In the last few years, the tools to collect and store data of any volume have become available to even the smallest mobile development organizations. The efforts around open source tools and the increase in mobile analytics products that are used to extract value from data suggests that the use of behavioral data to measure and drive business decisions has finally become accessible to organizations that choose to collect the information in the first place.  This means that targeting promotions, recommendations, and advertisements based on traditional marketing cohorts is no longer necessary, as the ability to recognize and characterize returning mobile customers creates opportunities to engage with them in a personalized and more context-driven manner.

A key barrier to creating successful systems for contextual targeting lies in collecting the right data to extract signals from.  Additionally, for data to be useful, it has to be in an intelligent format, highly available, and able to be summarized and visualized appropriately. By gathering the right data in the right manner, you can begin to build cross-session customer profiles that divide users into important cohorts (e.g. those who have made large purchases vs. those who have not) and that also differentiate users with otherwise similar aggregate behaviors into useful segments (those who spend $1,200 per year in your online store may make one huge checkout in late November or twelve smaller once-a-month $100 purchases). This cohort creation based on organic or business rules allows for more relevant targeting opportunities.

1-1Mobile_finalCreating cohorts through the collection, analysis, segmentation, and categorization of user data is only the first portion of a highly effective targeting equation. There is also an incredible amount of context present at the point which data is gathered that is traditionally ignored or available far too late due to the latency between the collection systems and updates to a targeting and promotion system. While static profiles traditionally used for targeting, when it even exists (e.g. age, gender, home zip code), may be useful, the rich contextual elements of “user in cohort ‘rich_suburbanite’ just searched for ‘restaurant’ and is in downtown Seattle” add the context needed to make a more intelligent, relevant (and therefore higher value) recommendation to that user. This use of contextual targeting is a win-win-win for the user, the business, and the ad network: The user gets a useful suggestion, the restaurant gets a likely missed customer, and the ad provider gets higher CPC rate for the targeted ad. It’s this intersection of context with profile based information that makes for a more highly relevant recommendation leading to a better user experience and higher monetization for advertisers. Increasingly, mobile has become a more personal experience than the traditional PC with people now spending more time watching their phones than TV  and nearly 60% of people admitting they don’t go an hour without checking their phones.  These trends highlight the necessity that information, content and ads on mobile must be more targeted or risk being irrelevant and discarded. Without the ability to recognize, update, and act on this context in real-time, any promotion system is limited to targeting based on either a static profile or worse.

Lastly, as someone who deals with data every day, I can’t stress enough that expanding the ways in which we incorporate the behavioral data we gather is imperative to remaining competitive. This means that as much as possible, data in its raw form must be thorough, unambiguous, and of the highest quality. Organizations that prioritize data as an asset  will be rewarded with the opportunity to create rich customer profiles and contextual targeting that provide opportunities for improving mobile customer satisfaction and monetization.