Ray Eitel-Porter is managing director of analytics for Accenture UK & Ireland. He has a deep background in strategy consulting, with LEK, and C-level business executive experience in software, and in clothing and jewellery manufacturing.
He has specialised in data analytics since 2009, and contends that data management has assumed a more elevated strategic role in corporate organisations, under the sign of ‘digital’.
What do you see as the trends in analytics in 2014?
Overarching is the trend towards digital. Enterprises are focusing more on more on what it means to be digital. Now, analytics is just a part of that, but it is a part. Analytics is big thing now because of digital. Massively increasing data volumes are also a function of everything starting to go digital.
If we take predictive analytics as an area where companies would like to do more than they can, how are your clients organising themselves to fill that capability gap?
There is a significant gap for most companies between their aspirations here and what they are able to do. Board members are now interested in analytics, and business and IT strategies are becoming the same thing.
For many companies, understanding what technology can do is becoming a key part of strategy as such. I don’t think it is the enabler it once was, but rather it is part and parcel, especially for the bolder, larger companies.
It used to be the internet start-ups but it is now large companies with the muscle and resources to make fundamental changes who are interested. Insurance is a good example of that with telematics. They realise that this is how the industry is going to turn on its head.
How are your clients organising for analytics?
There are three aspects to it. We are seeing an increase in the appointment of a chief officer of some sort – there isn’t a single title yet, it could be ‘data’ or ‘analytics’ or ‘insight’, but a very senior level person, at board level or one down.
And then there is a realisation that getting a data layer that is controlled and managed, which the whole business can call upon, is very important to get scale. We are seeing a move away from small pockets of analytics.
I can think of three clients for whom we are doing analytics strategy projects. That is taking a step back to figure out how to organise not just the data, or analytics or BI, but the change management that goes with it all, and the governance. Leadership is so very important here.
For more on how to organise for data analytics
- Organizing the BICC Part I: Move to the Middle
- Organisational design of data analytics sparks brightest and best
- Organisational design for analytics function needs governance
And the other thing is analytics as a service. Increasingly, companies are looking to move quickly by buying in services, similar to software as a service, where you get the latest ideas and tools all the time.
The analytics strategy projects you mention. Are they looking to bring in a CDO?
One is, the others are not at the moment, but they could end up with that.
Is the rational core of the excitement about big data the traditional disciplines of data management, business intelligence, and so on?
Yes, many use the term ‘big data’ to mean analytics in the broadest sense. Big data in its true sense is not applicable to many companies. It’s been more of a standard to fly, and a wake-up call. What we are seeing in parallel are pockets of experimentation in big companies with big data. You will increasingly see a federated landscape.
The view of five or ten years ago that everything had to go into one big data warehouse is changing. Using Hadoop for quick exploration, say on social media data, is more the thing.
How important is building a data science capability?
It is very important. Even if a company decides to buy in data analytics as a service, they will still want to have a certain core capability of data or management science themselves. Companies do struggle to find good data scientists.
Is that because the qualities companies are seeking in data scientists are hard to find in one person. Is it not better to think in terms of teams?
Yes, if I think of my own team here, that is true. Broadly speaking, we have the PhD level scientists and then you have the business consultants: data scientists and artists. The latter can understand how you apply analytics to a business problem.
No one jumps from doing nothing to advanced analytics. It is a gradual progression
I think you need that combination. Those who blend both are very difficult to find, though they do exist. Our educational system does not generate enough of those, and we are starting to talk to academic institutions to encourage and support.
For instance, people from our analytics centre in Dublin run courses with the Irish government about STEM subjects in schools to enthuse pupils about Maths.
How do you see the role on traditional corporate IT evolving in respect of this vogue for analytics?
At one of our clients for analytics strategy, the gathering, curation and understanding of information – the data warehousing, the BI, the advanced analytics – fall under the chief data officer, while IT remains responsible for the run of everything, which is a massive job to do cost effectively
Much of the pressure to do more with data has been coming from boards in the past few years. IT is responding to that. But the flipside is that there must be boards who are sceptical about the value for data, and IT will gripe about that. Is the picture changing?
Yes, it is. The popularisation of big data and analytics has helped with that, their getting into the kinds of publications that business people read. Five years ago, senior business people had to have a lot explained to them.
But that clutch of big data reports from Gartner and especially McKinsey changed matters. The summer of 2011 was the turning point. Twelve months on from that, the conversations were about doing things. And this stuff captures the limelight more than IT’s ‘keeping the lights on’.
How does Accenture organise its own data analytics practice?
We take a broad view. We do have experts in the different disciplines – MDM, ETL, BI, predictive analytics, and so on – but we take a holistic view, calling it all ‘analytics’. You could be more purist about it, but when I talk to clients they are on a journey.
No one jumps from doing nothing to advanced analytics. It is a gradual progression, so I think it makes sense to treat it all as one capability.
Recession does encourage a laser focus on efficiency. That was helpful for analytics
What will the industrialisation of analytics look like?
This is crucial. If you don’t industrialise it, you don’t capture the full value. For example, take the ‘next best action’ in a call centre in a telecoms company. You will only add value if you are intercepting every call coming in, with the analytics running on a permanent, real-time basis.
What makes this tricky is getting people to change their processes. Typically, a client will want to test how accurate is your analytical model, predictively. Very often, they underestimate how difficult it is to get the business to change.
Machine learning is also emerging as a real player, but it is still early days for that. It requires a ‘data artist’ to decide if there is anything that can be done with a pattern that has been revealed.
Ultimately, isn’t the pursuit of analytics for competitive advantage a recipe for frustration?
Well, it is bound to be temporary, but that is true of any type of competitive advantage nowadays. If anything, companies who do get an advantage from analytics, because it is new and talent is scarce and people haven’t figured it out, there is an opportunity to gain an advantage for longer.
The global digital leaders like Amazon, eBay, Google are still years ahead. Take Netflix versus Blockbuster. Or the early days of Capital One.
And you are seeing, as Accenture, mainstream companies piling in?
Absolutely. They are somewhere from beginner to medium in sophistication. But they do have resource and talent to bring to bear. Soon they will find themselves left behind by their traditional competitors if they do not get into analytics.
Do you think it possible that just as the focus on the value of data developed through the recession that, as the economy gets better, so too the focus on data will diminish?
That is interesting. Recession does encourage a laser focus on efficiency. That was helpful for analytics. Maybe we will now see more emphasis on growth and more creative uses of big data.