Wednesday, July 31, 2019

Customer Data Platform (CDP) – Introduction and Market Overview

Are you interested in Customer Data Platforms? But, what are we talking about? We read about: Data, Insights & Actions everywhere nowadays. Data must be transformed into insights and those insights into actions. As a BI specialist, I was always a little dissatisfied with the possibilities of following an insight.
Datapriest CDP Data Insights Actions
The data is „trapped“ in a data warehouse, updated daily or at most hourly with a data model, following the usual marketing channel, affiliate and/or campaign hierarchy. On the other hand, there are many new activation options such as on-site personalization, push notifications or email automation. But all these new tools and channels are only as good as the underlying data. After some research I found what I was looking for, a so-called Customer Data Platform (CDP for short), a relatively new term in the marketing tech environment.

CDP vs. DMP vs. CRM System vs. Marketing Cloud

If you are as naive as me and believe that there is a common definition for Customer Data Platform, you will soon be disappointed. There are very different views (as always) of what a CDP is and what distinguishes it from other systems such as CRM systems, DMP’s or Marketing Clouds. The cause of the many definitions is mostly from the technical origin of the various CDP vendors. Whether it’s a web analytics tool, tag management system, classic CRM system or campaign management tool, all now offer a „360 degree customer view“, but of course they have different features and strengths depending on the starting point.
Datapriest Customer Data Platform Architecture
From my point of view, the following features characterize a Customer Data Platform:
  • specifically developed software solution for the goal of „360 degree customer view“
  • different first-party data sources combined in a common, consistent storage
  • based on a customer-centric data model
  • consisting of anonymous customer profiles, known leads and existing customers
  • synchronization and matching of anonymous IDs and internal customer IDs
  • segmentation can be done by platform users (i.e. CRM Manager)
  • open system, which can supply data to other BI systems, ad networks or marketing tools
  • campaign execution tool built in to activate these segments
Since the term Customer Data Platform is still quite new, I would like to distinguish CDP from existing systems.
DMP: Data Management Platforms usually work with third-party data, that is with anonymous profiles based on cookies, usually with a short lifetime (i.e. 90 days). Most of the activation takes place in ad networks via a Demand Side Platform (DSP). Here too, the boundaries are blurred and many vendors now integrate first-party data. Examples of DMP solutions are Mediamath, Adobe Audience Manager, Lotame or Adform.
CRM system: Customer Relationship Management Systems (like CDP’s) focus on the customer profile. Customers have different features, touchpoints and transactions. However, as a rule, no customer journeys are mapped by anonymous users in a CRM system, and the activation options do not extend beyond email and the telephone. Many CRM systems also focus on B2B, i.e. business with companies, not with end customers. Examples of CRM systems are Hubspot, Salesforce, Zoho, Microsoft Dynamics or SugarCRM.
Marketing cloud: Here the demarcation is not always easy and you have to dig deeper. Marketing cloud vendors such as Adobe have the challenge of integrating a wide variety of (mostly acquired) solutions. Here, customer IDs are often exchanged between modules without having to do their own persistent data storage. Since these systems are not always equipped in the base with the 360 ​​customer view, they usually require an internal or external IT effort. For this they offer a lot of application and activation possibilities. Examples include Adobe Marketing Cloud, Salesforce Marketing Cloud, Oracle Marketing Cloud, IBM Marketing Cloud or Marketo.
To summarize, one can comfortably say that a Customer Data Platform has its purpose and the use cases it provides make sense, but the CDP can not always replace all other existing systems directly.

The 360 ​​Degree Customer View

As a rule, the following data is imported into a Customer Data Platform:
    • Clickstream data: Web or App Tracking Data including Pageviews, Sessions, etc.
    • Behavioral data: Event tracking including „Add to Basket“, „Checkout Steps“, „PDP Views“
    • Customer data: name, email address, telephone, address, gender, date of birth, …
    • Transaction data: purchases and product data
    • Campaign Data: UTMs, Sent Newsletter/Opens/Clicks
    • Customer Care Data: Live Chats, Support Tickets, …
    • Offline Data: Shop Visits, Shop Purchases, …
    • Enriched data: e.g. Socio-demographic data from Acxiom, RFM models, Net Promoter Score, surveys

Why is a 360 Degree Customer View Needed?

If we start from the classic e-commerce business model, we do not have an infinite number of levers to turn. A Customer Data Platform can optimize all 3 relevant areas of the e-commerce business model:
Datapriest CDP Retain Acquire Convert
  • Customer Acquisition: Lower Customer Acquisition Costs (CAC): A wide variety of CDP audiences can be synchronized with ad networks such as AdWords and Facebook for Include, Exclude, or Lookalike audiences to reduce CAC. Targeted advertising across multiple touchpoints also aims to reduce the very costs of the customer journey itself.
  • Customer Conversion: Better Conversion Rates (CR): Through onsite personalization and the right actions at the right time, the onsite conversion rate can be improved.
  • Customer Retention: Higher Customer Lifetime Value (CLV): While customer acquisition costs quickly reach their minimum limits on the marketing side, many companies have the potential for customer lifetime value. Monetization of their customer base through targeted CRM measures are not enough.

Application Scenarios of a Customer Data Platform

Enough with the theory. Hopefully nobody would get the idea to set up a CDP without knowing which specific use cases they want to utilize, in order to create business value. Here are a few ideas and suggestions:
  • Facebook Custom Audiences: Whether retargeting specific customer segments or creating lookalike audiences, Facebook is great for sharing data with a CDP.
  • AdWords/Criteo Custom Audiences: The exchange is also exciting for SEA and display. For example, granular excludents can provide more positive ROAS (Return on AdSpent).
  • Email-Automation:email, SMS, push messages or even offline mailings, will all require a level of creativity. Which customer segments need to be activated when and with which measures? Whether software vendors in a test phase or just before contract renewal, email automation is the key to higher customer lifetime value and is not utilized enough. A classic case here is the Abandoned Shopping Cart mail.
  • Newsletter: A “one size fits all” approach to newsletters is a thing of the past. Nowadays, it’s possible to insert personalization and product recommendation widgets into email newsletters.
  • Onsite Personalization: There are no limits. Different website content, depending on location, purchase history, respected products, channels, click behavior, devices,etc. Also, upselling through referrals or the delivery of dynamic banners that automatically optimize per customer, fall into this category.
  • Customer Care Infobase: Actually, after a few clicks, we already know quite a bit about our anonymous visitors. Why not give this information to the Customer Care Team when the user contacts you via live chat or phone. The team can offer higher quality service to the customer.
  • Voucher:: You should not always offer each user a voucher, minimizing margins. Why not keep the value of the voucher dynamic and only offer those users a discount where they have a positive influence on the purchase decision.
  • Onsite Advisory: Ask questions and use recommendation engines to advise the customer onsite and deliver value for the customer.
  • Internal Process Automation: Avoid sending csv files via email, slack & co. Automate the daily data exchange between departments and companies with API capabilities.
  • Product Data Feeds: Create Dynamic Remarketing Ads with products, the visitor saw on your website. Facebook, AdWords and Criteo need product data feeds for this usecase.

Define Requirements for a Customer Data Platform and Make the Right Decisions

In my view, it would be wrong to say, that vendor XYZ will fit for every company. The vendor which is most suitable depends on a number of factors that each company weighs differently. I’ve always been a fan of classic vendor selection. The following steps are necessary for this.
Datapriest CDP Vendor Selection
  • 1. Define concrete use cases, functional (features) and non-functional requirements with all stakeholders and departments (so there are no nasty surprises after the decision of colleague X who asks if feature Y is possible).
  • 2. Market Research: Make a list of all relevant market participants. If rough requirements are not met at first glance, the vendor already falls out.
  • 3. Vendor Matching: Match the leftover vendors with your needs and select 3-5 companies to be shortlisted. Be in contact with these vendors.
  • 4. Vendor Deep Dive: Benchmark these vendors against your requirements. Take a look at the product presentations, product demos and trial versions. Get offers now.
  • 5. Final Decision

Here Are Some Suggestions for Possible Requirements of a Customer Data Platform

Data Collection:
  • the CDP must realize its own first-party data generation (own tracking)
  • the CDP must use the tracking data from vendor XYZ
  • the CDP must capture every single pageview of the visitor
  • the CDP needs to be able to process campaign parameters
  • the CDP must be able to capture custom events
  • the CDP must provide a data export for our data warehouse
Data Integration:
  • the CDP must be able to integrate data from the backend system (MySQL) via SQL access
  • the CDP must be able to import data via CSV/JSON/XML
  • the CDP must have a connector for tool XYZ (i.e. Mailing Tool)
  • the CDP must be able to receive data via API
  • the CDP must be able to receive data via JavaScript tracking on the frontend
Data Storage:
  • the CDP must hold tracking data for X months
  • the CDP must hold customer data for X months
  • data storage must take place in the EU area (data protection)
  • the data must be quickly retrievable (response time and data freshness)
  • data storage must be technically scalable for potential growth
  • the data must be exportable (Data Ownership)
  • the segmentation of customers should be able to be made according to schema XYZ
  • the data points can be selected when segmenting a dataset.
  • the ability to select multiple data point fields when segmenting.
  • the segmentation engine should provide a preview of audience size
User Activation:
  • the creation of complex automation should follow a flow view
  • the CDP has its own mail editor/ESP
  • channels XYZ must be connected via connectors
  • the activation of the audience should take place for channel X directly in the tool
  • the reporting of campaigns should be in real-time
  • the CDP must have A/B testing capabilities
  • the CDP must be able to use recommendation engines
  • the CDP should have attribution modeling capabilities
  • the CDP provides clear reporting/dashboards
  • the CDP gives you the opportunity to answer individual questions
Non-Functional Requirements:
  • the monthly budget of X Euro should not be exceeded
  • the CDP is performant
  • the CDP is user friendly
  • the CDP has good documentation
  • the CDP manufacturer has good support
  • the CDP manufacturer is from the EU
  • the CDP has no negative impact on page speed
  • the CDP has real-time architecture