Drew Rockwell, CEO of Lavastorm Analytics, discusses big data trends and how operators can overcome the inherent challenges.
Eurocomms.com: What are the greatest opportunities for operators to benefit from big data?
Drew Rockwell: The opportunities to gain value from big data are numerous. With new opportunities appearing regularly, I can’t say which application areas will ultimately hold the most value; however, I can say that we see a lot of activity in customer-facing applications, where traditional data sources and analytics have been relatively limited in scope.
We have seen successful initiatives in the customer experience area, including initiatives to implement more profitable business models and pricing schemes, measure the impact of specific events on net promoter score, and collect feedback on new and planned products.
In general, big data can help an organisation in one of three ways: it can reveal new insights; it can transform business processes; and harnessing big data can provide better situational context so that business people can make better decisions.
In practice, the initiatives that operators have in place are small scale – do you agree?
To answer that question, we have to first be on the same page about what a big data initiative is. The invention of the term “big data” came after many big data projects were successfully completed. Lavastorm Analytics, for instance, has been involved for years with very successful and high-impact projects, including a number in the fraud and RA areas, that are processing high volumes of data and fit the big data definition when it comes to variety and volume.
These projects were initiated to solve a specific business problem and weren’t company-wide initiatives, so while I agree that up to now big data projects have been applied to specific uses and in that sense have been “small scale”, I think they have been very impactful and have been large in terms of the data volumes they have handled.
Equally, is it true that many are struggling to show a positive RoI from their big data initiatives?
That hasn’t been our experience, but I have seen signs of that issue in broader industry surveys. You need to have the appropriate infrastructure to handle the data volume and data variety that it entails, but that’s not enough on its own.
You also need to have the right business approach – one that allows you to experiment and explore with little risk or cost so that you can discover the golden nuggets. There are some well-known big data applications, such as revenue assurance, where there is a recipe for success, but in other application areas learning where the real value is stored is an investigative process.
Success, therefore, requires a flexible, rapid approach that allows you to make mistakes and explore freely.
What are the other major challenges that operators are coming across in your experience and how should they try to overcome them?
There are two other challenges we often see. One is the need to integrate all these different data sources and manage the high degree quality differences that goes along with integrating “cats” and “dogs”. This requires operators to use tools that can not only handle the volume, but ones that are flexible and fast enough that they don’t compromise RoI.
In many cases, these tools aren’t currently in use at the operator so it requires expanding the toolbox. More and more we see these tools being put in the hands of the business analysts, not IT because IT can’t keep pace with the demand for changes and updates. So creating a plan for how you are going to divide the responsibility is critical and operators should look to implement a more collaborative process where IT helps enable the business with, say large volumes of data, but the business is empowered with tools to explore the data and even add new data sets that they need to explore alongside the ones carefully cared for by IT.
The second challenge is related to the skills required to aggregate and analyse the data. With more analytic projects to complete, many organisations are finding they are capacity constrained by the people that have the skills to work with the data. Operators can alleviate this issue by hiring data people (if they can find them), training more employees to be data people, or expanding their toolbox to include tools that empower more people (even non-data people) to work with the data.
More than likely it will require some combination of these steps. People may not be available to hire or, with increased demand, they may cost you more. Training employees to have greater technical skills requires a long-term time horizon so that you can see the benefits of that investment.
What is the future of big data?
Lavastorm’s Analytics 2013 survey of 600 analytic professionals showed that 35 percent of companies were expanding their analytic investments because of big data. Given this, I expect a rapid maturation of the use cases. Further, I expect to see its ability to improve customer relationships, change business models and improve overall decision making will be borne out in large scale over the next five years.
Lavastorm Analytics and KPN are presenting a webinar at 3pm on Wednesday 11 June titled "Gain control of your financials, don't let your systems control you". Click here to register