Changes to the centralised technology department model could bring significant benefits.
ALMOST nine in 10 respondents to a survey by Foundry last year believed data would fundamentally change their business in the next few years.
But if data is so important, why are data projects so slow?
The reason is that we organise technology teams and projects in the wrong way: around a centralised technology department.
Modern centralised technology departments have evolved beyond just updating Windows, requisitioning laptops and providing internet service.
These departments provide a broad array of technology capabilities from generic innovation services, data services and cybersecurity, to document control and robotic automation of all kinds.
These teams contain great capabilities that can be highly useful in the modern world.
Their specialisation comes at a cost, however. By providing general services to everyone, they become specialist at nothing in the business.
This creates a great chasm between the technology team and the parts of the business that actually generate money.
The impact of this is that technology teams will almost always start with little knowledge about what other departments do.
This will manifest in projects taking more time, and in project outcomes being slightly misaligned from what was needed.
Good project managers bake this time into the plan, but there is another way.
Technology is getting easier to use every day.
Gone are the days that necessitated rooms full of boffins typing to create the technology equivalent of a Shakespearean play.
This presents a huge opportunity for leaders who can exploit this.
Rather than engaging the technology department to deliver a project, a savvy leader can train their crew to be just as capable with technology.
The idea of having teams embedded in the business has been around for a long time.
In 2019, however, consulting company Thoughtworks embodied these ideas for data teams into a concept called the ‘data mesh’.
The key idea of data mesh is that company departments would have complete control over the data in their department, or in data mesh terminology, their data domain.
Now, about four years on from its inception, the idea is starting to gain traction.
So, what are the benefits of a data mesh?
Consider a manager’s worst nightmare: a dashboard showing their metrics incorrectly in the red.
Yet, as data is shared and copied and transformed, it’s possible for bad data to make its way into a report.
A big benefit of data mesh is that departments can control their datasets, allowing them to trust they will see the red signals before their boss does.
Another benefit is speed.
The wonderful thing about technology is its low cost to create prototypes.
This tight feedback loop is vital to a good outcome.
A domain-driven team can expect faster projects with better outcomes and consequently lower costs.
Global pharmaceutical and diagnostics company Roche was an early adopter of the data mesh concept.
Roche wanted to leverage the vast amounts of data that it held, but it was impossible to imagine a 100,000- person company becoming data driven if only 100 tech people could use data.
Roche now successfully harnesses data across its business, spinning up domain teams as departments need them.
Data mesh is growing in popularity among companies here in Perth.
It’s likely we will see a big proliferation of data domain teams in the near future.
The best part is that, since technology is getting easier, these domain teams can be built from within the department.
I think it’s going to be an exciting time in the data space over the next few years.
But I wonder if the key idea of data mesh, which is about domain teams doing technology, will spread to other technology domains.
Perhaps in the future, technology teams won’t exist at all.
Maybe technology teams will be absorbed by procurement, becoming vendor managers of the tech that the company builds upon.
Only time will tell.
• John Vial has a PhD in robotics and has spent the past several years leading teams in major Perth businesses focused on AI and robotics.
