Will the real user please stand up: Identifying and targeting audience segments
Mobile advertising’s promise lies in truly enabling marketers to connect with the right individual at the right time and in the right place. The User is key to targeting the right individual and is central to driving the mobile advertising revolution. The technological challenge lies in identifying users and their personas as they engage with a multitude of internet-enabled devices including smartphones, tablets, consoles in cars, smart-watches, and wearable devices like Google Glass.
So, how does one really go about identifying the user? The process of analyzing user data to engage the right user can be broken down into several steps:
- Data Ingestion & enrichment:A mobile device will share over 50 signals when requesting for an ad.Not all of the shared data is immediately usable. For example, a publisher may choose not to share the location data of its users - it may still share their IP addresses without divulging personally identifiable information. Data enrichment vendors help increase the semantic meaning of the data points collected per user enabling better and more accurate inference of profile and personas. In the above scenario, one of many data enrichment vendors helps us map IP addresses to geographical co-ordinates that allows for richer targeting. Further, external data partnerships are leveraged enabling advertisers to engage branded data segments specific to their industry.
- User Identification and de-duplication:The data from multiple sources is processed and filtered to determine unique users coming from various publishers and integrations. The efficacy of InMobi’s dedupe algorithms enables us to improve our data signals by over 20%.
- Inferring Personas:InMobi’s terabyte-scale processing layer mines billions of data points collected daily to help infer profile and personas for users. By analyzing multiple data-points including ad requests, clicks, impressions, and ad interactions, InMobi leverages its own first-party data to identify personas based on apps installed, frequently-used apps, locations frequented, etc.
The above activities have resulted in InMobi categorizing over 691 million unique users in over 17,000 audience segments to suit the diverse targeting requirements of advertisers. Interestingly enough, at no point in this process is the user’s personally identifiable information made available to InMobi. The advances in mobile technology will continue to push the envelope in interaction design and with it will come a flood of even more data-points. The thought that excites me is what we can infer from that data and how we’d go about intelligently defining audience segments valuable to advertisers.