From Macroscopic to Microscopic: Artificial Intelligence enabled audience profiling in MOMAGIC

In Artificial Intelligence

Our Goal

“Sell the right product to the right customer” is a dream of every marketers and advertisers. As a leading digital marketing company in Asia, MoMAGIC is approaching this ultimate goal by launching two data-driven solutions: TrueReach and TrueInsight, which is capable of understanding and targeting audiences from both macroscopic and microscopic point of view. In this article, we will share the core ideas behind our solutions and some insights we learned when playing with large-scale real data.

MoMAGIC TrueInsight: Data Management Platform

As the backbone of all MoMAGIC services, TrueInsight integrates large-scale, heterogeneous data sources (e.g., AD-request, web behaviors, etc.) and transforms those high-frequent, noisy dataflows into structured datasets.

In order to complete those challenging tasks quickly and accurately, we first carefully designed the data processing pipelines of TrueInsight in a parallel and distributed manner to guarantee its performance scalability, which means TrueInsight can process large-scale data from multiple sources without scarifying its performance. Besides, we developed a generalized data adapter which is capable of handling most of the data formats and transferring those heterogeneous data into tables with expected schema to enable its feature scalability. Finally, we developed a reliable data/task monitoring mechanism upon virtualization technologies to fulfill the maintenance scalability , which allow TrueInsight to replaces failed components, or requests the missing data intelligently. With above 3- dimensional scalabilities, we not only ensure the reliability and accuracy of current data flows, but also enable the flexibility to integrate more data sources in the future.

Data flows of MoMAGIC DMP, TureReach and TrueInsight

The Intelligence of MoMAGIC TrueInsight: Audience Explorer

Audience Explorer helps you to grasp a group of users’ preference and trend in few seconds. It illustrate hundreds of dimensions such as demographics, residence, device, application preferences and even lifestyles.

Since Audience Explorer is tightly combined to our TrueInsight, you can arbitrarily select a group of users (i.e., segments) and begin to profile your users. You can drill down to any dimensions that you interested and understand more about your audiences (e.g., what is the preference of Noida Residents, who is active in the morning and using Samsung Note 9?)

Last but not least, we built interact dashboard as shown in figure 2 to demonstrate all those information and make it easy-to-understand in a simple glance, which provide you clear, real-time trends during the whole ad-serving cycle.

Audience Explorer

The AI Essence of MoMAGIC TrueReach: Deep Neural Network Enabled Segmentation Recommender

“Machine learning is a field of computer science that gives computers the ability to learn without being explicitly programmed”, (Arthur Samuel, 1959). In MoMAGIC, we formulate a segment generation process as an AD-recommendation task and generate customized segments for your campaigns intelligently.

We have designed and developed a 2-winged deep neural network segment recommender (DNN-SegRec), which modeling user preferences on different campaigns by learning the latent correlations between user attributes, campaign creatives, and background contexts in billions of AD-logs.

Figure 3 shows the structure of our model. We implemented pipelines which trans- fer AD-logs into training set by encoding user and campaign information (collected in TrueInsight) with thousands-dimensional vectors, and then input those vectors into the model for learning above-mentioned latent correlations.

Those latent correlations are difficult/impossible to obtain with human efforts, (e.g., how a user’s IP address sequences, his/her APP usages correlated to his/her click-through rate, given sentences in a creative?) yet they are the keys of mining business values from large-scale data. We conducted 2 off-line tests with the DNN-SegRec and its CPI surpassed manual targeting by 31%-34%, and more importantly, the generated segment kept most of the conversions.

2 – winged deep neural network segment recommender

User-Centered AI Marketing

Aim at intelligently profiling/targeting audiences from macroscopic to microscopic points of view, we designed TrueInsight and TrueReach with a user-centered principle: the best way to find the ’right’ audiences begin with understanding user needs, and the best way to run a marketing campaign is to match, or even arouse those needs. We also built those AI marketing solutions upon state-of-the-art data science and software engineering technologies to make sure that our clients and partners will be enabled with an informative and insightful capability to enhance their competition advantages. We are ready to be your smart second brain in the AI marketing and business intelligence environment!

(The article is authored by Mr. Jack Ho, Senior Data Scientist; Mr. Steven Chiu, VP – Product & Technology; Mr. Jason Wu, Chief Strategy Officer and Mr. Rune Hsiao from MoMAGIC Technologies, Taipei.)

You may also read!

Why is FaceApp Being Accused of Selling User Data?

If you’ve been online on any social media platform recently, you must have seen photos in which people look much older

Read More...

The Obstacles That Are Stopping India from Being an AI Powerhouse!

The Niti Aayog is set to head a national program on AI which is supposed to work on the

Read More...

UPI: A Comparison between New and Old Online Payment Systems

The first online transaction was supposedly marijuana, which was sold by some Stanford students to students at MIT over

Read More...

Leave a reply:

Your email address will not be published.

Mobile Sliding Menu