Discover new ways of connecting with a CDP
A collection of learnings, thoughts and tips from my Salesforce Marketing Cloud (SFMC) Customer Data Platform (CDP) certification journey.
We as consumers invest a lot in a brand.
And yes we all know the brand is not a living identify, but for our time commitment wouldn't it be nice to get a little bit of:-
Yes, [replace-with-your-fav-brand-name] I get you and I understand where you are at and together let's work out some next steps to shared awesomeness!
Well, that is where a customer data platform (CDP) fits in.
The purpose of this article is to discuss important CDP considerations from a developers perspective, it is written in such a way to appeal to a broad audience for anyone interested to learn more.
Key takeaways:
- CDP is about customer data integration, identity, AI, speed & experience management quickly and at scale.
- Even without a CDP platform, you can start preparing with a CDP mindset now.
- Hyper personalisation and connecting in the moment helps to deliver engaging experiences.
Let's get started.
CDP is about customer data integration, identity, AI, speed & experience management
Historically data used for digital direct marketing has been sourced from disparate systems. The issue with this approach for organising customer data is that information about a customer can be siloed and not easily accessed.
Moving forward to today, organisations both big and small, engage with consumers through multiple channels and have rich consumer data flowing and available from e-commerce channels, cloud based platforms and apps.
The opportunity for business is to best engage with consumers in the right channel, with the right message at the right time.
Enter the CDP.
The idea behind a CDP approach is to make the data from all systems highly available for making data driven decisions that allow for hyper personalised and engaging communications.
The 5 forces driving the adoption of a CDP
There are 5 forces driving the adoption of a CDP.
These consist of customer data integration, identity, AI, speed & experience management. Data integration and identity is about the ability to receive and stitch custom data from many platforms into a single view of that one unique customer. AI, speed and experience is about learning from past behaviours to deliver a more meaningful message with the customer at the right moment.
Consider the following to understand the differences between a CDP, versus a no CDP approach.
- Customer Data Integration - with CDP data comes from the platform whereas with no CDP data comes from the channel.
- Customer Data Identity - with CDP the audience is a person whereas with no CDP the audience is the channel.
- Customer Data AI - with CDP AI is making personalised recommendations when they make the most sense whereas with no CDP AI is reacting to past events.
- Customer Data Speed - with CDP next communication in the best consumer engagement moment whereas with no CDP next engagement can be days or hourly.
- Customer Data Experience Management - with CDP marketing communications are driven by experiences whereas with no CDP marketing communications are driven by batch prospecting playbooks.
Even without a CDP platform, you can start preparing with a CDP mindset now
STOP & READ THIS FIRST => There is nothing new about the idea of a customer data platform (CDP), in fact what drives CDP is what has and will always drive a data driven marketer, which is to - EFFECTIVELY AND EFFICIENTLY UNDERSTAND AND ENGAGE WITH THE CUSTOMER.
Even if you are not yet ready to implement a CDP, understanding the CDP framework can help if for nothing else than to provide a best practice for organising your Marketing Cloud platform. Following are a collection of thoughts, important considerations and tips for implementing a CDP that are all worth a consideration regardless of where you are on your data driven digital direct marketing journey.
1 - Start
- First things first - get your team ready.
- Identify stakeholders - who are the decision makers, and who are the marketing managers and data specialists that will be impacted by the implementation of the CDP?
- Identify users - these are the people that will ultimately turn the CDP into reality and will typically include a CDP admin, potentially some of those stakeholders - e.g. marketing managers and specialists.
Simply watching TV, ordering takeaway or making that next big purchase involves a myriad of customer data points consisting: what were sales messages prior to purchase, purchase options made, location, weather, personal preferences, and the list goes on.
TAKEAWAY => Before you start preparing for a CDP, become data aware - everything you do, ergo the customer creates data.
2 - CDP Ingredients (Inputs)
Organise what systems your are are going to get data from.
The ingredients for a CDP are all the various data that are integrated together and typically consist of the following:
- Customer relationship management (CRM) system data - components typically consist of contact data with unique identifier (e.g. email, mobile number and/or account id), along with non-unique identifiers like name and address.
- Engagement data - components typically consist of data about the contact that is stored in marketing databases, e-commerce systems and with an email service provider (e.g. open, click usage data, time on page, etc.).
- Point of sale (POS) data - components typically consist of purchase history stored in accounts and ecommerce systems.
- Preference data - components typically consist of the special ingredients contained in customer preference centre systems.
- Service interaction data - components typically consist of data about problems and issues derived from customer service systems.
- Behavioural data - components typically consist of the more granular customer details that are available from various systems and apps.
- Other important data - data that is used to make informed decisions.
For each of these CDP ingredients (data inputs) there are 5 key characteristics that are going to be important to consider. According to the Salesforce Marketing Cloud (SFMC) CDP there are the five to be on the lookout for, which consist of the 5 Vs namely:
- Velocity - from a CDP perspective can ingest file or streaming data at any speed. In practice when preparing for a CDP this means that the data must be available for automatic consumption, for example should be available on SFTP or AWS as a file that can be streamed into the CDP or be accessible via an API.
- Variety - from a CDP perspective can clean, sanitise and transform data, then map reformatted sources to a standardised CDP data model. In practice when preparing for a CDP this means that the data must be structured data that aligns with a schema.
- Veracity - from a CDP perspective can provision customer unification and provide a system of truth for identifying customers. In practice when preparing for a CDP there must be an identifier on each data input record that can be used to relate it to all other data in the CDP.
- Volume - from a CDP perspective can store data at virtually unlimited scale. In practice when preparing for a CDP there should be nothing stopping data from flowing into the CDP at scale.
- Value - from a CDP perspective can connect for personalised real time engagement with a marketing, advertising or analytics platform. This is about how value flows out from the CDP to the intended audience. In practice this means that connected downstream systems must have scalable interfaces to allow messaging and communication as required.
For a CDP to be successful, it is required that all stakeholders are onboard. In practice data will need to be in the CDP for decisioning, therefore the goals and priorities among all teams and departments need to align with this requirement.
TAKEAWAY => Corporate data silos are a threat to CDP success.
For a CDP to be a single source of truth (SSOT) and able to make good recommendations it is always important to have a correct picture of how the customer relates to other data in the system.
TAKEAWAY => Customer identifiers are required across all data inputs.
3 - CDP Recipe (Strategy)
The CDP recipe (or strategy) is about how all the inputs are combined and consist of the following:
- Unique customer identifier - key for making it all work is that there is a unique customer identifier for each record inputted into the CDP. The identifier has been mentioned before, and is mentioned again because it is just so important. Data in a CDP must relate to other data in the CDP. If there is no plain or compilable identifier on the data record that relates the record to the whole, then there is no value for that data to be brought into the CDP. Often there will be a number of strategies for creating a unified view of the the customer and ordering rules of most important identifier should be applied to determine the order in which customer identifiers are considered.
- System integration - mapping inputs to the CDP unified data model allows the focus to be on drawing insights from data for marketing purposes. In the case of the SFMC CPD, first consider preparing an inventory along with a data dictionary of all data inputs and how they will be ingested to map to the CDP data model. Next, these data inputs are mapped to data streams. Then finally, the data streams are mapped to the standard customer 360 data model. When customising the data model careful consideration should be given to ensure that key attributes of the model are defined in such a way as that multiple inputs can be mapped to them.
- Segmentation - consider all the smaller specific target audiences of your universe of customers. Segmentation allows you to target messages across channels. Segmentation depends on data insights which can be inferred from formulas applied to single or aggregate values calculated from many records contained within the unified CDP data model.
- Automation - use automated processes to make it easy to continually keep the data in your CDP up-to-date.
- Tap into everyone's expertise and be ready to change - refine your strategy with learnings from your customers and people in your organisation that know how to engage their audiences. Information to collect is what products and services customers have liked in the past.
Inputs get mapped to the CDP data model, therefore it is important to know your CDP data model. Ensure that all inputs are mapped correctly to the CDP data model and also when setting up inputs, ensure that the input fields have the same types as their counterparts in the data model.
TAKEAWAY => Get CDP data model mapping correct first time because making changes can be a significant undertaking.
Inputs get read by people so use familiar words for what IDs that people using the CDP will be able to relate to.
TAKEAWAY => Only use an ID as a label when it can be easily understood by the person reading.
4 - CDP Product (Output)
Continuing the food preparation analogy, the CDP is the oven that mixes all the ingredients and the food that is produced is consumed by your audience across email, mobile, social and digital channels.
This is where you mix up your ingredients to produce an engaging customer experience. Mixing all these who, what, why and whens together into one place allows marketers to do promotions to customers at the right time in the right way and that is where the magic really starts to happen.
The final step is to deliver personalised messaging using your customer preferences and segments on all channels.
The best practices sequence for preparing segments consists of first preparation, then data availability, data cleansing, data validation and finally publishing.
Calculated segmentation use cases that can be inferred from the CDP data model include:
- Customer Lifetime Value (LTV) - The total of what the customer has spent from when they first became a customer.
- Recency Frequency Monetary (RFM) - Used to determine scores based on recency, frequency, monetary or other values for targeting marketing.
- Affinity Scores - A score to determine how well an individual product fits with the needs of the individual customer.
Near real time segments that can be inferred from streamed data may include:
- Fraudulent activity detection - Typically the use of rules monitoring for unusual behaviour.
- Action Informed Engagement - Triggers to start specific activities that will assist the customer on their journey.
- Lead Scoring - Assigning a value to determine the worthiness of a lead based on an action.
Outputs are triggered based on the segmentations that have been enabled so give some time to getting a thorough list of segments organised to help make connecting with the customer a worthwhile and awesome experience for your audience.
Hyper personalisation and connecting in the moment helps to deliver engaging experiences
Typical CDP digital marketing use cases
Following are some typical digital marketing use cases and where a CDP can be leveraged to provide additional value.
-
Awareness - uses the data about individuals stored in the CDP to understand who to communicate with marketing communications and advertising based on recent engagement activities.
- Basic email engagement - Use case example, only target individuals who have open emails in the last month.
-
Engagement - uses previous campaign data stored in the CDP and results for call-to-action (CTA) campaigns across relevant segments and channels.
- Customers at a location - Use case example, find target individuals based on their purchase location.
- Unsatisfied customers - Use case example, find target individuals based on their response to Net Promoter Score (NPS) campaigns and/or sentiment analysis of support cases.
- Alternative channel engagement - Use case example, only target individuals that have stopped using email and have started using mobile.
-
Acquisition - uses first-party data stored in the CDP to target look-a-likes for advertising on social platforms.
-
Onboarding - uses historical data stored in the CDP to help customers get started with personalised insights in new customer journeys.
-
Cross-sell/Upsell - use historical purchase data stored in the CDP to suppress and target customers with new opportunities and offers.
- Exclusion of recent purchasers - Use case example, only target individuals with re-inventive offers who have not made a recent purchase.
- Date bound purchase behaviour - Use case example, only target individuals when they have received a purchase threshold over a certain period of time.
- Data lineage - point of sale purchases - Use case example, only target individuals from a data source where a condition is met, e.g. online purchases worth over $1,000 in the last year.
- Seasonal spenders - Use case example, find target individuals who only make purchases at a specific time of the year.
-
Loyalty - use an individual's customer data to stored to the CDP to improve retention.
-
New loyal customers - Use case example, only target individuals who have started making a certain level of purchases in a recent time period.
-
Highly engaged individuals - Use case example, find target individuals based on specific rules that qualify them to be considered a highly engaged customer.
-
I have put this article together to share my thoughts, learnings and tips from my SFMC CDP certification journey.
I know that over time I will make an updates as I learn more about the CDP platform. Subscribe to the influencerTips blog for updates.
Your questions and suggestions are what helps me learn, and I look forward to reading comments.