Have we lost control of data?

Data. Big Data. Bigger Data warehouses. Even bigger Data lakes. The face of data has taken many forms over the years. Where you store it has changed. How you use it has shifted. It’s hard to keep track. Decades and decades worth of data, and different processes has created data chaos. The current architecture is now monolithic, centralised and has lost its way. It’s time for a shift in mindset that will put data back on the right path.


This is a contributed article by Karl Hampson, CTO Data and AI at Kin + Carta Europe.


The face of data within an organisation has expanded over the years and continues to evolve rapidly. Almost certainly, you’ll recognise that where data is stored within your business has changed. As well as how you use it.

It’s hard to keep track.

The accumulated volume of data, storage architecture and the different processes surrounding it have created data chaos. The current data architecture is now monolithic, centralised and has lost its way. But organisations are only ever a mindset shift away from regaining control. And there has been plenty of talk that the answer could be the hottest trend in data right now - data mesh.

Data mesh, or data mess?

Data mesh can be thought of as the sequel to big data, although the approach is still very much in its infancy. But what is data mesh? It is the idea that rather than having all your data in one place with no clear ownership, it is instead a federated model which serves data-as-a-product and is linked by a web of universal data standards that can harness collaboration across multiple data domains.

But it’s perhaps more useful to focus on the ‘why?’ and the ‘how?’ of data mesh than being distracted by the ‘what?’ and getting bogged down by a technical explanation of its architecture. Ask yourself ‘why do I want to change data management in my organisation, and how can we go about it?’.

Unless you’re working at a heavily-funded startup, it’s unlikely that you’ll be starting any data optimisation project with an architectural blank sheet, and a blank cheque to match. In reality, most companies are not in a position where they can simply pivot to a data mesh – years of complexity and legacy acquisition prevent an entire rebuild of the aeroplane mid-flight. So focusing too much on exactly what a cutting-edge data mesh architecture would look like is largely a waste of time.

The real benefits of a data mesh

Fortunately, however, the benefits of data mesh architecture can be achieved with a strategic plan and creative, collaborative thinking. This means you shouldn’t be looking at data optimisation as a predominantly technical problem to solve but also as a cultural issue.

In the same way that cloud success is not predicated on being a native adopter, the cultural model around the complex software stack is where the true value of data lies. The same level of transformation and frictionless collaboration that has been achieved in managing the modern software stack needs to happen with data.

And as with any cultural reboot, securing internal buy-in is a prerequisite. Achieving this requires the ability to provide a compelling demonstration of the value of data.

An effective data hierarchy

To turn the opportunities that data can provide into tangible outcomes (which you can then demonstrate internally to secure buy-in) you will need a model for establishing data maturity.

At the base level, this begins with a modern cloud data platform that can support priority workloads, while also enabling federated and domain-driven access to data-as-a-product. This means data is packaged up and delivered to relevant departments and team members in an immediately accessible way, facilitating effective data management.

Building on these platform capabilities, the human aspect of data governance should drive trust through data quality, completeness, regulatory compliance and discoverability. This is about making sure high-quality insights are available at the coal face when required.

The democratisation of data assets across all business operations can then lead to analysis and experimentation, where end users are empowered to collaborate. In turn, this unlocks new value and use cases from the insights data can provide.

Remove silos for value at scale

A key aspect of the data-as-a-product mindset shift is ensuring that interdepartmental collaboration is incentivised. Too often, this kind of collective innovation is absent from individual targets, which leads to people using data only to further their own ends. At best, this kind of data usage mindset will only deliver results on a very small scale (and often just for the individuals themselves).

Most organisations are already able to demonstrate data success at this level, but the real benefits lie in showing how it can make changes at scale. An effective way to do this is to revisit a data story that has failed at some point in the rollout process (every organisation has at least one of these) and consider how things could have been done differently. Would it have worked with a different mindset and approach? If the science was right, would more effective and product-focused packaging of the data have made a difference?

By analysing where previous attempts have fallen short, the path to a powerful data mindset becomes clearer, and order can be brought to the chaos of data.

There is scope for every organisation to get more from its data by ensuring that those who need it are able to access it, understand its potential, and are incentivised to work with others in putting it to the best possible use for the company as a whole. Best of all, this cultural shift isn’t dependent on having the flashiest tools or newest architecture – the key forcing function is an alternative perspective and a willingness to change.


As Chief Technology Officer of AI + Data at Kin + Carta, Karl Hampson brings over 30 years of experience working with data. He has occupied multiple senior roles in data and digital consulting businesses, including MD, Board Director, CTO and owner. In February 2020, Hampson became Global CTO for Kin & Carta Data Labs, working with Data and Applied AI Capabilities. He is responsible for positioning and strategy, as well as acting as servant leader to data practice leadership teams in Europe and the US. His primary focus today is on applying emerging and cognitive technologies to achieve speed to value in delivering business outcomes with data.