Planning for Success

We’ve taken a look at why you may want to implement a CDP, how this data can be unified and how can you assess vendor capabilities. Now we take a look at the key considerations in planning a CDP implementation or migration project.

The key areas we will focus on for planning are:

  1. Prioritisation of Use Cases
  2. Data Ingestion, Transformation & Storage
  3. Profile & Audience Implementation
  4. Platform Integrations
  5. Reporting & Understanding Success

Prioritisation of Use Cases

Take your time and make sure it is the best implementation you could hope for…said nobody. The reality is a CDP implementation requires a significant investment in time and resources therefore pressure exists to deliver a positive return within a reasonably short time frame while building new capabilities for the future.

Returns may be achieved through increased revenue from campaigns with more accurate targeting and personalisation, time savings due to more efficient data processes or cost savings in removing redundant technology and paid media suppression.

We assess and prioritise use cases based on a scoring system that considers impact (financial or people) and time to implement. Within our most successful projects we’ve started with high scoring sub-projects at a range of effort levels. The quick ones lock in ROI and keep everyone happy, while in parallel we build the foundations of high complexity and high impact initiatives that in a few months will deliver lasting change.

Depending on the nature of this mix and the selected CDP vendor, we usually see initial use cases being achieved within 2-6 months of the project kick off. A CDP program is at least a 2 year project, so we recommend prioritising within that time horizon.

Data Ingestion, Transformation & Storage

What data is required for your use cases?

You will need to decide on what data points are feeding the new system to allow you to achieve your required unified customer view and the activation of that data. In this planning you will need to consider the following:

  • What are the key sources?
  • What is the quality of this data and does this need to be improved before being used? I.e. are there missing data, are values consistent where required, are invalid characters used, etc.
  • What identifiers exist as these will be essential for joining data?
  • How recent is the data and how regularly is it updated? If you have a key use case which requires near real time activation based on a user interaction but the data is only updated once daily then this process will need to be updated.
  • What data requirements does the CDP vendor have? What processes are required for data ingestion and it may be necessary to transform the data prior to ingestion for some vendors whereas others are able to ingest raw data and run ETL workflows within their systems.
  • What are the cost implications for data ingestion, processing and storage? This may determine how frequently data is sent into the system, how long it is stored for, what processes are used to transform the data and how regularly it is activated.
  • What requirements do you and the vendor have? I.e. will you be exporting CDP data into your own storage and if so what processes will be required for this.
  • How long should customer data be stored? You will need to consider governance policies, the impact on costs as some vendors charge on volume of profiles stored and the relevancy of old profile data to you.

Profile & Audience Implementation

What is the definition of a customer?

We find that within the same business, different teams and systems will have varying definitions of a customer, particularly where multiple brands or regions exist.

A definition for the customer will need to be created that works across the business and conditions can then be created for how the merging of profile data will be carried out. Essentially this will determine what key identifiers need to be present across data sets to create and merge profile data. Profile unification was covered in detail in our second article so have a read if you missed it.

What profile data do you need to capture to achieve your highest priority use cases? This will be a mix of attributes and event data, such as product affinities, lifetime spend, location and behavioural events. Identifying these will ensure you are able to plan the required work for ingestion of this data and configuring the CDP to process and store these data points as required for profiles to enable you to analyse, retarget, suppress and personalise activity.

Finally audiences will need to be planned and built for your activation use cases. Attributes, behaviour and activation timing will all feed into this and the way this is approached will differ depending on your CDP vendor.

Platform Integrations & Journeys

What platforms do you want to connect to and what data do you need to ingest to those platforms to achieve your use cases?

In most cases out of the box connection options exist across vendors however you need to understand the details of these connectors and how data is sent to these destinations. We’ve worked with vendors that advertise out of the box connectors for the tools our clients use but in reality the connectors do not work as required. In these situations additional work needs to be completed for custom integrations with these tools.

Once the destination systems are connected you will need to configure the data that will be pushed to those systems. Generally this involves either pushing a user ID into a specific list or segment for retargeting or suppression in journeys and campaigns or updating data fields for a user profile.

Again consider your use cases and ensure that you are tackling the highest priority integrations first. This will allow you to start actually using the data that you have available and generate a return on that investment quickly while continuing to work through the inevitable long list of systems and data requirements that will build for a CDP.

Reporting & Understanding Success

How will you determine the success of your use cases?

The CDP allows you to report and analyse data from this new unified profile perspective but before any use case is implemented you will need to plan how you will judge its performance.

Some use cases will be more easy to understand such as incremental revenue attributable from campaigns and journeys using CDP data. However others may be slightly more difficult to understand in reality when considering cost savings in media spend that haven’t been affected by other factors or resource savings in new efficiencies in data processes.

Having clear methods to assess each of these will ensure that the focus remains on activity that is having the largest impact and that can continue to drive your roadmap.

Moving Ahead!

There are so many moving parts to a CDP implementation as all aspects of your data, systems, processes and teams need to be considered. While this article does not provide an exhaustive list of steps in planning, considering these areas in your planning should provide you with a comprehensive list of deliverables that can be planned and prioritised.

An agile and pragmatic approach is an obvious requirement to ensure that teams can remain flexible in delivering each of the building blocks to unlocking your use cases. But my main advice is don’t overcomplicate an already complicated task. Constantly refer back to your prioritised list of use cases and look to implement the MVP for each and iterate from there!

If you’re about to embark on a CDP project then get in touch with us. This is a uniquely complex topic that at the very least we can give you some tips on.

Get in touch