27 Jan

Breaking Down Mobile Ad Targeting

Mobile advertising technology is continuously improving. The ever-changing landscape is a result of consistent improvements to hardware and software, fragmented components and multiple operating systems available in the market, updates to wireless infrastructure, shifts in industry regulation, and consumer privacy concerns as a result of a lag in market education on the innovations aforementioned.

The result: niche solutions to target users on mobile devices.

The following is a breakdown of the five methodologies which I see are most prevalent.

1. Premium Publishers, Premium Advertisers

Early desktop display networks managed the quality of both the publishers’ inventory and advertisers’ creative, similarly new mobile networks are now gaining momentum by using similar tactics. By maintaining the quality of the network, mobile ad networks are simply aligning quality creative with premium inventory to ensure a certain caliber audience is exposed to the campaign. Due to the assurances the closed network presents to advertisers, the publishers can charge based on Cost Per Milles (CPMs) – versus Cost Per Action (CPA).

2. Bid Request & User Profiles

Prior to desktop display inventory being bought and sold on a per-impression basis, direct buys were the norm. A marketer was essentially buying impressions in bulk in order to have the ads seen in a specific context (e.g., on ESPN.com) – similar to traditional, print media. Some additional targeting options were available (such as geography, browser type, time of day, etc.) but they still limited the opportunity to truly optimize performance – bidding based on individual impressions, taking into account viewer’s demographics and purchase intent.

Mobile focused companies profile users based on the device-level attributes available, similar to audience segmentation techniques on traditionally seen on desktop. As a result, inventory available on highly attractive consumers’ phones/tablets sell for a premium via Real-Time Bidding and Programmatic buys. Two differing, but potentially complementary, techniques are described below:

i) Bid Request Data: These attributes are exposed in real-time in bid requests from publishers. While the data varies depending on in-app inventory versus mobile web, the following are exposed in both: operating system (e.g., iOS, Android, Windows, etc.), device hardware (e.g., iPhone, Samsung Galaxy S3, etc.), contextual details (e.g., app name, IAB category, publisher site, etc.), language (limited to browser), and geolocation (based on availability & limited in accuracy – lat/long). Traditional desktop targeting qualifiers available for direct-buys like the time of day, are also available as variables to the advertiser when setting up buying parameters on mobile. Relying on bid request data with no further data decoration is very inaccurate – I find that marketers are making assumptions based on simple intuition (i.e., owners of HTC GE are wealthy due to the cost of the hardware).

ii) User Profiles: Persistent first- and third party cookies on desktop enable historical user profiles to be built based on past impressions served, conversions, and historical browsing patterns. Due to the rapid depreciation of cookies on mobile web, tying browsing behavior and historical events to form a profile centered around a cookie as an identifier is time better spent. However, both iOS and Android have persistent identifiers, device IDs (IDFA and AdID/Android-ID, respectfully), that act as anchors around which to form consumer profiles. Different players in the space approach building historical usage and performance profiles from various datasets. Flurry utilizes in-app analytics to determine user personas that encompass user interest. Admobius logs bid request data from a variety of Supply Side Partners (SSPs) and performance data in hopes of forming a comprehensive view on a per user level.

3. Omni-Channel Approaches to Expand Desktop Budgets

Educating media buyers takes time. However, minimizing market education on mobile tech and simply providing the option to expand an existing desktop buy to the same audience on mobile – well, that is an easy sell. Through i) econometric modeling (a.k.a. statistical IDs) and ii) registration data (a.k.a. user login and/or personally identifiable information [PII]) desktop users can now be found with varying amounts of accuracy across multiple digital touch points. Enabling a desktop segment to be targeted across screens means that established and trusted desktop data sources can be leveraged for omni-channel campaigns – or simply mobile only buys.

4. Location

As I shared in my post on Data & Context in 2014, the rapid adoption of mobile devices means wide availability of location data about consumers, their travel patterns, and shifts in those patterns. Businesses like PlaceIQ & Factual rely on location signals to target users, build individual profiles associated with device IDs, and determine attribution for a media campaign based on shifts in consumer foot-traffic. The industry is embracing this shift, and, with the recent release of Apple’s iBeacon, we expect that location data will become more accurate and actionable.

5. Social

Social platforms have a vast amount of both structured and unstructured data on their users. From declared attributes – such as age, gender, education, profession, location, close friends and family to inferred attributes – including social interests, travel habits, political views, diet, etc. – Facebook and Twitter are well positioned to optimize which advertiser’s creative to display to whom, and when.

Both networks have successfully introduced ad units to their desktop sites and mobile applications, but neither have successfully transitioned their extensive datasets from their owned inventory to external supply sources. However, with Twitter’s acquisition of mobile exchange MoPub, it can extend its audience targeting, via mobile users and their associated static device IDs, to in-app. Similarly, Facebook has recently announced testing its own mobile ad network, which will result in improved accuracy of native mobile ads, provide greater reach for Facebook advertisers, and help mobile publishers monetize their apps.

[Thanks to Andrew Eifler & Tara Johnson for contributing to and editing this post.]