Mobile marketing is not a random process, as industry veterans would tell you. With programmatic advertising, business intelligence data analytics becoming more than marketing jargon, advertising is becoming a concept that thrives on both art and science.

There’s a clear shift towards data driven campaigns. The creative and the distributive aspect of it all are being adapted to suit the new programmatic methodology. The creative management platforms, the DSPs are all trying to adapt to data invasion in mobile advertising. Leading to  focussed advertising  that gets you desired results and revenues. The first step in configuring  these campaigns is building audience segments and TGs from the existing data or accumulated BI. The process is as old as the advertising industry itself, the only difference is the technological shift  that has come with internet, data and  connectivity across the globe.

Audience Segmentation Through Ages:

Through ages, science and deduction has gone into gauging popular media for specific audiences.From print to television and now towards digital. In the days of print the targeting was  simple, based on gender specifications and social typecasts. The successful placement of a campaign was limited to sociology and gender studies and a bid on the first few pages, intermediate leafs/page three or sport section depending on the target audience’s likely interests .


Audience targeting through ages: Monroe for westwood luxury cosmetics
Westmore targeting the female audience, but making sure the males spot the ad.

Since press was local and vernacular, the advertising was aimed at groups bound by factors common to geographies gender and culture. In the age of satellite television, prime time Tv became the premier buy for advertising firms, visibility was the name of the game and the customer was spoilt for choices. Internet took over the reigns. With digital media coming in words such as data and analytics became commonplace for brand advertisers. The audience had grown, the industry too. Multiple generations, nationalities and user segments emerged out of nowhere.  These led to more complicated times and advanced mathematical  algorithms for derivation and platforms such as ours for maximizing outreach and narrowing down an international audience into specific segments.

The Mobile Audience Segments:

Mobile Audience : A Fuzzy or Crisp Set?

So how do BI firms get to this data? Usually a set of users is built on loose relations. Logical patterns that may or may not be verified by concrete facts or  frequency. These patterns could be attribution data that has no insights into the usage patterns for apps or this could be the result of ambiguous behaviour and purchases over a short period of time. The quantity of data alone doesn’t help. The quality is of utmost importance. Fuzzy Sets have lots of random assumptions that may result from short term conversion patterns. They may or may not repeat themselves if we check their periodicity over a longer time span. Even for machine learning algorithms to work over them they need periodic checks and frequency measures.The data has to be granular and verified with both slicing(user centric) and dicing (time oriented) practices.

In Mathematical terms if a user is an element in a segment and an event that connects them is an associative relation between all group elements (users) then to declare this data as a classical set or a crisp set the event has to be well defined. For online purchases mobile conversions to label user/device ids as defined elements in a target segment. This is where the algorithmic approach applied by an analytics platform comes into play. While multiple data analytics and BI tools work to extract the same information the algorithms that are used on the top of this data aggregation determine the effectiveness of the tool. It’s justified to say that all user sets are fuzzy till they are justified both by time and narrowed specifics (in some cases more than 40 parameters are applied). So, how do we condense this approach for quality in lesser time and turn these fuzzy sets into crisp sets? The answer lies in two cs: Conversion and Content Consumption.

Data Collection : How To Make Data Driven Decisions and Leverage Them?

Over a relatively smaller period of time mobile audience behaviour can be tracked legitimately from content consumption and transactional conversions(from third party sources) put together:

As an example let’s take the case of an ad campaign of a branded budget hotel. The user groups are narrowed down to business travelers, travel enthusiasts around the year and families ( mostly in holiday season). That’s a broad criteria for all users but how do we specify this further?

There are a number of bookings from different device ids and how will one ever figure out that which belongs to which sub segment in this category? To establish a pattern of a device id owned by a family guy one will wait at least for three annual- bi annual holiday seasons to verify the pattern. Similarly for a business traveller one will have to verify at least three to four successive flight bookings/ travel desk transactions in a short span of time. But what if these bookings are made by a travel desk agent and not the traveller himself? How does one figure out if the user is a genuine catch to air budget room advertisements and what exactly is the user looking for? The answer lies in the content consumption patterns. Clicks and conversions may present a uni- dimensional picture.

However,content consumption through various sources completes the picture. An adventure traveler may be reading different blogs, visiting different sites and looking at niche options that cater to only adventure travellers. Similarly a business traveler is going to be looking at local maps and connecting routes in a city, reading around the more convenient places to access, restaurants and meeting places. The family guy is spotting usual tourist places and well booking patterns will show that he has made multiple bookings in one transaction.

Advertising tracks sources for the click, profiles the source, the time of click, the publisher. Imagine having this information verified from multiple sources. So you know now which users belong to what sub segments thus creating a narrowed audience profile. This is the head start and now you can decide which segment can you appeal to with what kind of offers/content/ discounts. This is how we master building audience segments that help maximize revenue per installs or revenues per clicks.

What is this sorcery? It’s not in-app analytics or attribution data or conversion data alone. It is a more holistic approach to data. The kind that can be implemented using a more holistic tool such as a DMP.

That’s how you accumulate quality data from the right sources and verify it granularly in a small time span.That explained, the rest of the programmatic methodology chain can put your ads in the right space with the right content for the right set of audience.With mTraction DMP you can trust that your audience segments have been built to precision to extract maximum from your campaigns. Our DMP is not limited to the app analytics use case, It works on the transactional behaviour and tracks content consumption on single or multiple devices thus imparting a broader view in a shorter period of time. This broad perspective then can be used to establish specific user sets over the same media or customize audiences with look alike elements and better researched profiles that are more likely to respond to a specific product’s appeal.

Acquiring the right audience is the first step to a successful, profitable ad campaign. Get the right start to streamline the rest of your mobile advertising strategy. Start a discussion to know more or simply leave a query with us on partners@affle.com.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>