• Home
  • >
  • Robert Bates, Head of Decision Sciences, Currys

Robert Bates, Head of Decision Sciences, Currys

Describe your career to date

I’ve taken a more winding route across all aspects of data science (or analytics as it was called when I was starting out) and looking back at each stage I realise I’ve developed different skills and ways of looking at things from each one.

 

After my PhD, I worked in management consulting (strategy), which emphasised the importance of structuring the problem and the need to take stakeholders on that project journey with you.

 

I then moved into commercial finance roles, which taught me the power of setting appropriate business targets; not only do you get what you measure but you have to balance ambition with achievability. It may be the best idea in the world, but if the business doesn’t believe it (and won’t sign up) you won’t get the project funds.

 

At Currys, I moved back into pricing and marketing analytics roles, which provided experience of the wider business operations and emphasised decisions are rarely clear-cut. Breaking down the key dependencies by applying systems thinking means you can ask the right questions and manage compromise.

 

Finally, the past few years have been spent driving engagement and encouraging use of analytics across the business and leading a team to do so.

What stage has your organisation reached on its data maturity journey?

In some ways this is a trick question – few parts of any organisation will be exactly in sync, and the speed of advances in technology and platforms mean that you’re always evolving and looking for new ways of applying innovative technologies.

 

However, I believe we have solid foundations across the board – there is strong sponsorship and belief in the value of data and insights, coupled with a strong culture of innovation. This means we’re able to identify and utilise the value of data-driven insights across the business and, while there will always be areas for improvement and increased maturity, we’re able to get there quickly through agile methods.

 

Tell us about the data and analytics resources you are responsible for

Within Currys, we have a relatively small central analytics team at present – around 10 to 15 individuals – in addition to functional analytics teams who are dedicated to support in an individual area.

 

This central team supports a mix of larger, cross functional projects and those which require more specialist data science and modelling skills which may not be needed on a day-to-day basis, allowing us to benefit from combining in-depth operating knowledge with cutting edge technical skills to develop people across the organisation.

What challenges do you see for data in the year ahead that will have an impact on your organisation and on the industry as a whole?

It’s hard to miss the bigger stories relating to ML and AI in the news (ChatGPT). These and the continued improvements in computing power and ML-based toolkits create an impression that harnessing data is easier to do and implement than it actually is. This is one of the biggest challenges – how do you manage the expectation gap between what the business thinks is possible (almost anything, right now), what is possible (pragmatic, impactful solutions), and what the more technical data scientists want to develop (the perfect solution).

 

It’s the ability to successfully balance these competing tensions that has the greatest impact – there’s real excitement for innovation in data but without demonstrating benefits through delivery of pragmatic, actionable solutions this can be easily lost. To do this we need to develop a layer of ‘data translators’ who span the commercial and data worlds and such individuals are rare at present.

Have you set out a vision for data? If so, what is it aiming for and does it embrace the whole organisation or just the data function?

Within Currys, our overall goal is to ‘help everyone enjoy amazing technology’ and this extends across the data function where business partnering is critical to understanding the use cases and unleashing the power of data across the organisation.

 

The other day one of my colleagues observed ‘Data without understanding is just IT’ and I agree – we collect, protect and use data to drive change in the business, otherwise it’s just an expensive asset we never see the benefit of. Central to ‘helping everybody enjoy amazing technology’ is transforming how we use this asset, and how we enable the data translators to become agents of change within the business.

 

Have you been able to fix the data foundations of your organisation, particularly with regard to data quality?

I don’t think any business, anywhere, will be in the position of being able to select from perfect a data-set. There will always be gaps somewhere, a new feature not yet coded or simply differences in interpretation of an event – it’s part of life in the data science world.

 

From a decision science perspective, I see the challenge as identifying the level of data completeness required and tolerance permitted on a use case by use case basis – focusing upon the agency being granted to the model and the present levels of accuracy/certainty in the outcomes being predicted. This moves the question away from ‘have I got perfect data quality’ to ‘what do I really need to know, and when’ allowing you to focus upon the key data flows as you move into model production and ML Ops. 

Robert Bates
has been included in:
  • 100 Brands 2021 (EMEA)
  • 100 Brands 2022 (EMEA)
  • 100 Brands 2023 (EMEA)