The ability to understand the interests of each individual, better than anyone else, makes a difference.
This is the case with Netflix, which has made it a competitive technological advantage. As a matter of fact, it owes its success to an algorithm that analyzes the interests of its audience, guiding the viewing experience with tailored recommendations, but also orienting the production of new, original content by the company.
Most Analytics platforms used by companies, however, determine a user’s interests solely on their navigation path, just tracking their visits to content.
On the other hand, big digital players such as Facebook, Google, Amazon, Instagram and Netflix go beyond this simplistic and mechanical collection of interests with a more consistent one, capable of evaluating the different interactions of users with content: the Interest Graph.
This digital representation of the interests of single users, or of groups of users, is structured as a hierarchical graph based on three elements:
- nodes: the characterization of interests
- connections: the link between interests
- the number of users who share a certain interest.
Three steps lead to the creation of an Interest Graph:
The user navigates the website and explore the content hosted there.
Data about each action performed by the user is collected. HITs are fully exploited: those behaviors and interactions activate a monitoring code, signaling the Analytics system to record them.
We are referring to indicators such as pageviews, social shares, comments, clicks, video views, time on page, traffic sources, internal and external searches, purchases, etc.
The Interest Graph processes every HIT collected, defining the single interests and the relationships between them. It is not misdirected by any volatile interest (e.g. a single pageview): the result is a normalization of the weights of the various interests, a precise picture of each visitor’s tastes.
They can then be divided into two types: inside interest, the ones shown on a specific site, and outside interests, the ones collected on the network – a group of websites - the Interest Graph applies on.
Why apply Interest Graph to e-commerce?
Thanks to the IG, you can identify which users are closest to conversion, and set up targeted remarketing. Similarly, you can avoid wasting time and resources with those who have not shown interest.
- SEO strategy
By understanding what the interests of Search Marketing personas are, you can plan an editorial calendar with a SEO strategy based on what users really enjoy on your websites or network.
- Personalization - A/B testing
Being able to determine which users are most influenced by your campaigns allows you to identify the clusters with the highest return on investment, maximizing communication effectiveness.
- Behavior prediction
Pinpoint the user behaviors/interests most useful to your business: you can use this information to set positive or negative remarketing actions.
- Marketplace analysis
With the IG, you can understand which product categories your users are most interested in, thus getting ideas to reorganize your e-commerce structure to highlight those types of products.
- Predict the interests of new users
The data collected can be used to create a model that, based on previous interactions, will predict the interests of newcomers, and react accordingly.
Among the platforms that can be used to determine IGs is DataLysm, the Data Intelligence solution developed by 3rdPlace.
There are also Content Intelligence solutions (the Saas DAM THRON is one of those) which perform a classification of content through AI algorithms, associating descriptive metadata to the users who have visited content: tag = user interests. In this way, it is possible to know which editorial characteristics (topics, formats, etc.) have pushed users to convert the most.
This data, if integrated with your CRM, allows you to keep an eye over the entire navigation path of users (even of anonymous ones), so nothing is left to chance: what you get is a complete, single customer view with which to feed targeted marketing initiatives.
Q: Why do you think it’s better to use a targeting function based on user interests, rather than on socio-demographic characteristics alone?
A: Working exclusively on a socio-demographic targeting may lead to hit users that are out of target, not aligned with our objectives, generating dispersion.
I’ll try to explain it with an example: imagine I am BMW, and I want to push the sales of BMW X5 SUV by addressing people with a certain spending power, and targeting only men and women between 35 and 60 years old. But his could be too wide a target: in it, we’ll likely find users who are not completely interested in such a car.
But if, instead, I add the interest in BMW and in similar competing brands (eg Mercedes) and the interest in SUVs to the socio-demographic segmentation, I’ll hit a more precise target, with minor dispersion and the chance to invest my budget in a more structured way.