Too often when building anything with a technology focus, the method is to buy the new shiny market tools. However, if you are to build an adaptable, cost-efficient and scalable data platform, you need to plan first. But plan what? And why?

In the second edition of Data Means Business, co-author Jason Foster and I introduced the use of business capabilities. Business capabilities allow you to define everything your data platform needs both now and in the future. The why is explained by a quote from the book: “In the rapidly evolving landscape of modern business, organisations need to continuously adapt and refine their strategies to stay competitive. One powerful tool that supports this is the business capability map.”

To build this map properly, you need to understand your organisation and its use cases, but to ensure adaptability and scalability, this also requires collaboration.

Where to begin: look to your use cases

With a defined set of problems to solve in mind, you should look at what parts of the data platform are needed to meet your organisation’s immediate needs. At this point, it is also useful to understand what other parts of your organisation will use the capabilities being built. One of the key advantages of knowing what you need for a complete data platform is to create an understanding of the dependencies between data capabilities.

Ideally, you would select use cases that are small, well-defined and can be completed quickly. While we would never recommend starting with Master Data Management, if there were no other way to begin, the requirement for core data management, i.e. data quality and governance, would have to be considered in the scope. There are dependencies that will derail agility and adaptability if not considered up front.

The eight components of a complete data platform

So, what does a complete data platform contain? Our view involves eight key components defined below:

• Global data modelling represents data through the business lens of your organisation. This is where you define what data, process and ownership are applied in your organisation. It is a non-technical view of the organisation. Traditionally, business analysis and business architecture would be undertaken here.
• Data management: the governance, processes, policies, guidelines, standards, controls, compliance and rules on how to manage data. These are the core foundational capabilities needed to run and manage data effectively.
• Data source consolidation: the internal and external data which is available and interesting for consumers and external parties. Here, you understand the various sources of data and apply the global data model analysis to ensure it is understood in the right context and uniformly.
• Data integration: this connects and moves data. This is critical and needs to be thought through strategically. Often, data is moved and pushed around in an inconsistent and ad-hoc manner. This is the enemy of adaptability.
• Data storage for insights, analytics, AI and other models and sharing. This is largely a technical are which houses the data to be transformed and used in various formats and mediums (i.e. data warehouse, financial and risk models and all AI models).
• Data access is the authorisation and management of access to data safely and securely. This is often a place where a strategic view is needed to resolve the patch work of access definitions and roles for different systems. Data is a good place to start building robust access that can be used later for all systems in an organisation.
• Data provisioning is a way to link data and applications together. Here, reuse and consistency are the key to enable robust quality data shared with a clear understanding of lineage if appropriate.
• Search and consume gives you a way to visualise, report and make data and insight available. This is like a well-organised supermarket where all data products and services can be found, checked out and used. The contents are understood and in a perfect world quality-marked, so data is used appropriately and with the right level of confidence.

Building a scalable solution

Now, while you won’t build out a full-scale platform for your use cases, you will need to make sure you have core capabilities in place that can scale when needed. This highlights an important need to be able to start small when selecting your tooling to deliver the data capabilities.

This ensures that you can build small use cases to test and define what is needed i.e. when ingesting or moving data, what are the requirements for classification and data quality. This also ensures the costs are minimised and should help provide the ability to provide a value basis for the solution you create. Remember, the solution should be scalable and will be applicable to other areas of your organisation.

We started by understanding the business capabilities of the organisation. We then created small and defined use cases that had a value. Now, while technology is needed, it is important not to forget that creating your business capability map and agreeing on your use cases will require collaboration across the organisation. Why is this important?

This helps reinforce a key concept of adaptability that Jason Foster and I explain in Data Means Business, using a formula: agility x collaboration = adaptability.

An adaptive data platform is built understanding all the constituent parts that allow a full ecosystem to manage, define and utilise data in your organisation. While technology choices are important, for the ecosystem to work well, it needs organisational alignment, and the right operating model fit for the culture you operate within.

While this may need to change over time, it is dangerous to assume that if you built it, the organisation would adopt the necessary change.

Read more

How can good customer data platforms help businesses? – The market for customer data platforms (CDPs) is booming. Since 2016, the number of CDP vendors has shot up from 23 to 96

How to create customer-centric data platforms – It has become pivotal for data leaders to improve their data warehouse offerings to match fluctuating customer demands