The CDP, or Customer Data Platform, brings together all data sources and make the data directly accessible for marketers. You might still think of Customer Data Platforms as emerging, but there is a lot of experience in the market ever since David Raab first coined the term back in 2013.
You might consider selecting a great CDP, especially when first putting together a new Martech stack, or use a CDP modernize your current data infrastructure.
We asked 7 experts in the field of CDP and data-driven marketing to share their secrets and help those considering a CDP.
The results a brand can expect from a CDP
Being able to predict the results from implementing a CDP might be key to a successful implementation. We all like to see ROI on a MarTech investment. Now that quite a few companies have started using CDP’s, there is some experience in timelines and results.
So we asked our experts:
What changes can a company expect thanks to having a CDP in 3, 6, 12 and 24 months?
|David Raab, founder of CDP Institute:
What changes can a company expect thanks to having a CDP? Immediate changes will be less time spent on data preparation, followed by deeper analysis as analysts’ time is freed and as unified data becomes available.
By the twelve month mark, you should start to run marketing programs that use CDP data and be getting measurable business results. By 24 months, you should be ready to manage cross-channel programs with the CDP as the connector between different delivery systems.
|Todd Sherman, CMO at Amplero:
The trend of CDPs incorporating additional marketing and general business applications as part of their base offering has already started and will increase significantly.
Also important is easier integration with the wide range of platforms that can both take advantage of the data and contribute additional results back into the CDP.
|David Chinn, president and CMO of Lexer:
3 Months – Integrated IN and OUT-bound feeds of data, single customer view, enriched customer profiles, deep insights into the underlying drivers of sales performance (acquisition or retention), a strategy to act on the insights backed by a robust customer segmentation. The first campaigns in market on Paid & Owned channels and a benchmark for the performance uplift.
6 Months – Highly targeted marketing is now operational. Stakeholders are empowered to self-serve insight and execute personalized campaigns in significantly reduced timeframes. Customer data and insight becomes something that precedes campaign strategy, informing what campaigns are run rather than finding the best audience for an existing campaign. Acquisition audiences in Paid Media are tuned to high LTV looks alikes targeted with relevant products providing a materially reduced CPA and higher ROAS. Customer are being intelligently engaged on both Paid & Owned channels with relevant cross and up sell offers to improve LTV. Low value or Churned customers are suppressed out of audiences to reduce media waste.
12 Months – Increased focus on the customer leads the development of rich personas and differentiated products and services tailored to those customers.
Improved operational efficiencies are creating opportunities for cost saving initiatives or new value-generating activities. Data science and analysts are spending less time on low value list queries and are performing more sophisticated customer analytics.
Marketers are more agile, testing and learning in each campaign and have increased campaign volume and audience numbers.
The efforts around acquiring high LTV prospects and targeted cross & upsell activities have increased customer LTV growth providing the business confidence to move away from ROAS based acquisition investment. The business now takes a longer view providing a CPA cap optimized to each customer segment first-year LTV significantly increasing acquisition volume.
24 Months – Companies are more creative and adventurous with customer data. They are capturing additional information from customers earlier in their engagement in order to offer differentiated experiences.
Automation is a key strategy to create additional bandwidth for further innovation and development.
|Jordie van Rijn, Email marketing expert and Martech selection consultant:
A first implementation of a CDP is often what is called a lighthouse project. An initiative that is expected to have a signal effect for numerous following projects.
A CDP can be used to shake things up! And brand that are on the road to change towards more data-driven marketing or restructuring of current efforts around a central database.
Start small, with one or two use cases that have a direct impact on profits or internal processes and can be implemented quickly. Quickly can be quite quickly, like a few months. This will help to increase internal support for data-driven marketing as a whole.
You can communicate it as “…everything after that will be gravy.”
By now having the infrastructure in place, the first use cases can be built out, or more uses added. You can expect, like with all change, some resistance to organizational and workflow changes at first.
|Kelly J Waffle, Head of Digital Strategy at Hinge Marketing:
Depending on the channels used in their multi-channel campaign and the average sales cycle, a company should see increased campaign results in advertising and website engagement at the shorter end of your timelines. Results from social and email typically take more time.
|Todd Belcher, Founder of CDP resource:
Organisations can expect organizational change. Lights will be shining in areas that lights have not been shining in.
Hopefully, new data products as well.
CDP use cases to get excited about!
You have to use data to get value from it A CDP has Data Accessibility by Design, meaning a CDP brings together data sources and opens up the direct use to marketers. The data can be put to work in a lot of ways, but there are cases where a CDP really shines.
So we asked our experts:
“What are CDP use cases or examples you’re enthusiastic about?”
We are really interested in understanding how discounting can be used as a lever to drive sales growth and to influence buying behavior. Using the CDP to calculate the mean discount rates per customer and then analyzing customer value and profitability is hugely impactful.
One customer of ours just segmented their customers by mean discount rate and used this for targeting during a current sale. They direct their customers to the products and categories that match their discount threshold.
|Kelly J Waffle:
A news media company got great results with a CDP-driven approach. As you know, most of these companies struggle to maintain, let alone grow, their subscriber base–as there are so many options now to get news updates.
They wanted millennials to relate and subscribe, but had an aging publication. With the CDP, it created an awareness campaign specifically to university students by targeting:
A predictive engagement scoring model is used to serve up offers based on a reader’s subscription status.
Using behavioral scoring, this company started pushing out content to readers based on preferred relevancy and consumption format. It also created content hubs based on specific topics driven from modeling. The campaign also ran segmented ads on Facebook, Google, and Snapchat.
With data and a CDP as the “hub,” you are able to connect “spokes” with email, mobile, social, website, and advertising to dramatically build awareness and generate new subscriptions. I am beginning to see similar stories in retail, financial services, and other industries.
Some CDPs have put significant effort into making it easier for brands to connect to data sources on the input side and also the marketing automation platforms. That approach adds considerable value to existing customer – and prospect – related workflows.
Also interesting is the identity resolution and the speed with which it can be accomplished, that also adds value to those campaign workflows.
We’ve talked to a lot of customers over the last year who are either implementing CDPs or planning for it. They often turn the discussion to “how will they drive value from the CDP” and “how quickly can we do that” once it is in place.
This comes down to being able to harvest the vast amounts of data in an efficient and effective way and quickly leads to the idea of using some form of AI or machine learning to accomplish that — something that can understand the data and either put it directly to use or make it a lot easier to understand and use. They are looking to decrease the time to value.
|Jeremy Muras, Director Digital at Lion Capital, former VP of Digital Commerce for Burberry
Longer and deeper discounting now seems to be ubiquitous across retail and our portfolio is no different. By identifying audience sensitivity to discounts we hope to be able to be more targeted or surgical with our promotional activity.
A good example is PerriconeMD that bundles offers and promotions like most skin care brands. Customers acquired via different channels may be value orientated. We want to identify our best, highest value loyal customers and bring them closer to the brands. While also surprising and delighting them with appropriate offers.
For our value customers we continue to provide great value, but look to structure content and offers based of what we know from our CDP to improve their experience and value.
For everyone else in the middle we target them more closely to increase purchase frequency and by reducing the amount of generic promotions also protect the brand from the risk of being overly promotional.
In terms of customer acquisition, there is also an opportunity for our brands to be more efficient and effective in how we acquire new customers. With our CDP we can define and identify high lifetime value audiences and their characteristics and build audiences via performance channels that resemble these real customers. This makes our marketing spend more efficient as we become more targeted, but also ensure we focus on acquiring the most valuable customers.
Finally, where we have brands that are looking to extend their product offering out off their core categories in to adjacent categories, we want to evolve messaging and target based on what we learn from customers who have or are making that transition. For example, a brand that sells primarily denim and is now focused on building a broader lifestyle business.
|Jordie van Rijn:
I am a big advocate of keeping campaigns as simple as possible, (but not simpler). Yet some campaigns are just too important not to experiment, go a bit more complex and test several approaches.
I like the example of how HelloFresh used a Customer Data Platform for their winback campaigns. They could use many channels for that, but which combination is most effective?
For example: Facebook retargeting via custom audiences, email marketing, mobile push notifications, SMS, their own site(s) and integrated personalized offline letters.
Through the use of CDP they were able to coordinate the different channels, and do cross-channel campaign testing. They tested several channels against each other and also different combinations (like the winning combination with email + direct mail). Just setting this up would have been an enormous effort without a CDP.
I’m enthusiastic about all CDP use cases, of course. But I’d say matching identities from online to offline channels in retail may be the most exciting, because it’s so important and so difficult to do without a CDP. I should make clear that this is a very hard thing that only some CDPs can do internally. Others rely on external services. What all CDPs do in this context is give you a place to store the linked data – without which, making the linkage has very little value.
I am enthusiastic about data scientists and marketers working together. There are many use cases, but I think an area that gets overlooked with MarTech hype is the hard work that data science and marketing teams do to carefully tune models and the use of models that feed into dynamic segmentation.
Hopefully, new data products as well.
How to know if a CDP will live up to its promise?
Many companies might be considering a CDP implementation, but not really sure that the envisioned results will be met. A CDP can be quite an investment, so not something to take lightly.
To help all those that are in that situation, we asked our experts:
Is there any way to test / find out if a CDP will bring the spoils expected? What would be a good way?
The CDP plan should have a number of business goals – along with planned ROI-type analyses – to make sure the design and implementation work per the overall strategy.
|Jordie van Rijn:
It is not only the question if a CDP could bring you the expected value. But also what the best CDP is to bring you the value (and maybe more). Some might be a better fit than others, so that is what to look at during a selection.
I’d be diligent because the tools can be quite expensive and impactful. Make sure you have your use cases and a business case worked out at the start of the software selection to a rather detailed level.
Ask the CDP vendor for references of companies that have done a similar thing or had a similar set-up as yours and talk with the references about their implementation. This will give you a better notion of what to expect and many marketers would be glad to share their experience with you.
Make sure to pick a CDP vendor with strong credentials who can show the business case for your company in the beginning of your relationship. We, for example, load all the customer, order, email and product data into the CDP and analyze it to uncover the drivers of sales and growth opportunities to act on in the first 30 days.
Be realistic, get your best people to put together a 1, 3, 5+ year plan, and hire an impartial consultant to review it.
|Kelly J Waffle:
I would baseline the campaign performance of various connected systems such as email, mobile, social, website, advertising, commerce, and more. If possible, I would start out with some sort of pilot CDP program.
You want to have a clear idea of how hard it is to build and access the unified database.
But most of the value comes from using the data in analysis and marketing programs. This takes additional time and results can rarely be estimated in advance. It’s important to realize that CDP is a foundational technology, like your Web CMS or email system, which only makes it possible to do things that create value rather than creating value directly. You might have some cost savings from users spending less time on data preparation and extracts, but that is a very small piece of the picture and not the reason most people buy a CDP.
When should having a CDP be a priority?
Any customer-centric company would love to get all that customer data together in one place. So that it can be used to run campaigns, and get insights on customers’ behaviors and needs. Some say that it is critical for companies serious about data-driven marketing to be successful. But there are a lot of balls to keep in the air for a company at any given time and strategy dictates to make clear choices.
So we asked our experts: When should having a CDP be a priority?
From the start and always. So much data is being generated through many different activities that there needs to be a way to best take advantage of it. Competition in almost every industry is becoming fiercer and the ability to understand and act on data is already a significant differentiator; that will only increase over time.
We have worked with all types of companies at a variety of life stages: start up, high growth ecommerce, challenged wholesale moving to DTC and mature omni-channel.
The greatest determinant of success is willingness. As long as the company is committed to change it should be a priority.
|Kelly J Waffle:
Messy, siloed data is still a major marketing problem and over one-third of marketers say their inability to integrate data is the biggest impediment to the success of their analytics team.
CDPs can help, but most CDPs and CDP users are at the early stages of maturity so implementation has its own set of considerations that may affect various teams or the whole company. The questions go on and on. I would recommend that you make up a detailed list of considerations and questions and have the team answer them honestly.
I think a CDP implementation can be a priority after a team can positively address all the considerations and can dedicate the time, resources, and budget to make the implementation successful. Make time to plan for success with a strategy. Don’t just jump to execution.
A CDP should be a priority when you have disconnected data that is preventing you from giving customers the unified treatment they expect. Interactions across channels are the most obvious application: for example, suppressing customers from email re-targeting campaigns if they’ve purchased a product in a store or on the Web site.
This is very common integration for CDPs that is often not otherwise possible. Connecting online channels to call center is another high priority task that a CDP can facilitate. Problems with data quality for training predictive modeling systems is another priority application.
When not to invest in a Customer Data Platform?
Depending on your situation, there are plenty of reasons which now might not be the time to dedicate budget towards a Customer Data Platform. There is a contrast between the importance of data-driven marketing versus the importance of technology, so we asked our CDP experts:
What is a case where a CDP add little to no value, even if the company thinks it would?
Small hyper targeted audiences in Facebook is a challenge. The Facebook Ads API needs 1000 people to match to create an audience. Which can mean sending 2k customers to get a 1k match. If the volume is too small the audience will be rejected.
There are a number of cases where the CDP effort is an IT-driven project and is not as immediately connected specific business workflows. While that may be the way to kick off such a CDP project, the end-use goals should be clearer so that decisions made along the way can more efficiently deliver on them.
|Jordie van Rijn:
Say you have the business case in order and calculated that it would be wise to go look for a CDP. One of the most important pieces is to see if your (internal) organization is ready for it.
Is there enough buy-in to select – implement and execute on the CDP and the projects you were thinking to use it for. Ridiculously important and often misunderstood. Especially in organizations that have been struggling with their data projects in the past, internal support can be low (especially from people who lived through projects managed badly and disappointed in the past) or highly bureaucratic organizations. A big red flag for trouble down the line, think twice!
Are we leapfrogging a bit too far ahead? If the basics aren’t in place yet, it might be too early to change. In the order of things, get the basics in place first. Sometimes that means to get the data and tech structure arranged as a foundation. Often it means to get more product and services out the door first (just market and sell).
I also read in some articles about GDPR / data regulation and subsequent data management being a valid reason for data centralization with a CDP. It’s hard to make that case. If it would be the only reason for your organization to centralize your data, I’d not go there.
We still see confusion between CDP and DMP, especially in Europe. If you really need a tool to feed advertising audiences to exchanges and DSPs, a CDP is not the right solution. (There is one company, Paris-based Mediarithmics, that makes a credible claim to being both a CDP and DMP. They use proprietary database technology to make this possible.)
Any time a CDP is doing something that another separate category of technology is built around, you have to pause. Question whether or not the CDP is enabling or squashing access to “best of breed” technologies. To pick something specific though, I would head towards privacy and security. It’s not that the CDP adds little to no value, it’s that the value added is minimal relative to the importance of organizational approach to privacy and security.