While much has been written to help define the category, the customer data platform (CDP) remains one of the most central yet least understood components of the marketing technology stack.
David Raab defined CDPs as “marketer-controlled systems that use persistent, cross-channel customer data to support external marketing execution” back in 2013 (at the time, largely in contrast to data management platforms, or DMPs, which were focused on harnessing anonymous data for ad-serving purposes). Raab now estimates the category will exceed $1B by 2019.
Whatever you wish to call them, customer data platforms provide a foundational data layer that connects marketing execution systems without replacing them. That last part is critical because, as Morgan Stanley has reported, fully 91% of marketing technology purchasers prefer best-in-breed, multi-vendor architectures to “all in one” stacks. Customer data platforms offer solution-neutral hubs, whereas multi-purpose suites, or “marketing clouds”, tend to be more hub-and-spoke.
In this sense, customer data platforms have been around for a while. Previously, though, these repositories were built largely in-house, or by system integrators. They were traditional data warehouses, requiring expensive ETL software and human resources to support and maintain. Suffice to say, these systems did not fit well within marketing budgets, timelines, or skill sets. Nor did they play nicely with the new breed of SaaS and media platforms marketers were using to engage with highly social and mobile audiences at an increasing volume and pace.
Enter the customer data platform. The customer data platform provides a single source of truth, accountability, and control while accelerating a company’s marketing strategies.
Here’s 10 questions to ask to identify the right customer data platform for your needs…
1. What am I trying to achieve?
Customer data platforms can enable a wide variety of possible end benefits. A typical part of the purchase rationale, in fact, is not only what a company plans to do with the platform right away but also “optionality” to do more over time. Nevertheless, it’s helpful to lock in on one macro driver that’s of primary importance to you and your business:
- Is it to improve operational efficiency (e.g. reduce the strain on analytics and engineering resources)?
- Is it to accelerate growth (e.g. enable new and higher quality customer engagements, faster)?
- Is it to mitigate risk (e.g. address a compliance need, or mitigate the risk vendor lock-in)?
The next set of questions are for your vendors. By definition, every customer data platform should be capable of doing three things: collect data from a variety of sources, control the data in different ways, and connect the data to executional tools.
Vendor Abilities: Data Connection
2. Is the data clean and reliable?
A customer data platform creates no value if it’s not capturing data properly. This is not a trivial point. Your customer data platform will rely on proper tagging, formatting, and structure to collect good data. Some vendors put that onus entirely on you, while others bring a host of tools, best practices, and playbooks to ensure that what’s collected is not merely “garbage in, garbage out.”
3. Are the data collection methods relevant to my business?
Be sure to probe into the data collection methods of each provider; this is the front lines of where success is achieved or lost.
For example, if you are an eCommerce company, eCommerce specific methods matter a great deal. If your users can create lists or set preferences, you will need a solution that handles more complex data structures with attribute arrays. If you’re mobile-first or just trying to be mobile-competent, make sure that you’re able to capture all native app data.
4. Are the data collection methods relevant to my engagement channels?
Customer data platforms like to tout their ability to assemble data and build shared profiles from every source imaginable — web, email, in-store, CRM and more. By now, most of these channels have fairly mature methods and APIs.
You also need to consider apps across connected devices. Even if your product is not experienced primarily on connected devices, this is where the future of rich customer data is being created, and you need an easy way to ingest and unify this data with data from your legacy channels. To future proof your business you should typically avoid data companies rooted in legacy paradigms as the data opportunity is usually a subset of the total opportunity. For example, you wouldn’t use an online marketing database to handle your web needs; similarly, you shouldn’t use a web solution to handle your mobile needs.
Vendor Abilities: Data Controls
5. Is the data governance and security suited to my business?
Every vendor will tell you their data is “secure” and most of them will be right. The question is really, how secure and in what context do they mean by this?
If you’re an independent app developer, you probably don’t need enterprise-grade features such as single sign-on, role-based permissioning, and heavy encryption, but if you work for a large company (or aspire to build one) you very much do.
To save on costs and make the code easier to maintain, certain customer data platforms combine lots of small datasets across multiple clients into a single database which should be a huge red flag to any enterprise customer. Meanwhile, other providers have been architected from the ground up to physically segregate customer data at the client level as well as support complex account hierarchies.
6. Can I control the who, what, where and how of data?
Data filters are an important feature of any customer data platform. They allow you to to get full control of your data and more easily manage privacy (e.g. limiting or preventing certain identity data from being sent) and cost (e.g. allowing you to send a sample set of data).
You should examine the granularity of these controls: can you filter at the event level (e.g. ticket purchases) or the attribute level (e.g. ticket purchases made in Boston). Can you filter by partner (e.g. sending each partner only the data they specifically need)? Can you sample and throttle data (e.g. sending a particular vendor a 5% sample of your data for testing purposes)?
7. Can I get access to all of my data, at any time?
The two most common reasons why you would want “all of your data” are for business intelligence or when you are switching vendors, including to a new data platform provider. A customer data platform provider should provide an easy way to “replay” all of your data into new systems to avoid cold-start problems, as well as be able to send data on an ongoing basis to the cloud data warehouse provider of your choice.
Vendor Abilities: Data Connection
8. Will I know more about my customer as a result of using this platform?
Some customer data platforms are essentially one-way streets, which collect data and send it out to executional systems of engagement. Others bring data back “in” to the platform from these systems of engagement for further analysis. Only a small number of customer data platforms go beyond “assembling” data and provide other value-added services such as profile enrichment (harnessing outside intelligence sources) or predictive modeling to reveal value-additive insights. Know the difference.
9. Can I connect to all the partners that matter?
Having hundreds or even thousands of integrations available doesn’t matter if they are not integrations with tools your business already uses, or might want to use. While there are a limitless number of open APIs available today, the best ones are often not open to the public and require certification. This is true of APIs from marketing clouds such as Salesforce, Oracle, IBM, and Adobe as well as media platforms such as Snapchat, Pinterest, and Instagram.
10. Are the integrations high-quality & action-oriented?
Piping data in and out of analytics systems is relatively straightforward. However, if you’re looking for a platform that can send data for the purpose of driving action, such as a personalized email or a mobile push notification, you need to dig in and understand specifically what the “integration” entails. Some platforms, for example, may claim they are integrated while in reality they are only integrated at the level of sending insights, not at the level that can trigger actions. To take a mobile-specific example, if the customer data platform is incapable of transferring push tokens, then it renders useless the executional system’s ability to send push notifications. Hamlet realized only too late, “My thoughts be bloody, or be nothing worth!” Don’t let this happen to you.
And Last But Not Least…
If you have done your diligence and asked these questions, you are almost ready to move forward with your decision. Don’t (yet). No matter what your findings are, you should never skip the step of checking references.
Check that the companies being put forward as references are comparable, and then call them! It’s critical that you speak live with customers and end users and ensure that their data needs are of comparable size and complexity to yours.Especially when it comes to mobile, there’s an enormous difference between businesses with a significant mobile presence and everyone else.
During your reference checks, probe also to see if their organizations match yours in terms of the rigor of legal and security requirements, and the diversity of marketing use cases.
Since the customer data platform will be responsible for sending data to partner services in your stack, you should also check references with those tools you plan to integrate. Have they worked with the particular data platform you have in mind before, and what was their experience?
After you’ve taken these steps, you will probably have narrowed down the field to one obvious solution. It’s time to call your rep, get the paperwork signed and start implementing!