The idea of Audience Enrichment has been around for decades in the list-rental business, but it means much more in the offline/online world of data today.

It is no longer theory or even past tense; it is reality. Technology rolls on and affects all industries and in turn, how we do our jobs. In marketing & media, you can trace a path from direct marketing analytics to digital marketing analytics to the current machine learning-led predictive analytics trends we see shaping business operations today.

The progression has been led by advances in computing power, storage capacity, and advances in data science.

As prescient as Gordon Moore was when he made his 1965 prediction about the future of computing power, the announcement of the Netflix Prize in 2006 generated innovation & collaboration around applications of big data, the likes of which had not previously been seen in marketing analytics.

Even with the advances in data enrichment science, when it comes to data, more is not necessarily better. Marketers understandably want to make the best use of their own data, regardless of the quantity or quality of third-party data availability. So how can you increase the value of your data?

There are opportunities for marketers to strengthen the value of their data through various approaches, but it is important to consider the trade-offs associated with each type of enrichment.

1. List Appends via Direct Match:

A deterministic solution. Additional variables are appended to your user-level data to create more descriptive profiles of users. These include attributes such as demographics, auto registrations, lifestyle, and category interest data. Many of the sources used are publically available and can be enrich via deterministic matches.

Trade-Offs

• Massive scale can be achieved
• Match is typically done against a total US adult file of 240mm records with physical addresses
• Offline data is the primary source, which is not refreshed as frequently as online data
• Demographic and lifestyle data is widely available and different sources do not always match each other
• Data must be on-boarded to digital ID’s in order to make actionable for digital campaigns

2. Digital Matching via Persistent Digital Identifier

A deterministic match using some form of a digital ID such as a login to an online account, cookie or device ID.

Trade-Offs

• Less-scale as match rates to other digital ID’s will lower total number of users from the source file
• Allows for greater insight into the consumer journey via cross-device graphs
• Cross-device graphs enable digital targeting & activation
• Sharing 1P data creates an opportunity for data leakage, data ownership rights need to be explicit
• Enrichment is limited to known attributes from web-browsing, app usage, location data, etc.

3. Look-A-Like Models (LAL) using Online/Offline Data

A probabilistic approach providing additional scale to find buyers that look like your target audience. Seed data, otherwise known as the truth-set is modeled. New users, fitting the profile of the truth-set are identified and appended.

Trade-Offs

• LAL models can be quickly translated for use in digital campaigns
• Requires common data between truth-set and larger data pool.
• More data in common produces better models
• Demographics are widely available across data-sets but are known to be poor predictors of behavior
• Truth-sets often consist of loyal or high-volume buyers, which means non-buyers are excluded

4. Social Affinity Audience Enrichment

A probabilistic approach relying on public brand engagements across social platforms.

Trade-Offs

• Massive global scale with tens of thousands of brands eligible to be used for enrichment
• Social affinity learning can be transferred to online & offline data
• Uses same recommendation engine as Netflix and Amazon
• No digital ID or sync required
• Mapping process of taxonomy is a manual process that can take one to two weeks

The Affinity Answers approach to data enrichment puts consumers first. It highlights details previously hidden from view, expanding your data’s potential. For example, Dunkin’ Donuts can leverage their fans interest in Blake Shelton and the Baby Bump app for insights and planning, but not if these insights remain hidden.

Contact us to uncover what’s hiding in your data.