All communication service providers (CSPs) have silos of data that could yield pure gold. Nick Booth warns that, in the rush to extract it, there could be a lot of casualties.
To paraphrase one of the experts quoted below, there’s gold in CSPs’ data sets. The individual elements scattered about the storage landscape contain information on behaviour, tastes, location and demographics that, if combined by a marketing alchemist, could make CSPs richer than Facebook.
Gold, like big data, has many applications beyond sales. Both are vital components in making business
systems run more efficiently, with gold being a component in computers and electronics to speed the
accurate transmission of information. Gold has dental and medical applications – as indeed does big data, which provides the basis of the nervous system and immune responses of the increasingly organic and self-regulating networks of the CSPs.
In other words, there’s so much you can do with big data it’s terrifying. It is literally causing panic in some companies. “Many departments are under pressure to boost their customer revenue and optimise their network and even to improve the customer experience,” says Ravi Palepu, the head of telco
solutions for management system vendor Virtusa.
The biggest mistake a CSP can make is to run all their programmes at the same time, says Palepu. First
CSPs need to decide which of their ambitions – improving customer experience, boosting revenue,
improving operations or creating pro-active marketing – is their priority. “Running all the different streams in parallel will be counter productive,” he adds.
Having decided on which seam of data is to be mined, the next priority is assembling the tools and the personnel.
Contrary to popular marketing messages, many companies have impressive data mining systems like
Hadoop and SAP’s HANA, but haven’t worked out what they really want from the data sets, nor do they understand the domains or know what they want to do, says Palepu.
Too often CSPs lack the expertise needed as there’s a small talent pool, says Matthew Roberts, director of marketing at Amdocs’ big data analytics and strategic innovations division. “Data scientists are in high demand, particularly those with a telecoms specialism who can understand the data in context,” he says.
Meanwhile, the staff charged with striking gold are having to blend data from different sources and make sense of it. “It’s like a cocktail, blending different ingredients and you don’t know what works until you’ve found the winning formula. The trick is to find it fast before you drown in data,” says Roberts.
If it’s difficult to find the staff for that ambitious deep dive into the data, why not create a more realistic target? While other CSPs are tying themselves up searching for the meaning of life in their data, it might be a lot more worthwhile going for some quick wins, according to Ben Parker, chief technologist for Guavus.
“When panning for gold, you do it in the river, sifting out the excess, and bringing home the nuggets – you don’t shovel all the soil into a truck to take back home and sort, that would be a waste of time and money,” he says.
Yet this is what many CSPs do when they take a store and query approach to data analytics. “By applying streaming analytics at the edge, sifting out the useless data and keeping the nuggets of gold, CSPs are better set up for success,” says Parker. The best way to immediately improve network operations, customer experience and revenue streams is to work on the data streams not the stored silos, Parker argues. It’s a more practical way of showing the top management that results are being achieved.
Another practical approach that CSPs – and the companies that promise to solve their big data problems – could take is widening the talent pool by making analytics easier.
Not many people can create the algorithms for interrogating big data, but why should they anyway, asks Steve Bowker, the vice president of technology and strategy for Teoco. “Once the data has been interrogated with a specific use case others can make use of it,” he says.
Big data isn’t as scary as everyone makes it out to be, argues Andy Stubley, the vice president of sales and marketing at SysMech, which creates big data apps for CSPs.
“You just need to know what you want to get from the data,” says Stubley, “half the battle is asking the right questions.”
The idea that big data can only be understood and translated for practical use by data scientists is a
common misconception anyway, says Mark Davis, Citrix’s senior director of product marketing. “Yes there’s a shortage of big data scientists but a new class of big data analytics platforms makes the insights more widely available.”
For big data’s benefits to be fully realised, more people need access to insight, which includes marketers and customer care organisations, who will want immediate and continuous use, Davis adds.
That may be true, but that leads to another intelligence conundrum: there are so many big data solutions out there. How do we make sense of them all?
