Digital transformation needs business leaders to develop new skills

Digital transformation

Q: What are the big things you each have in mind in regards to digital and the change and impact it is going to make? From a Computer Weekly perspective, our readers are mostly corporate IT. So is digital transformation a genuine thing for CIOs or is it just supplier hype?

Gah-Yi Vahn: I started thinking about big data before it became a buzzword back in Berkeley, California, where things tend to be a bit more ahead. I’m increasingly interested in the misalignment between existing capabilities and what is being deployed in the business world. There is a huge gap.

Strahan Wilson: [Eat is in] an interesting industry from a data perspective. We are not influenced to any great extent by the internet. You can’t “Amazon” or “Google” food. Big data has been a buzzword for a long time without becoming a reality. It’s only in the past two to three years we have started thinking about it. Working with Blue Yonder has opened our eyes to [what is] possible.

One challenge is that it requires a different skillset to process big data than to analyse small data – that which can be done in Microsoft Excel and that which can’t. You have to be able to make sense of the noise in the system. 

What does that mean for the future of the finance department? Is it a skillset to bring in-house or beyond that, which you can expect of accountancy training of a chartered accountant? Or are you better off building relationships with [suppliers] that operate in that area and build those relationships?

Christian Benson: I’ve seen things move on hugely during the past four to five years with the advent of big data. I think analytics has the power to be transformational in business.

As far as we are concerned, there is no question – the whole go-to-market is focused around digital transformation with analytics underpinning all of that.

Big data analytics

Q: So big data analytics is a function of the transformation to digital. Is it a driving force that gives big data meaning?

Benson: I think it does. There will always be cases where you don’t need analytics, but it’s the foundation for digital transformation and the value you can get from data, which you couldn’t get before.

Rakesh Harji: I’ve spent the past 25 years in the IT industry and I started as a computer operator and systems engineer. I have a similar background in that regard to Christian. I’ve seen tech move.

The internet is transforming the way companies communicate with customers because of multiple platforms and channels: online, in-store, catalogue, mobile, loyalty programmes and franchises. It’s exponentially getting bigger. There is a tremendous opportunity if companies can capitalise on data and become more data-driven as a business – it will change.

And that’s all good, but where digital transformation will come into play is where companies will be able to automate decision-making. We believe 99% of decisions can be automated. If you are able to do this, it will help you become a digital business.

Board discussions

Q: Suppliers talk all the time about “digital transformation”. When you are in board meetings, do you sit around saying, “We will have digital transformation”?

Wilson: Do we do that every week? No, but as part of a strategic planning cycle we ask, “Do we take advantage of the way the digital world is moving?”. We haven’t had to find a competitive response to online players. We are not like some industries, such as fashion, where no doubt they are seeing a big push to online – they have to respond. We have the luxury of not having to respond in the same way.

Instead, we look through the lens of competitive advantage. While the market is more protected from online, we still have fierce and ferocious competition in our market, so what can we do through digital channels to keep a step ahead of the market that offers us competitive advantage? 

We don’t use the language of the supplier, we never use the words “digital transformation”. But we do need to get better at forecasting or better at communicating with customers. That is digital transformation, but done with clear business objectives in mind.

Digital communication

Q: Do you do social media analytics?

Wilson: I have done in the past. Am I pushing it at Eat? No, but we do email communications and use social media as a platform with customers. The way I see Facebook and Twitter in our industry is that it’s a complaints channel, so it’s important. It’s hugely valuable in terms of brand building.

Digital communication is very important. I came from a business that measures marketing budgets in the tens of millions. Digital offers Eat a low-cost way of talking to customers without spending £10m on above-the-line marketing.

Beyond Excel

Q: Do students have the skills required to build digital careers, and do they really need much more than Excel skills or is it just a percentage of them who do?

Gah-Yi Vahn: At London Business School [LBS] and other top business schools, there is a core course on data analytics that teaches Excel skills, spreadsheet modelling and risk analysis. This is fairly standard. 

It’s a course that is received very well. LBS’s course is called “Data Models and Decisions” and it’s one of the courses that is appreciated the most. Students say they are using the skills straight away, especially those who enter consulting. We are in talks to develop an elective course on big data and there is interest from students.

Wilson: There are interesting questions around that step beyond Excel. Is that something the generalist could ever require as a skillset? I am slightly biased, as my experience is talking to the Blue Yonder team who are at the more technical end of the skillset spectrum.

I came through the classical route. I consider my Excel skills to be quite good, but I know there is that gap between what I can do in Excel and what the Blue Yonder team can do. What is the right skillset for anyone trying to analyse big data – is there an “Excel plus” model or do you want a different skillset and the data scientist route?

Big data platform

Q: What are companies saying about big data?

Benson: A lot of them are looking at building a data science capability, especially clients who have a clear idea of what to get out of it and understand the need to commercialise their data more. They are potentially defining, designing and implementing a big data platform, which takes a particular set of skills rare in the market, especially if it’s Hadoop-based. It’s pretty difficult to find those people.

Once that platform is up and running and maintained, data scientists can extract the value. Sometimes those two tracks are misaligned because it has taken longer to set up the platform than thought. 

Finding a really good data scientist is difficult. It requires someone with a strong statistics background, business engagement skills and creativity to find the cases and applications – and someone who evangelises.

Very often, big data initiatives can meet a bit of scepticism around the value it brings to the organisation, so part of the job is selling it by showing the value and convincing the business to engage with you, showing that you can do something useful.

Harji: For some of the processes we address – operational processes and demand-based forecasting – it isn’t just about building the model. You have to run them 24/7 to fulfil that back to the business. 

If you are in a retail business, demand forecasting must be done on a daily, granular level. Data patterns captured are changing constantly, so models need to be redeveloped.

Having data science as a service and automating processes is much more of a capability for customers to consume, instead of investing in infrastructure and people trying to build the models for their business – that takes time. They are looking at a service-based approach.

Wilson: As a customer, this is the second time we have tried predictive forecasting. First, we paid a consultant to come in and make a model. It was built and, “Oh, yay, it’s working today and it will keep working” – of course, it doesn’t. What do I do? Bring the consultant back in at high cost or recruit talent? It’s improbable we can afford them. How do we manage them? Train them?

I wouldn’t know how to recruit or train and we wouldn’t be able to keep them. They would be sat on the side in a silo of the business wondering what to do. 

Then it’s a no-brainer. There is no practical way for a business our size to get into data science unless it’s through what we received from Blue Yonder – buy everything, not just the platform. There is limited use if it can’t be maintained.

Take our store managers – what is their core capability? Their reason for employment is customer service. We want them out talking to customers, but because of the tools and systems we have, they are effectively trying to do predictive analysis in their heads and it’s taking them a huge amount of time.

What is the future like in terms of the integrated digital transformation in our business? Let’s take all the decisions that don’t relate to being nice to the customer and automate them.

Harji: For a 14-day forecast [at Eat], it works out to one million predictions a day. Machines are better suited for such decision-making, taking into consideration weather, promotions and events. These factors can be taken in, but Excel can’t factor them.

DaaS benefits

Q: But doesn’t getting data science as a service erode competitive advantage?

Wilson: I’m not sure you would ever erode competitive advantage. Am I eroding competitive advantage if I’m hiring the greatest minds in the field? I don’t think I am. Maybe in the hedge funds industry you would, but in retail – where it is not core – there’s not a single retailer out there, including all the majors, that wouldn’t benefit from data as a service (DaaS).

Creating new roles

Q: Are MBAs joining companies to be data scientists, or are they joining as managers or setting up on their own?

Vahn: So what does this mean for MBA students? Not becoming data scientists. But they will have to learn what is going on in the space for understanding and appreciation. For example, finance went through the data revolution first and now it is standard for investment banks to have in-house [data science] teams. MBAs who work there appreciate having that capability in banks, working side-by-side.

For a small organisation that can’t have it in-house, but know technology and services exist, you of course seek it out to add value to your company. 

If I were to guide my younger siblings towards the best career in the next 10 years, I would love to see programmes that combine business, engineering and statistics. It’s something that is in conversation, but they can’t be set up as quickly as we want. 

Ideally, my students going through combined programmes will be able to create roles for themselves. I can see that in the future, but for now, we are striving to teach them to appreciate the technology already available.

The productivity puzzle

Q: Will the move to digital and developing data competencies help with low levels of productivity in the UK?

Harij: If you can automate a high proportion of processes, then it allows a team to focus on areas that are important to the business. Those productivity goals can be improved on. Productivity will play a huge role in this space.

Wilson: The relative decline in the UK – are people in the UK getting lazier? No. There are increases in employment in my sector – hospitality has been hoovering up people. 

Productivity levels are low as we are in a labour-intensive business where we serve customers physically. Tills being automated was a big driver in retail, but we have been plateauing ever since, as there is only so much automation you can do.

Quantitative analysts in banks, on the other hand, can generate millions of pounds of profit with one person. The productivity per head is huge. We need the banks to pick up the slack in the UK.

Benson: There are huge opportunities in manufacturing for increasing productivity through automation and analytics. Even in manufacturing, which has lean principles applied a long time ago, we are still seeing huge benefits through analytics.

Transformation of the CIO

Q: Computer Weekly has covered the rise of the chief digital officer or data officer [CDO]. Does that mean the obsolescence of the CIO? 

Benson: We are seeing that. The nature of our relationships are changing because of the advent of cloud and software as a service [SaaS] and the fact that, with something like Qlik, they sell to lines of business rather than the IT department. They are selling to the marketing or sales department. Lines of business owners are now empowered and don’t need a capital budget to be able to go out and buy IT – they can purchase a subscription from Salesforce.com.

And so, to some extent, IT is being sidelined and having to reinvent itself.

They don’t just want to be seen as keeping the lights on and infrastructure running. Potentially, the roles may be merging – or certainly working in tandem – where the CDO is more of an advisor, a mentor or ideas person who understands the potential of the market. 

In that scenario, the CIO becomes the executor and implementer of that vision and strategy. Otherwise there is a huge amount of experience you are not tapping into. The CIOs have seen the market evolve and have relationships with software suppliers and they understand the pitfalls. 

The SaaS model is great, but you have limited flexibility to make it operate the way your business does.

Wilson: The move to SaaS is moving IT into the CFO role. All SaaS decisions we’ve made have ultimately been taken at the board level, through me rather than IT. Historically, IT has been all about development risk and construction risk: Are we going to get this live? Are we going to spend a lot of money and have it go to waste?

Now I don’t have that implementation risk. The IT director’s role is to discover [what is] possible – talk to the suppliers, cut through the hype and understand, “Is it something we want or is it just a sales pitch?”. 

The IT director has the technical knowledge that will always have a role in-house, which allows you to see through it all and say this is valuable. 

Effectively, what IT does is bring it to the board and say, “This is a really useful service for us”. That advisory role is always going to be important. There’s also always going to be a data integration role and advising on as-a-service provision.