Seeing the future – analytics as a competitive edgeby
By Degerhan Usluel, KXEN
Analytics software has become so much part of the landscape in customer lifecycle management that it's hard to imagine life without it. We frame a question, curse the delay while our highly-paid analysts build a model to answer it, and we get a result of sorts. The problem is that the answer is increasingly too simplistic and too late.
It's not our tried and trusted tools that have changed, but the business climate. Before, we lived in a world of relative certainties. Now, we are challenged by globalisation, commoditisation and lower margins. Marketing has shifted to centre-stage as organisations realise they not only must they react much more quickly to customer needs and wants, but they must also be smarter about what they sell, when they sell, and to whom.
The conventional thinking underpinning one-to-one marketing of concentrating on the top 20% of most profitable customers is giving way to a new understanding in which companies once dismissive of lower-value customers now know they must embrace all but the least profitable of them. The new thinking is that to sell less but to many more customers - known as the long tail - has become vital to bottom line health.
But while some of the other tools used in sales and marketing have become more sophisticated, providing the enterprise with a mass of rich data, conventional statistical analytics has failed to evolve. It cannot handle the new flood of data, proves painfully slow to use, costly to own, and is unable to deliver the sophisticated, highly granular results needed to manage today's customer lifecycles.
The result is that enterprises remain information poor, making decisions on their product mix based either on stale information, or, in frustration, little more than guesswork. Industry commentators have likened it to driving a car by only looking in the rear view mirror.
But a change is coming that is being dubbed extreme analytics. A new analytics technology that delivers results in hours rather than weeks, at a fraction of the cost, and with an unprecedented level of granularity, is overturning old concepts of modelling, changing the way companies exploit customer information and, as a by-product, changing the traditional role of analysts too. It is based on breakthrough thinking by Russian mathematician Vladimir Vapnik.
His statistical learning theory has rewritten the rule book on business intelligence and data mining. Instead of a single analytical model demanding weeks or months of fine tuning, it allows a large number of models to be created very quickly.
Where conventional analytics is limited to considering a few tens of variables, the new analytics routinely works with thousands, so there's no longer a need to apply assumptions and preconceptions – all the data can be used and the quality of the result increases. Where before analysts have had to spend weeks manually preparing data, extreme analytics automates the data preparation process, cutting the time required down to a few hours.
In presentation too, extreme analytics is markedly different to old-style analytics. No longer are results delivered as impenetrable tables of figures. Instead, they are outputted as clear charts and graphs that avoid statistical jargon and are easily understood by business people across the enterprise.
All this means extreme analytics has a remarkable democratising effect, enabling more questions to be asked by more people – and delivering the accurate answers today - not next month. Companies can afford to build hundreds of models with thousands of variables and, for the first time, get timely, accurate and highly detailed answers that add real insight to their customer relationship management programmes. The greater granularity means companies can address much smaller groups of customers and prospects more accurately than ever before, with offers and propositions that are consistently relevant.
Powergen, one of Britain's leading energy suppliers and part of E.ON, the world’s largest investor-owned utility, is using extreme analytics to help target customers and prospects with the most appropriate energy packages for their needs. The first time it used an extreme analytics model, it saw a 20% uplift in sales and a direct £150,000 saving in mailing costs while retained 300,000 customer contact opportunities for future campaigns. Powergen has gone on to embed extreme analytics at the heart of its customer relationship management programme.
Direct marketing specialist Cox Communications uses extreme analytics to produce 1600 response models for marketing campaigns in 26 regional markets from a data base of 10 million customers and 800 variables. As a result, direct mail responses rates have risen from 1.5% to 5.5%.
A major European wireless communications company placed extreme analytics at the centre of a complete analytic environment to support the easy creation and deployment of churn and cross-sell/up-sell models. A Teradata warehouse was built with a customer analytic record containing over 2500 meaningful variables, with monthly aggregates updated on every billing cycle.
With data preparation for each model reduced to virtually zero, they are routinely creating more than 700 predictive models a year for multiple tasks including marketing, customer price sensitivity analysis and channel preference. Consistently high quality models are being produced by the business analysts and the company is able to automatically deploy models into production and re-score customers whenever necessary. Overall annual churn has been cut from 26% to 21%.
A leading direct-to-consumer insurance company is building hundreds of models on click-stream data, enabling its e-commerce agents to close six times the amount of business using real-time predictive scoring compared with the previous rule based systems.
These examples – just four among over 500 users and growing fast – illustrate the ability of extreme analytics to change the way companies relate to their customers. They also point the way forward for companies wanting to survive and prosper in today's business climate – deploy extreme analytics and use the quality of customer relationships to sharpen competitive edge.