Discovering the Power of Social TV Data for Advertisers

By Pablo Gerstenfeld, Director of Research & Analytics, Seevibes
@PabloGerstenfel

Most Canadians share their points of views and interests daily on various social networks – over 20 million Canadians hold active accounts on Facebook and 7 million Tweets.

As Seevibes crunches big data from social television over the past 4 years, we can conclude that at least 35% of Canadian Facebook users have engaged around TV related content and 65% are on Twitter. At the same time, 18% of Canadians engage with brands on Facebook and 21% on Twitter. This represents a valuable opportunity for companies to gain a better understanding of existing clients and to attract new customers to their brand.

While most of this data is public & accessible, there are tough challenges due to the vast amount of interactions generated every day.

Harvesting pertinent data

Traffic on social media sites is so large, it needs to be filtered and structured. On Twitter, we can infer someone’s preferences by analysing his Tweets/retweets and the accounts he follows.

Example:
Screen Shot 2015-04-15 at 10.18.48 AM

Therefore we know that @nadia_fordham has expressed her opinion towards the brand Coca-Cola. But does @nadia_fordham like or dislike this brand? Running a sentiment analysis will help. We can thus infer that @nadia_fordham likes this product.

We identified that @nadia_fordham also follows the TV show “Masterchef Canada”. Being a follower of a certain brand or TV show would mean a positive interest towards it.

Putting the puzzle together

Now, we’re able to say that @nadia_fordham likes Coca-Cola AND “Masterchef Canada”. We can eventually infer that she has an interest in Cooking Shows on TV and may also participate in cooking from time-to-time. From this, we are able to highlight general trends about consumers to help advertisers make better decisions.

We know that people engaging on TV shows are way more numerous and 3 times more engaged than people engaged only on Brands, and 79% of profiles engaged on brands also engage on TV and they generate more than 85% of all impressions.

By understanding the TV preferences of a certain brand’s consumers, we can target an audience that is naturally more engaged. In order to discover general affinities we suggest two KPIs.

Affinity rate: The percentage of Audience A that is also engaged towards Audience B.
Example: 15% of Masterchef Canada Twitter audiences show an affinity towards an “automobile” brand.

Affinity Index: But is 15% high, low or can it be considered a normal affinity rating?
It’s essential to put it in perspective to another comparable population. A good option is to contrast the Brand Affinity for one TV show against all other TV shows regarding the same brand.

Examples:
Masterchef Canada: Auto Brand Affinity 15%
Dragons’ Den: Auto Brand Affinity 8%
Average TV shows: Auto Brand Affinity 10%

At the end, Masterchef Canada shows 87% more affinity than Dragons Den and 50% more affinity than the average of all Canadian TV Shows, which makes it the best target for any automobile social media actions.

How advertisers can apply Brand Affinity analysis

Knowing the type of content that your customers prefer, will help you select the most suitable media mix for your brand or product.

When launching a social media campaign, focus on these profiles that show a specific interest on a certain product. Sometimes, the preferences of an audience are less obvious.

A Brand Affinity study helps identify complementary products or content appreciated by your audience that present a higher engagement on social media and hence, will help in prospecting.

Social TV data is a gold mine for advertisers. However, understanding your client base is more important than ever when “Marketing Smart” in the digital age.