In November 2019 perhaps you caught this article: “To CDP or NOT – 3 tips – then you decide.” The main takeaway – the CDP space is a quasi-market with a mixed bag of firms coming from different lineages and different levels of capability, maturity, and focus. The conclusion: buyer beware and standby.
That was BC – Before COVID-19. Since then, what hasn’t changed about the world? And like everything in 2020, the CDP market was not immune to upheaval. And although the basic premise for adding a CDP into the Martech stack is still the same:
- Help resolve customer identity
- Rationalize and manage customer data
- Make that data accessible to other systems
…what’s changed are the vendors involved, and their core and extended capabilities, which are substantially different 18 months later.
Most markets appear as nebulous categories, and the CDP market was no exception. But as buyers and vendors evolve, dust settles, and the picture becomes clearer. Still, two important aspects of what a CDP should supply loom large and are worthy of close inspection. Namely, providing customer recognition/identity management and distilling the right (and righteous) customer data into meaningful insights.
Considering those key features, let’s explore a few of the big changes since November 2019:
- The huge marketing cloud players entered the market: Adobe, Oracle, and Salesforce
- More consolidation took place, with small CDPs swallowed up by an interesting mix of companies
- Perilous new milestones reached for third-party cookies and stealth consumer tracking
Stick with me. You’ll get insight into these three changes, three tips, and some final thoughts.
Marketing Cloud Titans Enter CDP Fray
Adobe Real-Time CDP
In early 2020, Adobe entered the CDP ring with Adobe Experience Platform’s Real-Time CDP, promising to “Combine all individual and company data — internal and external, known and unknown — into a standard taxonomy that can be activated in real-time.”[i] A tall order indeed.
Although certainly set up to collect digital data by way of Adobe Launch & Analytics, Adobe’s aggressive mission to combine “all data” for B2C and B2B across known and unknown, lacks focus and gives reason for pause.
Adobe has fared well in providing digital marketing data & support for early-stage customer journey activity, with its first-generation web analytics and tag management (by way of its Omniture acquisition over 10 years ago), followed by its subsequent purchases of Demdex (third-party cookie data-management platform), and marketing automation firms like Neolane (B2C) and Marketo (B2B). Yet with the third-party cookie tracking foundation crumbling as the final browsers outlaw it, they’ve had to look for another way. So far, that appears to be using CNAME record cloaking, which in effect is just a clever DNS hack to circumvent gaining explicit permission to track.
The ultimate jury and judge (the consumer) may not approve of this tactic (once they discover it). Further, with a shortage of direct access to first-party behavior data, customer analytics depth, and channel breadth, Adobe still struggles to develop deep customer understanding and natively/performantly enrich its customer profile. And other than collecting raw digital data in real-time, not much else about Adobe’s CDP is real-time and insightful. Adobe nonetheless plows forward with bold statements of real-time and unity that potential CDP buyers should take with a grain of salt.
As a major marketing cloud player, Adobe will eventually amass more digital data, improve its signal detection, and get more apt at activating those signals and audiences in acceptable ways. But for now, buyers should beware of completeness claims, tracking practices, data feeds and speeds, and external integration features.
Interestingly, googling with the term “Oracle CDP” yields a top result pointing to an Oracle whitepaper-like webpage espousing that a “customer data platform (CDP) is software that collects and unifies first-party customer data.” [ii] So far so good.
Reading on, the article mentions “first-party data” 11 times, never mentioning third-party data until the final punchline at the end, where the author claims that a Customer Intelligence Platform (CIP) is different from a CDP because it “incorporates anonymous, third-party data as well as first-party data.” It’s here that Oracle tries to differentiate its CDP, Oracle Unity, from all others. That differentiation attempt falls flat, and is oddly fascinating on three fronts:
- Oracle has almost no choice but to take this approach, since it spent $400m on BlueKai in 2014, one of the world’s leading third-party data trackers. As such, Oracle wants the buyer to believe they get a premium from contracting with a CDP that can merge third-party data.
- Oracle claims it’s not really a CDP, but instead differentiates as a Customer Intelligence Platform (CIP), and not just for marketing. Amusingly, in my June 2019 article I advocate for a CIP – The Final 4: MarTech Platforms and Ecosystems – yet with the middle letter short for insights about individuals attainted from first-party data, not general intelligence. Very different CIPs indeed.
- The reason for the demise of the cookie-based cottage industry and third-party data is that it was built on a house of cookie cards, gathering and brokering consumers’ data without explicit permission, and inherently unreliable as a good proxy for consumer intent & behavior – one of the major tenants for a CDP.
Given this, be careful with Oracle’s CDP (or CIP) solution, with its bias toward third-party data, paid media channels, and early-stage acquisition use cases. Purchasing one means buying into the value of third-party data and acquisition use cases, while not solving for data-driven, real-time 1-1 customer engagement use cases, deeper into the relationship, on owned channels.
Late in 2019, as the virus was unknowingly spreading, Salesforce began spreading the news about its new Customer 360 Truth, claiming it had a product with “a new set of data and identity services that enable companies to build a single source of truth across all of their customer relationships.“ [iii] And although at the time they didn’t call it a CDP, they were quacking as if it were one, and funny enough in April 2021 relaunched it as a CDP. [iv]
In 2019, in classic Salesforce fashion, they announced a not-ready-for-prime-time CDP-like product, C360, with pages of fine print. Like a theater stage with a kitchen viewed from afar, it might have appeared fully equipped. However, on closer inspection, some of the supposed appliances were but props with no cords to plug in, no motor to run them.
And even on re-launch in May 2021, they simply slapped existing separate products such as Tableau and Mulesoft onto the wrapping paper of the Salesforce CDP. Further, like most CDPs (except ones that come from the web analytics space, such as Tealium and Celebrus) everything is based on creating customer segments and sharing those in less than real-time for activation instead of taking an individual personalization approach and sharing in real-time.
Thus, rip off the cartoon marketing wrappers, and look inside the box and inspect all the parts for function and fit before buying.
CDP Market Consolidation
In addition to the entrance of the above big three, Microsoft and SAP also announced CDP solutions. Before November 2019, 18 acquisitions took place. Since November 2019, 8 more further transformed the CDP landscape:
- IgnitionOne bought by Zeta Global – December 2019
- AgilOne bought by Acquia – December 2019
- Evergage bought by Salesforce – February 2020
- Segment bought by Twilio – October 2020
- Exponea bought by Bloomreach – January 2021
- BlueVenn bought by Upland Software – March 2021
- Boxever bought by Sitecore – March 2021
- Zaius bought by Optimizely/Episerver – March 2021
What’s the takeaway? Dust is still flying in this market. And if you are betting on one of the 100+ vendors calling themselves a CDP to plug key gaps, especially in foundational areas such as identity & data management, consider whether their future is secure, and they’ll continue to go in the same direction, as it could impact yours.
The Calamitous Cookie Crisis – Customer Identification and Tracking
In January 2020, Google announced plans to end support for third-party cookies in Chrome in two years. Late-breaking news is that in June 2021, Google said they will delay until the middle of 2023. But cookiepocalypse is still coming. With less than two years until that deadline, ad-tech companies, and ad agencies alike are scrambling to find workarounds for web behavior identification and tracking.
Case in point – The Trade Desk and ad agency Publicis (who bought the database marketing firm Epsilon in 2019) are teaming on a digital advertising solution built around the new open-source identification scheme called Unified ID 2.0. Initially developed by The Trade Desk, Unified ID 2.0 obfuscates a consumer’s email address, using a technical hashing technique to protect consumer privacy.[v]
As of May 2021, The Trade Desk says it already has over 170 million profiles obtained with consent. But long-term success depends on an even bigger pool of email addresses (e.g., more consumers opting in than opting out), and that means enough publishers adopting the standard, and obtaining consumer consent. Since history has shown consumers will opt-in without reading terms and conditions, it may have hope, especially in places like the US and Asia, so stay tuned. My advice – read before you click, as it’s essentially agreeing to be a target of every participating company.
In addition, SAP and Akamai bought traditional sign-on companies Gigya and Janrain respectively, going the route of obtaining social sign-on solutions to gain access to customer identification and tracking capabilities. And although Okta, who acquired rival Auth0 in May 2021, hasn’t called itself a CDP (yet), they are a force in the customer authentication and identity space.
What does this have to do with CDPs? Well many ad-tech companies, formerly calling themselves data-management platforms (DMPs) during the third-party cookie era, now claim to be CDPs. Keep in mind, however, they built their solutions to manage third-party data and cookies and to target based on these spurious methods, and not on first-party data and known identities. Ultimately, without a strong first-party data foundation, those DMP CDPs have a limited shelf-life and are poor investments.
CDP Selection Tips
Tip #1 – Study their specialty
Keep in mind that all vendors started with a core offering. That tells a lot about what they’re probably good at. When interviewing a job applicant, there’s a reason why we inspect someone’s background (work history, school they attended), as it gives insight into how they’ve honed their craft.
No vendor (not even the big ones) will be able to supply best-of-breed capabilities to handle all stages of a journey, from the anonymous browsing steps to phases deeper in an authenticated relationship. Nor will they be able to major in more than a handful of the dozen or so capabilities the collective CDP market covers:
- Data collection
- ETL – Extract, transform, load (including cleansing and householding)
- Identity stitching and management
- Real-time data insights
- Predictive analytics
- Recommendations and decisioning engines
- Journey (cross-channel) orchestration
- Owned channel marketing automation and e-message services
- Digital advertising
- General (business intelligence) customer data activation
- Internal query, reporting, dashboards, and attribution analysis
Most native CDPs came up focusing on one or more of the first 3. And with no official CDP magic quadrants or waves by major analyst firms, many others have conveniently slapped the CDP label on themselves. So, decide where you have the biggest capability gaps and needs along the customer journey, where a data-driven solution will drive better outcomes and more value, find matches, and select accordingly. Also, if gaps exist mainly in areas 4 – 10, look beyond the CDP market, as there are a multitude of vendors not calling themselves CDPs that major in these areas.
Tip #2 – Demand real-time response times
When considering the claim of “real-time,” (which is a critical capability to take CX to another level) look beyond single components, such as the speed of data collection, or placing data onto a customer profile record. Instead, inspect the entire data/event -> insights -> decision journey and ask:
- “Can that entire trip be accomplished in an SLA (Service Level Agreement) under 200 milliseconds?”
- “Can the vendor do that at scale, for millions of customers and 1000’s of interactions per second?”
Why 200 milliseconds you ask? Because as a consumer, do you want websites you use to be slower? As a person responsible for the website, will you allow anything new to slow down page loads? I bet the answer to both is no. So if your new CDP is going to play a role in providing better real-time digital experiences, it better not take up much of the two-tenths of a second response time budget.
Tip #3 – Demand real-time insights
Look for a CDP that can supply real-time data insights, with a library of these for your industry. This looms so largely in reaping unfair benefits from a CDP investment because not many CDPs do this, and it’s how you’ll move the needle on customer experience. Can you do this today? Can you find customer behavior diamonds in the deep mines of digital data, surface it, polish it, and immediately pass it to a customer decision hub? Not many can.
For instance, detecting consumers’ heightened but fleeting interest in specific products, refining that raw data into curated signals, passing them to a decision engine in real-time, so it can trigger special and immediate actions. Very few CDPs can do these things – in that order – fast enough. An example: a consumer on a banking website, researching mortgages [again] in the final stage of selecting a mortgage provider.
So, look for a CDP that can solve this problem. There aren’t many. You’ll add something special and unique that few can do. Celebrus is one solving this exact problem: collecting the right behavior data, making sense of it in the form of a signal library, passing those signals to a decision authority in real-time, so it can act in the moment.
Big is not always better, but it’s always bigger. And although selecting a large outfit as a CDP provider gives some assurance that the solution will be around in a few years, that doesn’t necessarily equate to the best CDP solution. Doing business with a mega-CDP vendor rarely means faster, more seamless interfaces and deep expertise. On the contrary, expect bigger integration costs, longer wait times, custom work, and more patience required. And if selecting a big CDP is for “one throat to choke,” try finding that elusive throat inside a tech behemoth with 50,000 employees who have swallowed up 20 companies on the way to building their marketing stack and CDP.
Conversely, using a smaller player has its tradeoffs. Besides the risk of being bought, or folding up, inevitably their capability focus will be esoteric. So, carefully inspect core competencies. Look for a CDP that compliantly tracks customers, collects data in real-time, has a signal library fit for purpose, and can interoperate with a decision hub. That way, you’ll get differentiation leading to better customer experiences from your CDP investment.
[i] Adobe.com, https://business.adobe.com/products/real-time-customer-data-platform/RTCDP.html, June 2021
[ii] Oracle.com, https://www.oracle.com/cx/customer-data-platform/what-is-cdp/, June 2021
[iii] Salesforce.com, https://www.salesforce.com/news/press-releases/2019/11/19/salesforce-announces-customer-360-truth-a-single-source-of-truth-for-every-customer-across-the-worlds-1-crm-2/, November 2019
[iv] Salesforce.com, https://www.salesforce.com/news/stories/salesforce-cdp-innovations-make-customer-interactions-smarter*/, May 2021
[v] The Wall Street Journal, https://www.wsj.com/articles/publicis-groupe-signs-on-to-use-trade-desks-alternative-to-cookies-11617883217, April 2021