Predicting When to ‘Dial-up’ and Stay ‘Connected’ for Success with SQL MapReduce

I will be surprised if these days, mobile applications  are  not in the top strategy agenda of most industries, be it Banking, Financial Services, Retail and of course Telecom where it all originated. While much of the exhilaration associated with Smartphones and mobile applications (i.e. Apps and Apps Store) are being attributed to Steve Jobs, it is worth remembering that the ecosystem for mobile internet, mobile content and applications originated and evolved from Japan some 15 years ago to reach the culmination of the industry as it stands today.

NTT DoCoMo, Japan set the industry trend for mobile content and applications with i-Mode followed by J-Phone (a then joint venture between Japan Telecom and Vodafone) which led the industry for “point and shoot” digital camera integrated with the mobile phone, both of which are integral to the success of smartphones today. NTT DoCoMo’s i-Mode is a mobile content portal site that hosts digital content and applications from 3^rd party content and application providers that are offered as a service to mobile customers for a fee. NTT DoCoMo shares revenue with the content and application providers.

The success of i-Mode was followed by communication service providers (CSP) in every country around the world with their own version of mobile portal such as Vodafone’s VF Live!, Telstra BigPond, Optus Zoo etc. For instance, J-Phone's Sha-Mail, mobile photo-transmission service topped 5 million users within a short period after launch. i-Mode itself commanded some 4,000 registered content providers controlled by NTT DoCoMo (i.e. “Walled Garden”) which was soon followed by about 50,000 unregistered content providers (i.e. “Open Garden”) that resulted in proliferation of content and services in several categories including finance, entertainment, horoscopes, ringtones, concert tickets, navigation, music information and sales, employment,  restaurant information, virtual shopping, online books, m-payment etc.

Despite the phenomenal success and popularity of Facebook.com, Mark Zuckerberg admits mistakes made by his company in getting the mobile strategy right. In fact, mobile phone demands a distinct and different strategy and  is not a portable substitute for the desk-top PC is more aptly described by an associate professor of business at Kobe University's Research Institute for Economics and Business Administration, Jeffrey Lee Funk, in his book titled, “Mobile Internet: How Japan dialled-up and the West disconnected”.

Jeff Funks points out that the success in the fixed-line Internet has blinded many Western nations to the possibilities inherent in the mobile Internet. Jeffrey Funk’s book contains intriguing results from research of the Japanese market and the underlying success of NTT DoCoMo's i-Mode service and a comparison of the Western markets.

Perhaps, the success of NTT DoCoMo and J-Phone can be attributed to the results of careful strategic planning, predictive business modelling and fact-based decisions. In a nutshell, the mobile Internet exploded in Japan due to a superb market feedback loop between the initially chosen content, mobile device capabilities, device prices, packet networks, business models, careful segmentation and user targeting. At first the users were young and the services provided were simple. As the user base expanded, content became more complex, and the devices became much richer.

According to Wikipedia  there are some 675,000 apps available for Android, and the estimated number of applications downloaded from Google Play was 25 billion, but not many are willing to pay for the apps. Apple boasts similar numbers for iOS apps.

Which of the billion mobile applications will dial-up for success and which ones will generate revenue are big question marks now facing the CSP. Without the use of a crystal ball the CSPs need to know which business model will win. The good news is that CSPs have the capabilities to combine different services and applications on demand that will create highly profitable incremental revenues. More importantly, they can use advanced analytics to answer the following types of business questions on the multi-structured and highly complex Big Data:

• Based on a customer’s entire buying history across all product categories, what is the next best offer with respect to call plans and bundles?
• Based on a customer’s buying history in the Apps. Store, what is the next Application most likely to be purchased? e.g. Smartphone Back-up Service

CSPs can use data mining techniques such as collaborative filtering which business analysts can use to find items or events that are frequently paired with other items or events. A related function, Minhash, a probabilistic clustering method can be used to assign a pair of customers to the same cluster with probability proportional to the overlap between the set of items that these customers have bought. This relationship between customers is useful to identify upsell opportunities such as Smartphone applications and accessories to generate additional revenue.

Teradata Aster makes Big Data analytics and MapReduce easier to use through a set of SQL extensions called SQL-MapReduce. By leveraging the 50+ Out Of the Box MapReduce analytic functions, CSPs can exploit their customers’ multi-structured behavioural data to derive immediate business value.

Listen for the dial-tone and get started with SQL-MR to harness the complexity of mobile apps!