Wine Interest Chart


Understanding what interests people is the lifeblood of Facebook – be it sports, our friends or even Wine. This keeps the content (and ads) we see on facebook relevant to us.

Each interest is inevitably connected to other interests- For example, intuitively, we know that if a person is interested in wine, a topic like food pairings, cheese would likely come up. In contrast for “beer” the associated concept and context would more likely be along the lines of opening a nice chilled one with pizza and the latest sports match on TV.

Let’s use Wine as an example.

If you play around the interactive chart on top – (click on a blue dot and drag) you could see all the interests Facebook think are connected to wine. Be it direct relations like “cheese”, “gastronomy” or indirectly like other alcoholic drinks such as beer.

Based off this we could find hidden interest targeting that is possibly high ROI but just hidden.

So – how does this work?

On each user, Facebook keeps tabs on our interests. You could find yours here:

For example for myself, my kickstater addiction is showing up strongly on my interest:

Facebook Ad Preferences

These are determined by quite a few ways by Facebook :

  • What you share on your timelines
  • because you liked a Page related to the interest
  • because you installed a related app
  • because you clicked on an ad related to the interest
  • Activities you engage in on and off Facebook related to things like your device usage, purchase behaviors or intents and travel preferences

These interest are then used by marketers to target users along with prefilled suggestions.

It is in Facebook’s interest to provide relevant suggestions to marketers – better suggestions mean better marketing results. Hence we could reverse engineer what Facebook knows about an interest and what other interest Facebook deems most relevant.

With the Facebook API we could get recursively suggestions and suggestions of suggestions …and so on… with a Breadth-first search Algo we appended nodes and connections to a network graph.

In this example for wine, it results in a graph for interests clusters around the theme “Wine”.

You could scroll to the top to interact with the graph and see the following clusters:

  • Directly related to wine is food related: Cheese, Gastronomy,Gourmet
  • Indirectly Beer is related and has its own very interesting cluster that include  pop culture (Movies, Eminem, Rihana) and other expected interests like sports. The linkage even extends to military.
  • A closely relate cluster also formed around wine culture themes like:Viticulture (study of grapes), Oenology(study of winemaking)
  • Wine tasting formed another cluster with associated events like wine festival, judgement of paris.
  • Indirectly wine is related to other alochols and this clustered to focuses of beer and harder liquors.

So why is such a graph interesting?

To name a few:

For the Marketing Strategist – the concepts found could also be used to build upon persona research and develop a more robust understanding of a targeted audience.

To a digital marketer working on Facebook buy, you could directly find relevant interests to your audience and target. This method would quickly uncover topics that are hard to find with the stock suggestion tool.

For a Search SEM marketer – the frustration with standard keyword tools is that the outputs are too direct and often you would miss interesting campaign builds. This helps with indirect terms.

For a Content Marketer – it opens new avenues to explore in crafting compelling experiences that your audience cares about.

The list goes on…

I hope you had as much fun reading this as I did hacking!