67 %
33 %
Information about MMA-SAOPABLO

Published on March 17, 2009

Author: ikensolutions


MOOGA: UNDERSTANDING THE USERArtificial intelligence applied to content delivery : MOOGA: UNDERSTANDING THE USERArtificial intelligence applied to content delivery Slide 2: Iken Solutions Iken is a software product company specialized in intelligent business systems backed by hybrid AI (Artificial Intelligence) techniques (expert system, case-based reasoning, neural networks and genetic algorithms). It is an IIT Bombay research spin-off. Iken solutions awards: Iken Solutions finalist satrtup in NASSCOM Innovation Awards 2008 iKen Solutions selected by Microsoft to participate in Le Web ´08 as one of the Top 10 innovative startups in the world Winner TIE CANAAN entrepreneur challenge 2008. Iken has been awarded with a direct entry into the Microsoft Startup Accelerator Program 2008 Silver award for “Best Technology Innovation” at the Mobile Content Awards 2008 in London. Slide 3: The opportunity Slide 4: 6.760.177.421 WORLD POPULATION Slide 5: 3.3 Billion Mobile phones Slide 6: 3.3 Billion in Perspective 1.4 Billion Credit card Source: Slide 7: Mobile advertising Source : "The next big wave in advertising is the mobile internet" "mobile will be a larger business than the PC-Web. But it will take a few years." Eric Schmidt Google CEO Experts have predicted that over the next 5 years, mobile advertising will become a $6.5 billion business. Mobile Advertising: After the Growing Pains Slide 8: The problems Slide 9: Source: Telecom-DATA based on data from operators and regulators. 57 % Latin America 87 % ! Source: Telecom-DATA based on data from operators and regulators. The unknown user!! World Mobile phones are Pre Paid Slide 10: Handset support (SMS, WAP, download) 90% Usage (SMS) 60% Usage (WAP) 20% Usage (download) 5% Sources: Telephia and M:Metrics: Warning: actual usage of technologies differs from technical penetration Actual content usage The “Long Tail” in the mobile space : The “Long Tail” in the mobile space Current Scenario – User Perspective : Current Scenario – User Perspective Everybody gets the same Slide 13: Learning from successful models: The evolution of digital business The Long tail : The Long tail The Pareto Principle: 80/20 rule, traditional pattern of sales concentration. C. Anderson (WiredMagazine-2004): The Long Tail: By lowering inventory storage, distribution and search costs, digital markets have the potential to increase the collective share of niche products, creating a longer tail in the distribution of sales. Now there are 2 markets to attack: Old Market: Top Sellers. New Market: Long Tail. What percentage of your items are you selling at least once a month? The Wisdom of crows : The Wisdom of crows James Surowiecki (2004): Wisdom of Crowds A diverse collection of independently-deciding individuals is likely to make decisions and predictions better than individuals or even experts. Lego: encourages its fanatical customers to design their products. Companies pay solvers from $10k to $100k per solution. More than 30 percent of the problems posted on the site have been cracked. The Many Are Smarter Than the Few !! Long tail + wisdom of crows in action !!! : Long tail + wisdom of crows in action !!! Amazon: over 25% of Amazon’s book sales come from books not available in brick and mortar stores. Rhapsody: 22% of sales are from songs not played on traditional media. 53% of 750k Rhapsody´s songs are streamed at least once a month. Netflix: 20% of DVD rentals are documentaries, B class movies and niche movies. Google: Niche advertisers provide Google over 60% of their Ad Words revenues Make everything available, help your customer find it and you will increase your sales. Slide 17:  What should we do ? Slide 18: MOOGA Understanding the user !! Personalization is the key ! : Dynamic personalized one to one offer Personalization is the key ! MOOGA personalization : MOOGA personalization Non-intrusive Must not involve explicit user input or action. Adaptive Continually learns about user individual interests and behaviour Dynamic Real time Intelligent Automatically derive individual user behavioural patterns Identify common behaviour across users “wisdom of crows” Able to predict subscriber’s future interests Context aware Understand the context of the user request Understands the subscriber’s environment MOOGA in action ! : MOOGA in action ! RESHAPE: A whole new mobile world reshapes and personalizes the subscriber storefronts on mobile for users REMEMBER: Your personal mobile friend and butler AI remembers what you like and dislike RECOMMEND: Intelligently knows your likes and dislikes Detailed profiling to ensure that the user is enjoying the moments and experience on mobile REFERRAL: Share the information Allows the user to share their findings with their networks REWARD: Set off the multiplier effect Let your subscribers be your sales agents and help to sell on the network Slide 22: Intelligent Ads and Premium Portals The ad is intelligently selected by the AI as the user has a preference for Hollywood movies. Clicking on the ad will bring the use to relevant WAP page Viral Distribution and Reward : Viral Distribution and Reward "64% of consumers will try something recommended by a friend. 69% of will forward something they like to from 2 to 6 friends.”  Jupiter Research Sept 2006 Slide 24: MOOGA Validation Slide 25: Case study Introduction : Introduction Sony BMG was looking to enter the market with a differentiated service that was unlike any other competitor offering. The approach taken by Mooga was seen as ground breaking and market leading. Mooga is currently administrating SonyBMG´s portal in Argentina, Chile and being launched in other Latam countries. This is the first mobile music portal administrated by artificial intelligence in the market. Overview : Overview Auto Context Classification of existing meta-data/information/classification can be used to automate the process of contextualization. The SONYBMG Contextual Framework has been automatically built with our auto-context scripting engine using SonyBMG´s existing metadata, enabling a rapid deployment. Overview : Overview SonyBMG´s portal structure Top 5 Section Premium placed content defined by SonyBMG Shows same content to each and every user (SonyBMG´s request). One and only section not being administrated dynamically by Artificial intelligence. Artists Fully dynamic section administrated by AI. Presents relevant content according to each user profile. All content presented in context. Mooga allows to mix dynamic and static sections to promote specific content due to marketing needs. Overview : Overview SonyBMG´s portal structure Context, context, context! Content is placed in easily recognised categories. Provides with ALL available content related to the category. Takes full advantage of IMPULSE increasing multiple transactions per session. Recommendations Fully dynamic section administrated by AI. Presents relevant content according to each user profile. All content is presented in context. In order to find a content, you must know what you are looking for. Mooga discovers relevant content for each user. Overview : Overview SonyBMG´s portal structure Search functionality Uses a “wisdom of crowds” ranking algorithm. Its the quickest path to access items of interest. Ranking is dynamic and continuously changes as the user activities evolve. Search results provide a 3 tier view to useful and relevant content as the crowd knows best what everyone wants! Overview : Overview SonyBMG´s portal structure Browsing functionality: Mooga uses a “Top Down” approach. This is the most effective path for a user to access items of interest. Every list have been structured using intelligence based on entire portal activity of users. Ease of content access coupled with implicit recommendations powered by Artificial Intelligence Slide 32: Statistical Validation Statistical validation : Statistical validation Give each customer WHAT THEY WANT, not what you think they need: Static Top5 made 25% of total sales meanwhile the Dynamic Storefront and the Recommendation engine (AI) administrated 75%. Up to 50% of users who clicked on a recommendation ended making a download. CONTEXT IS KING Almost 20% of users made multiple downloads from same artist. Statistical validation : Statistical validation Give customers INFINITE CHOICE and they will make infinite choices 90,7% of the content available was download at least once. 98,9% of artists got at least one download. Help your customers FIND RELEVANT CONTENT More than 22% of subscribers used the recommendation engine. 24,3% users made more than one download in a single session. Antonio PeñaGeneral Manager : Antonio PeñaGeneral Manager

Add a comment

Related presentations