Downloaded 135 times
Uploaded byTreasure Data, Inc.
Introduction to Customer Data Platforms
The document discusses the importance of Customer Data Platforms (CDPs) in bridging the data gap businesses face regarding customer insights and marketing effectiveness. It highlights the challenges organizations encounter in creating a single customer view, the need for unified data management, and the proven advantages of utilizing CDPs over traditional solutions. Furthermore, it emphasizes the necessity of real-time data access and the ability to connect various data sources to enhance customer engagement and marketing strategies.
More Related Content
![[DSC Europe 22] Lakehouse architecture with Delta Lake and Databricks - Draga...](https://cdn.slidesharecdn.com/ss_thumbnails/draganberic-lakehousearchitecturewithdeltalakeanddatabricks-221130080712-6e817e95-thumbnail.jpg?width=640&height=640&fit=bounds)
Similar to Introduction to Customer Data Platforms
![[Webinar] The Best Kept Marketing Secret to Achieving Complete Customer Views](https://cdn.slidesharecdn.com/ss_thumbnails/115definitiveguidetocdpswebinar1-210312170259-thumbnail.jpg?width=640&height=640&fit=bounds)
Introduction to Customer Data Platforms
- 1. INTRODUCTION TO CUSTOMER DATAPLATFORMS 0 1 / 1 1 / 2 0 1 7 D A V I D R A A B / R A A B A S S O C I A T E S
- 2. WELCOME TO THEDATA GAP 94% 14% Feel SCV is important Shared SCV in place Source: CDP Institute, 2016
- 3. WELCOME TO THEDATA GAP 74% 75% 76% 84% 14% 12% 13% 10% 0% 20% 40% 60% 80% 100% Matching customer across devices Understanding Customer behavior over time Tailoring messaging by channel Associating conversion events with marketing “Very important to growth” Strong capability Unified Customer Data: Needs Outstrip Capabilities Source: Econsultancy, Customer Recognition: How Marketing is Failing its Top Priority, August 2016
- 4. WELCOME TO THEDATA GAP Obstacles to Marketing Success 58% 49% 39% 34% 32% 26% 13% Single Customer View Collaborate Outside Marketing Budget for Martech Budget Over-all Train Staff on Martech Organizational Barriers Buy Latest Martech
- 5. WHAT IS UNIFIEDDATA, EXACTLY? 72% 57% 42% 33% 14% 0% 20% 40% 60% 80% 100% Combined customer data Central customer database Central customer database Central customer database Shared customer database Unified Data in Place Source: CDP Institute, 2016
- 6. WHAT IS UNIFIEDDATA, EXACTLY? Users for Single Customer View 70% 65% 51% 37% 34% 14% 13% Personalization Customer Insights Measure Across Channels Custom Offers Custom Service Consistent Treatments Loyalty Programs Source: CDP Institute, 2016
- 7. UNIFIED DATA REQUIREMENTS ProcessFlow for Unified Customer Data
- 8. UNIFIED DATA REQUIREMENTS ProcessFlow for Unified Customer Data
- 9. UNIFIED DATA REQUIREMENTS ProcessFlow for Unified Customer Data
- 10. UNIFIED DATA REQUIREMENTS ProcessFlow for Unified Customer Data
- 11. OPTIONS Architectures for UnifiedCustomer Data
- 12. CDP DEFINITION •Marketer-controlled systemthat •Builds a unified, persistent customer database •Accessible to external systems.
- 13. OTHER COMMON CDPFEATURES Cloud deployment Software as a Service NoSQL data stores Real time access Supplemental applications
- 14. REALITY CHECK Source: CDPInstitute, 2016 Obstacles to Building Single Customer View 41% 39% 31% 29% 29% 29% 17% 15% 14% Budget Extract from Source Organizational Roadblocks Other Priorities in IT Other Priorities in… Systems Can't Use Skills to Build Skills to Use Technology
- 15. IT’S NOT SNAKEOIL CDP Industry Statistics • 2,000 installations • $350 million revenue • $700 million funding • 24+ vendors
- 16. SUMMARY Customer Data Gapis Real Meeting all SCV goals needs unified, shared database (with direct access – not an EDW) Conventional solutions fall short CDP is a better, proven alternative
- 18. IN THE DIGITALECONOMY, IT’S WINNER TAKES ALL What do Google, Uber and Amazon have in common? 64% 87% of the global digital ad revenue of the US ride-sharing Amazon > everyone else combined
- 19. CHALLENGES: DISCONNECTED, DELAYED,INACCESSIBLE DISCONNECTED Most valuable data is outside your firewall, siloed in cloud services and line of business apps. DELAYED The half-life of your data is short. 54% of data’s business value is lost in two hours. INACCESSIBLE Gartner predicts 90% of deployed data lakes to be useless by 2018. $$$ Time YOU DATA
- 20. LIVE DATA ISTHE KEY TO A UNIFIED CUSTOMER VIEW CONNECTED Most valuable insights come from connecting data across multiple sources. CURRENT Act on customer data within minutes. Keep all historical data current to see the whole picture of each customer. EASILY ACCESSIBLE From marketing to sales, data must be everyone’s fingertips. Democratize your data, mobilize your teams. ✔ ✔ ✔ ✔
- 21. ALTERNATIVE: LIVE DATAMANAGEMENT
- 22. GREAT CUSTOMERS Gaming Media& AdTech Auto & Industrial IoT E-Commerce & Lifestyle
- 23. • Various onlineand offline data silos captured partial views of customers • Unable to leverage existing marketing technology due to a limited understanding of buying signals • Flexible platform needed to satisfy multiple stakeholders with varying skill sets, allowing granular, event-by-event data access. Pains POS DATA MEMBER DB AD PLACEMENT DATA
- 25. “Blasting emails toeveryone who tried samples or bought a particular product won’t lead to customer delight. Detecting a mood swing in a customer and changing the tone of push notifications does.” Kenji Yoshimoto, Lead Direct Marketing Analyst 20% BOOST IN CUSTOMER LIFETIME VALUE Among Customer Loyalty Program Menbers
- 26.
Editor's Notes
- #3 Nearly all marketers want a single customer view but few have it.
- #4 Not just CDPI – other survey, same answer
- #5 Is it really important? Or just something on a wish list? On our survey, it was #1 obstacle to marketing success. So, yeah, it’s important.
- #6 Let’s define this. When we asked people in the survey about the current state of their data, we got very different answers to slightly different questions. Basically three answers: 72% had some kind of combined data, including data that was shared directly between systems without a central database. From 33% to 57% had some type of central customer database, although we don’t know how they used it. Just 14% specifically said they had a central customer database that was shared by multiple systems.
- #7 What’s interesting is to keep that previous slide in mind when you look at how marketers expect to actually use the single customer view. The most common goal was personalization – something that needs just a little customer data which could be easily copied into the execution systems without directly connecting to a shared database. The next two, customer insights and measurement across channels, are analytical applications that do need a rich central database but don’t need it connected to execution systems. Custom offers also involves relatively small data volumes that could easily be copied to execution systems. It’s only when you get down to customer service, consistent treatments, and loyalty programs that you really need direct access to the shared database. So marketers actually have a pretty nuanced vision of what they want to do and, at least implicitly, are going after the low hanging fruit first.
- #8 So we say that maybe companies are dipping their toes into the unified view pool rather than jumping in head first. Still, to do the really good stuff – those consistent customer treatments – they do need a complete solution, which is to say a fully shared central database. Let’s look at what they need to build that. We can divide that into roughly three tasks
- #9 The first is loading – or, if you prefer, ingesting – data from multiple sources. A few key features here Connect to lots of systems Connect to lots of data types Check for quality Pre-process to make it usable Accommodate changes: new sources, new data elements, new data types
- #10 The second task is actually processing that data once you’ve loaded it. This is where most of the magic happens. Specific requirements are Standardize and transform so data is consistent and classified Link identities to find the same person across systems; that’s what unified data is about but it’s very complicated Do preliminary work to simplify access; that could be calculating trends or aggregates, finding patterns over time, or creating indexes. That’s another thing that can be complicated to do well.
- #11 Finally, we get to the part about making the data accessible. Tasks may include Formatting for real time access, which often means putting it in different tables from the main data store and organizing so data about each customer is already brought together Formatting it for analytical purposes, which may involve a different structure from real time access, since most analytical projects involve subsets of data about groups of customers rather than all data about a single customer Loading it into other databases that external systems know how to query or can read via APIs Publishing the metadata and other access tools that the external systems need This can be pretty technical but it makes a huge difference in whether the data is actually usable
- #12 Several ways to meet those requirements. Silos don’t work Data hub = moving data between systems; works with a few systems but creates redundancy and allows inconsistency. More for cross-system processes than single customer view. Doesn’t address identity resolution or need for historical data met by persistence. Examples: Boomi, Jitterbit, Zapier Data warehouse = loading all data into central database; is right general solution but EDWs are mostly for analysis; don’t usually support real time updates or access. Typically big corporate IT projects that take years if they ever get done. Marketing cloud / suite: combine customer-facing systems in single system with unified database. Great theory but in practice suites are built from acquisitions and lightly unified. May have table to link identifiers and maybe some very skinny profiles e.g. for personalization. Doesn’t allow easy analysis or realtime updates of events across systems. Anyway, most marketers use products from multiple vendors.
- #13 - marketer-controlled: not an EDW - unified, persistent customer database: unifies identities, stores at least some, includes identified customers [needed to unify]; not a DMP - accessible to external systems: not built for specific application (although many CDPs do have apps)
- #15 CDP’s don’t solve everything - Still need budget (although it’s cheaper) - Can’t fix existing infrastructure (to extract data or use results) - Organizational roadblocks and priorities - Some skills required
- #17 Summary - Customer Data Gap is Real - Meeting all SCV goals needs unified, shared database (with direct access – not an EDW) - Conventional solutions fall short - CDP is a better, proven alternative
- #18 Thank you so much. David talked about how a single customer view is the greatest unmet need that prevents companies from accomplishing many of their marketing goals. Treasure Data is a platform built to address that need. Let’s take a little bit of a closer look at what I mean.
- #19 This idea of a single customer view, as David said, is about the ability to unify data. And as we all know, companies that get this right outcompete the rest.
- #20 As companies rise to the challenge of competing in the digital economy, they naturally adopt SaaS tools to solve various problems. And a lot of these tools are fantastic. But they aren’t always exactly clambering to make their data connected and accessible. There’s a tendency, as anyone knows who’s tried to connect them, for SaaS tools to become siloed point solutions. This results in fragmented data and dependency of marketing teams on engineering. Treasure Data was formed to meet this challenge.
- #21 This gave rise to the idea of Live Data Management. Live Data is data that is connected, current, and easily accessible.
- #22 We wanted to make it possible for companies to outsource Live Data Management so that they can compete with the data giants, without the need to hire armies of data engineers to do so. Treasure Data was built from the ground up for this purpose. It enables companies to unify all their data sources into a single customer view that can be owned by any organization in the enterprise, and it provides shared, self-service analytics, with access to a unified view of all the customer data, to anyone who needs it.
- #23 We have a roster of amazing customers including Warner Brothers, Subaru, Pioneer, and Toyota. I want to zoom in on one case study that shows the power of Live Data Management in action.
- #24 Shiseido is the fourth largest cosmetics company in the world. In 2012 they formed a new website, Watashi Plus, that enabled customers to receive expert health and beauty advice and product recommendations. They truly had a vision of a personal relationship with their customers that would deliver a higher level of service across all channels, but they were blocked by the difficulty of achieving a single customer view.
- #25 Treasure Data was able to take all of these disparate signals from their first party data and automatically put them together into a single customer database. This data was then enriched with data from their partners, along with third-party data, to give them extremely detailed information to deliver information to their customers that was exquisitely sensitive to the moment-by-moment needs of their customer.
- #26 This enabled Shiseido to achieve their vision. Instead of marketing automation, they were after Customer Preference Management. Improvements in their advertising and customer communications resulted in a 20% boost in Customer Lifetime Value among their Customer Loyalty Program Members. Kenji Yoshimoto, the Lead Analyst on Shiseido’s Direct Marketing team, had this to say: (read quote)
- #27 So that’s a little bit about how Treasure Data’s Live Data Management Platform provides a powerful Customer Data Platform, and now I want to invite you to ask any questions you may have.