On August 22, the CDP Institute hosted a pair of roundtable discussions, one for APAC and one for US and Europe.  Each featured a guest speaker, who made a brief presentation followed by questions and answers.  Here are notes from the events.

APAC Roundtable: State of CDP in APAC

Guest speaker: Jayesh Easwaramony, Spectra Global and APAC CDP Maestros Collective.  Jayesh brings in over 18 years of experience in APAC and runs Spectra Global, a specialist firm that helps companies convert data into revenue by implementing customer data platforms to unlock growth. He is a well respected CDP expert and has experience across multiple domains – adtech, data, martech, telco and media.

Discussion summary:

CDP started in retail/ecommerce, then expanded to media/publishers, especially at digital-first companies where the focus is on cookieless first party data.  Next came digitalization in insurance companies, then banks, who realize their old systems are too inflexible; telcos, who have high data volumes; automotive/industrial; and fast-moving consumer goods (FMCG).  FMCG firms are global so CDP used in APAC is often selected elsewhere.

CDP in APAC heavily used by publishers to develop advertising audiences as part of the cookieless transition, to replace DMP.  Don’t see CDP used for paid media optimization except in Australia.  Now, greater use of martech based on known identifiers.  Strong use case for retargeting, especially in apps where you can reach the customer without cookies, and for sending segments to media to build acquisition models.  Clients in Europe are more educated and interested in paid media optimization and retargeting.  Acquisition use cases alignment with media agency: data regulation by brand, media buying by agency.

Very low use of data clean rooms in APAC.  Most people see clean room as a collaboration tool, not a replacement for third party cookies.  Exception is Australia where Netflix is using it for collaboration.  Expect will be more due to retail media and OTT.  Much less maturity in SEA and India.

Both marketing and tech/data teams need education on value of CDP.  Marketers are confused between CDP and marketing automation and Google Analytics for media.  Data and tech teams would rather avoid the CDP and instead use AWS, DataBricks, Snowflake data lakes and AI/ML models, possibly with composable CDP solutions.  These conflicting roadmaps slow the buying decision.  Need to find common ground, based on use cases including customer 360, personalization, real time triggers, omnichannel experiences. 

Buyers in marketing should synch their RFP with data architecture and strategy teams.  Important to have everyone who will manage and use the CDP involved in vendor selection.  IT teams look for API and backend features with heavy code-based operation, while marketers look for UI-driven operations.  Important for both to be in synch.  Vendors should deal with both teams.  IT tends to bloat the project because they don’t understand the agility the CDP offers, and are not interested in a small project.  They have to understand CDP is a different type of tool.  Sometimes IT just wants to get marketing off their back by giving them what they want. 

Understanding use cases is critical to success.  Companies like banks and others with good CRM have a good understanding of their needs.  Those that don’t often will get caught during implementation when things don’t work as they expected.  Outside consultants can help if they’re specialists and they focus on the areas of weakness in an organization, such as data quality and governance.  Often need to use different consultants for different tasks. 

Biggest trap is managing business change.  People assume there will be little change from previous marketing automation or DMP systems.  Companies often struggle once they pass the simple baseline use cases because the organization isn’t ready for the change.

Often hard to get a purchase decision in APAC because need board approval for $1 to $2 million purchase.  To get approval, revenue growth is strongest argument.  Optimization use cases get more scrutiny because it’s harder to prove you need it.  CDP has to fit into existing stack and show quick time to value. 

Many companies want CDP as service (SaaS) rather than purchased software.  CDP not locally hosted is a big blocker, especially for banks and telcos; although there is now more comfort with cloud instead of. on-premises.  Hybrid CDP is becoming more popular, where some parts of CDP are running on hyperscaler platforms (AWS, Google Cloud, Microsoft Azure) or they do heavy lifting on warehouse and export ‘benign’ segments to CDP.  Could build entire solution on hyperscaler but each deployment will be custom so is harder.  It’s also a problem to maintain a hyperscaler system since data and engineering teams go to next project, leaving few resources for continuing work of APIs and connectors.  Hyperscaler solutions also don’t get complete dashboards and end-to-end conversion tracking.  In any deployment, need constant measuring of results from new use cases. 

Cost of services is often greater than cost of CDP software.   Often need to increase campaign velocity to get value from CDP, and there may not be enough people to do that.  Service providers can help to align data and marketing teams.  Agencies sometimes take over most of the work for their clients, especially tech arms of big holding companies.

Is confusion about whether marketing automation is sufficient.  Companies with large marketing teams that separate segment marketing from product marketing recognize need for data to feed segmentation and journey optimization.  Also need separate analytics product to get richer data.  CDP can be the single source of truth for marketing, collecting data from analytics and marketing automation, and marketing automation reads CDP data and is just an activation platform.

EU/US Roundtable: “What I learned from auditing dozens of RealCDP systems”

Guest speaker: Vernon Tirey, CDP Advisors.  Vernon runs the CDP Institute’s RealCDP certification process.  He is a former CEO, CRO, CMO, and entrepreneur, who works with industry leading brands, agencies, and martech vendors including Schwab, Allstate, Apple, Nestle, Adidas, Acxiom, IBM, and the US Mint, to address CX performance and digital transformation.

Discussion summary:

History of RealCDP: dates to 2018, when CDP Institute and Vernon set up certification program to distinguish CDP vendors from pretenders.  Vendor certification process starts with a detailed survey of 250 questions, then a phone call to discuss the survey, an audit call to demonstrate that answers are correct, and a draft audit report which is reviewed with the vendor.  Answers are adjusted and finally the report is published on CDPI website.  Now about 40 certified RealCDPs. 

Industry history started when customer data lists were built on mainframes.  Marketing, sales and customer service applications then began to build their own data stores to get more control and measure what’s working and not.  Result was fragmented customer data, which CDP is born to solve in mid-2010s.  CDP brings together all data sources to create 360 degree view of customer for multi-channel and then cross-channel marketing.  Privacy and real time have become more important over time.    Data enrichment has popped up in last two years. CDP are getting more data from data warehouses, and unlocking that data to be more available to users.  Some CDPs make good use of AI for identity resolution, but what matters really depends on the user’s needs. MDM can be part of a composable solution.  Data lakes let users create different views, including separate MDM and CDP views.

Above the data layer is the shared services layer, which includes reporting, model building, robust analytics, personalization, cross-channel journey orchestration.  Journeys stretch from acquisition to purchase to multiple purchase to loyalty.  Content management services are growing because few companies do that well.

Last layer is engagement.  Many CDPs offer email services, some do digital marketing, web, ecommerce, offline sales integration.  Some CDPs include marketing automation, especially those that started out in that field and then added CDP capabilities.  Not clear whether companies are replacing stand-alone marketing automation investments with CDPs.  This is more likely to happen at smaller companies that can take advantage of CDPs with good marketing automation features even if they are not the most advanced.   Harder for giant enterprises to change marketing automation vendors.

There’s no best CDP because what’s best depends on the needs of the buyer.  Different CDPs have different capabilities: some are strong for B2B, small companies, data management, big enterprises.  Composable CDP fits CDP into services-oriented architecture which has been around for 20 years.  Core capabilities for ingesting, processing, storing, and privacy are similar across vendors.  Differences are in data management such as ID resolution and data hygiene.  Also differentiate in whether they have organic BI solutions and model building, journey orchestration, robustness of messaging such as email, SMS, web, and mobile purchase.  Differences relate to whether they started from engagement and moved down to CDP, or started from the data level and came up.  About 1/3 of CDPs in market today provide real time personalization.  This breaks down into real time data capture, real time ID resolution, and real time decisioning such as best offers updated during the course of an interaction.  About 20 vendors can do that.

Beyond features and functions, buyers need to know technical and business requirements, how you will do implementation (yourself or CDP vendor), how to assure adoption, how to assure operations and support.

First and second generation CDPs often being replaced with a new CDP.  Often they were never effectively implemented but in some cases the original technology did not work well.

Growth of martech has created many tools that are not integrated.  CDPs sit at the center of the martech stack and offer large amount of pre-integrated functionality, creating more simplicity.  There’s a danger that composability could result in greater complexity and fragmentation, or higher cost to configure and integrate separate tools.  Composable CDPs are pushing data but not doing two-integration with activation systems, which limits ability to do real time decisioning in omnichannel marketing.  Major composable vendors have introduced new modules for ID resolution, ingestion, and analytics to provide a more than just reverse ETL.  Buyers have to make sure those integrate well and they are mature enough to meet user needs.  By contrast, traditional CDPs have now reached a good state of maturity.

CDP provides grounded data to feed into AI, to avoid hallucinations.  We’re preparing the data infrastructure for real-time AI agents, which we already see in customer service.  In the next two years, expect customer agents will do the shopping and research for consumers. Customer agents will package information for buyers.  Most AI today is machine learning, such as packaged propensity models for purchase propensity and next best offer.  These are available in over half the CDPs.  Generative AI is starting to be integrated.  Agents are mostly coming from LLM companies including Meta, Google, Microsoft, not CDP vendors with a couple of exceptions.  Salesforce released an open source LLM recently.

CDPs can help with data privacy because they have had to build in security and privacy compliance, and the vendors have experts on staff who can provide advice.  Third party cookie deprecation isn’t necessarily a doomsday scenario because many companies are pursuing solutions.  Clean rooms let companies add third party and second party data to their first party data, to get a broader view.

CDP is changing division of labor within companies.  For journey orchestration, used to be separate email, text, advertising, direct mail teams working with their own data and measuring their own results.   Now there needs to be a central customer journey program design group thinking strategically to manage journeys across channels.  Also is growing role for marketing operations, which should be kept in marketing, not IT.