Breaching the Data Dam - Part 1
Why marketing teams need their own data platform
Marketing technology is a hot discussion topic at a senior level in most enterprises, particularly between CMOs and CIOs. A common question that comes up from the IT side in these discussions is “Why do I need yet another data platform?”
In reality, that’s the wrong question because:
IT does not need another data platform, Marketing needs a dedicated one.
Many enterprises have spent years and millions of dollars upon realising Enterprise Data Strategies that revolve around consolidation and centralisation. Key drivers of consolidation include cost reduction and simplification of the technology stack. Centralisation of data provides benefits focussed around improved systemic control and compliance.
Although these strategic approaches undoubtedly deliver multiple benefits across the organisation, there is usually one exception – the marketing department.
The problem with most Enterprise Data Strategies from a marketing perspective is that they are essentially driven by compromise. In many enterprises, there is an overriding need to create and maintain an immutable System of Record (SoR) to satisfy external reporting and regulatory requirements. This need results in compromises because every department in the organisation has different requirements, and it is impossible to meet them all.
In a digital landscape that is increasingly fragmented, with consumers using ever-increasing numbers of platforms and devices, marketers are becoming hamstrung by traditional Enterprise Data Strategies. Common enterprise data environments fail to provide marketing departments with what they need to drive the types of experiences that their customers and stakeholders are demanding.
Key signs that your Enterprise Data Strategy is not supporting your marketing programme
Your data lake is more like a dam. Modern marketing requires quick access to data in order to drive timely and relevant experiences. However, most enterprise data warehouses and data lakes require skilled analysts to access and manipulate the data that they hold - skills that marketing teams usually don’t possess. Consequently, marketing data requests often join a significant backlog for prioritisation and execution. This means that, in an increasingly real-time and reactive world, the opportunity to send a campaign can be lost.
The cost to add data is prohibitive. Systems of Record used for business critical and regulatory reporting typically cannot be changed without going through multiple layers of approval. This usually adds layers of cost and time to the process of adding data to the environment. Marketing departments struggle to justify the cost of sourcing data for campaigns that run for relatively short amounts of time, and the IT change timescales involved are often too long to make it worthwhile.
Your data environment is not designed to handle scale. Enterprise data warehouses and data lakes are often built to store and report on large core transactional data sets. However, even in the largest of organisations, the transactional data set may be dwarfed by the amount of digital data that can be generated by web analytics and marketing platforms. To handle this scale and preserve capacity for critical SoR data, digital marketing data can be heavily filtered or summarised, limiting its usefulness for driving campaigns. Speed is also an issue – most centralised data infrastructure is not designed to handle the real-time data feeds and decisioning required by modern customer experience programmes.
Not all marketing data needs to be centralised
Although there will always be a need to feed some marketing data back into the centralised enterprise data store (e.g. opt outs, business process and reporting data), there is a significant amount of operational marketing data that will never need to be centralised.
A lot of marketing data is only used by marketers, for example: email opens, complete web analytics feeds and paid media impressions.
Marketing data can be transient, in that it may only be required for a one-off campaign or for a short period of time (e.g. – seasonal, promotional or partner data).
Some data is platform-specific, for example: specific data tables that are required by a marketing platform for its operation but are derived from a larger data set.
Holding marketing data in a dedicated platform makes a lot of sense for marketers who need to run time-sensitive, reactive campaigns. However, it also makes sense from an IT and data governance point of view because:
It preserves the integrity of SoR data sets so that business critical reporting and processes are not affected by relatively unstructured and transient marketing data.
Marketing teams can have SoR data access limited to only that required by their marketing platforms. This improves auditability as it becomes easier to track what was shared with the marketing department and when.
IT capacity planning becomes easier without the ‘lumpy’ demands for data from marketing teams when large campaigns are being run.
You cannot drive a modern digital customer experience programme from the same data infrastructure that is acting as your System of Record.
The only real solution to the different needs of IT, Compliance and Marketing is to have a dedicated data platform for marketing. This platform needs to be fed from a variety of sources, including Systems of Record, to generate a single view of marketing data. The single customer view can then be distributed in relevant subsets (segments) to downstream marketing decisioning and delivery platforms, and then augmented with marketing campaign response and interaction information.
Only by implementing a dedicated Customer Data Platform (CDP) can enterprise marketers deliver the customer experiences that are increasingly demanded of them.
But what of the other elements of the modern marketing technology stack? A common question that is asked by CMOs is whether any of their existing marketing technology platforms can potentially be co-opted to act as a Customer Data Platform. In part two of this article, we’ll look at the other types of data platform in the marketing technology stack.