Site navigation

Leader Insights | Saving Data Science with Martin Thorn, Standard Life Aberdeen 

Ross Kelly


Martin Thorn Standard Life
Ahead of the fifth annual DIGIT Leader Summit, Martin Thorn explores the changing nature of leadership in data science.

The failure rate for data projects is alarmingly high. Gartner suggest that around 80% of data science projects fail, which according to Martin Thorn, head of data science at Standard Life Aberdeen, is hardly surprising given the misguided approach at many organisations.

A common stumbling block is the tendency for data science to be treated as special and viewed as a silver bullet to solving problems. And Thorn thinks it’s an issue we’re seeing more of in modern business.

An organisation has a problem or aspires to enhance and streamline operations, technology is perceived to be the all-mighty salvation and thereafter endless funding is thrown at data science teams.

That is, he adds, until they’re asked to present the fruits of their labour, when the outcome is rarely positive for the people footing the costs.

“I think that data science is not yet mature enough in many organisations, and as a result we’re treating it with a certain level of expectation,” Thorn says.

“It’s the idea that just give a team a huge vat of data, lock them in a room and then they’ll come up with the answer to all your problems. But when you check in with them a year later you find they’ve disappeared down a rabbit hole that doesn’t really help you or the business.”

The issue here lies in leadership and the perceived place of data science within the structure of an organisation. Thorn believes that many organisations still remain unsure over how data science teams should be led.

C-suite executives often view data science “as some kind of wizardry” and are scared of infringing upon what they view to be the pinnacle of innovation within a business – a team that operates independently and innovates organically without the need for traditional ‘leaders’.

This is what makes data science unique, but it’s also what inhibits both the team and the broader organisation.

“It’s a model focused on just giving teams money and leaving them to get on with things,” he says. “It’s as if we’re trying to prevent them from being sullied by commercial problems and letting them embrace what they’re doing.”

“But in no other part of technology would you just leave people to get on with it. In any other part of an organisation you’ll have management or some semblance of leadership to move things along and oversee things.”

Thorn believes this idealistic, laissez-faire attitude toward data science teams must be challenged and changed. If not, C-suites will find themselves throwing money away time and time again.

A business focussed approach

Long-term, organisations should view data science in the same vein as any other facet of the business, and leadership should reflect that, he says.

“Too many data science teams are run by data scientists, which I don’t think is the case when you look at other facets of many businesses.”

“Data science leadership requires the same approach as with anything else. We need to be looking at more mature areas of businesses, because you would never hire a head of development and expect them to be coding all day,” he explains.

“You would never hire a head of InfoSec and expect them to be on the tools. You just wouldn’t.

“That’s where data science is interesting. You can have a head of data science with ten people working under them and they spend their day coding. It doesn’t make sense and I don’t understand why we continue to let that happen.”

For data science teams to truly deliver value to their organisation, they require leadership figures that can cross the divide between data science and C-suite to contextualise things effectively.

Read more Leader Insights

Unfortunately, Thorn believes many data scientists simply aren’t suited to this commercially-focused role. The image of data scientists being shy, dry and laconic appears to be a worn-out cliché, he notes, but in his experience is still a common issue.

“With a lot of data scientists, the skillset that makes them great at being data scientists means it’s hard for them to speak to C-suite in general terms,” he says.

“We’re talking about a discipline that rewards and encourages a great level of detail and obsessive nature. And I think that it’s really hard for somebody to say, ‘how do you summarise the last years’ worth of work in one paragraph?’ – it’s a very tough ask.”

Fundamentally, Thorn says the C-suite craves the reassurance that their money is being spent in the right way, or focused in the right direction, and people who can rest in between both positions will be key to future success.

Balancing the middle-ground between the creative, curious realm of data science and the hard business realities will help establish a healthier relationship between C-suite and data science, he says. And at the core of this lay the softer skills, communication know-how and personable qualities that boardrooms value.

“It’s the translator skill that I think we’re often missing. It’s the ability to take the woes of the C-suite and weave it into something that we can do something with. But it’s not necessarily a skill that many data scientists have practiced, or are used to dealing with,” he says.

Varied personality types

Organisations are faced with two challenges when choosing leaders that span this divide, Thorn believes. The C-suite could find a data scientist who is gifted with ‘translator’ skills, but this is rare.

So much so that he refers to this type of data science leader as a ‘unicorn’. Elusive for many organisations but, when found, a gamechanger.

Google’s Cassie Kozyrkov or AWS’ Allie K. Miller are both prime examples of these types of data science leaders, Thorn says. Both are able to stand in front of audiences and explain complicated processes in simple terms.

By that measure, an individual would also be able to do so for a boardroom. And that’s valuable to both sides of the divide here.

Alternatively, Thorn says organisations should consider individuals who know “enough about data science to be credible and hold people to account” while also being able to explain and contextualise in simple terms.

“It doesn’t matter how many spreadsheets or slides you show someone. Generally speaking, a CEO is going to buy from a person. That’s what it comes down to.

“It’s a question of do I believe this person, are they credible and do I trust them? Because I don’t understand what they’re telling me, I can’t do what they do so I just have to go with that gut feeling,” Thorn explains.

“That’s the trick for data science leadership – or any leadership role, for that matter – it’s about building trust.”


Cultivating a culture which values communication, softer skills and the ability to talk in commercial terms will be key for data science teams in the future. As Thorn explains, the maturation of this field will, in time, diminish the level of flexibility data science teams are granted.

“At some point in the future, many companies will start killing projects. They’ve essentially flushed millions of pounds down the toilet and what have they found?

“Oh, we’ve found some interesting stuff and our scientists have had a few white papers published – I’m not sure that’s the commercial reality that a lot of organisations are looking for.”

As data science becomes more commoditised throughout businesses, Thorn thinks the people that are going to stand out are the people with softer skills.

“If you look at what’s happened to other coding disciplines, they generally get offshored, but what is still onshore is the massive value-add, problem solving-type customer engagement skillset,” he says.

This changing landscape of demands and skillsets does worry Thorn. For many data scientists, the inability to adapt and evolve new cultural norms puts them at risk of being commoditised and replaced.

DIGIT Leader 2021 Virtual Summit | Join the Discussion

Martin Thorn is set to feature at the fifth annual DIGIT Leader Summit, where he will be further exploring the changing face of leadership in data science.

For more details and information on how to attend, please visit:

Ross Kelly

Staff Writer

Latest News

Cybersecurity Finance
Cybersecurity Editor's Picks
Climate Editor's Picks Energy Featured
%d bloggers like this: