
Here are the typical “out of the box” capabilities of the main four types of tools, alongside an explanation of why they might be important to you.
Cross-Device Identity Resolution: Does the system possess the ability to match a customer’s identifying details across channels and devices to create a Single Customer View?
All Data Supported: Some systems are better suited to handling either first-party (CRMs) or third-party data (DMPs), but CDPs and DWs should be capable of both.
Lifecycle Management: Is the system useful at all stages of the funnel and does it lend itself to lifecycle campaigns?
Real-time Processing: Does the system process information as it is received or in timed batches?
Ownership: Which department typically owns the system day-to-day; sales, marketing or IT? As we’ve noted elsewhere, CDPs are built with marketers in mind
Activation & Engagement: Does the system seamlessly integrate with the vast majority of engagement channels?
What are the use cases for a CDP?
Putting all of this theory aside for a moment, on a practical day-to-day level what will a CDP help you to achieve? Below are some industry-agnostic examples of how our clients use it to deliver tangible ROI. Primarily, each of them involves in-depth, real-time audience preparation (or segmentation) based on the comprehensive data stored within each unified profile. Because these profiles are pieced together from a variety of rich first-party data sources (including behavioural, transactional etc), it allows for greater creativity, and accuracy, when composing a segment.
By setting the right combination of attributes, you can speak to an “audience of one”, and deliver a personalised message at any stage of the customer’s lifecycle.
- Reducing churn
It’s not simply businesses operating on a recurring subscription model that must constantly monitor their rate-of-churn. Any brand that is reliant upon repeat purchases, or a publisher that wants to drive visitors back on-site time after time, can use a CDP to identify users that are at risk of never returning.
- Cross-sell and upsell opportunities
From a revenue perspective, one of the biggest appeals of a CDP is the potential to upsell to existing customers. With all of your data centralised, it’s easier to identify and action cross-sell opportunities. These typically come in the form of personalised product recommendations (with “customers who bought this item also bought…” messaging), but within the travel and tourism industry, for example, we have clients who upsell additional experiences (whether that’s car retinal, tickets tourist attractions or restaurant reservations) based on key attributes from the customer’s profile.
- Cross-channel marketing automation
One of the many drawbacks of having data siloed within each engagement tool is that it restricts your ability to create campaigns across the various channels you might be using. So your email campaigns have no awareness of what’s being sent via SMS or push notifications, for example. You run the risk of over-messaging your customers, which is not a good idea, but you also don’t get an understanding of which channels each customer prefers.
A CDP can give you that level of insight and more, acting as the intelligence engine (or brain) deciding which channel within your engagement stack to use and when.
- Enhancing customer loyalty and increasing wallet-share
According to the classic Paredo formula, 80% of your revenue from just 20% of your clients. If you were able to identify this 20%, and knew exactly which channels they preferred to hear from on you, what campaigns would you send them?
How can a CDP help me stay GDPR compliant?
Brands are swimming in highly sensitive private data, which they are responsible both for keeping safe and using only as permitted by each individual customer.
When you are storing data in multiple locations your ability to fulfil these obligations is compromised. Firstly, each storage system is a potential weak spot in your cybersecurity framework. A single, safe CDP reduces the chances of a data breach.
Of course, it is paramount that when assessing vendors you investigate their security credentials. Not all adhere to the same standard as the Xtremepush platform, which is ISO-certified and regularly penetration-tested.
Additionally, CDPs make it easier to update, delete and export an individual’s data should they ask you to do so. And having the data centralised also makes it easy to send deletion or update requests to any additional third-party systems in our ecosystem.
Under GDPR there is a clear requirement to provide all customers with a comprehensive record of every piece of data that you store on them. This has proven to be a time-consuming, difficult task for organisations with siloed data, one that is also prone to error.
Primarily thanks to the Single Customer View, CDPs simplify and expedite this process allowing marketing and customer support teams to act fast without the need to involve the IT department.
The next step for Customer Data Platforms and the marketing-tech industry
One thing we’ve pointed out a few times in this article is that CDPs by themselves are not enough to actually send a message to a single customer, let alone orchestrate an entire campaign.
Yes, they are incredible when it comes to preparing the data needed for personalised, targeted engagement, but they can’t execute it.
Up until now, marketers have needed to invest in some form of multichannel marketing tool (typically an MMP, MMH or a combination of both) in order to really benefit from a Customer Data Platform. As a leading vendor in both the mobile and multichannel marketing spaces, we have seamlessly integrated with a number of leading CDPs including Segment and Optimove.
We recognise however that the martech industry has been moving towards the emergence of One Platform, which combines the capabilities of various software products. CDPs will increasingly be absorbed into other technology categories either through mergers and acquisitions, or built in-house by vendors themselves.
We’re proud to be at the forefront of this. The Xtremepush customer engagement, personalisation and data platform is built around an open architecture that has allowed our development team to evolve its capabilities in line with, and in advance of, industry trends. Most recently, we have added CDP functionality to our offering, creating a seamless, unified solution for understanding, execution and decisioning.