AnalyticsMining Social Data to Create a Content Strategy

Mining Social Data to Create a Content Strategy

You can apply a smorgasbord of social data – from Facebook, Twitter, Google, Pinterest, and tools – in your strategic work to inform decision-making. But one of the most useful, and relevant currently, is in helping steer content ideas and strategy.


Social is a hugely exciting space right now. For businesses it gives us, for the first time, the ability to connect with customers individually, and in an incredibly targeted way.

For search and content marketers it isn’t just the platform but also the data behind it that makes it really exciting. So how can you make use of what will become the most valuable available pot of data?

You can apply the data in much of your strategic work to inform decision-making, but one of the most useful, and relevant currently, is in helping steer content ideas and strategy.

Big Data

Big data startups are crawling all over the social space right now, and for good reason. Social data has the potential to unlock vast amounts of insight about brands and associations, giving businesses information that they have craved for years.

Facebook, Twitter, Pinterest, and more serve up a smorgasbord of info for those that know how to access it and where to look.

This kind of data can help you build out audience and persona profiling pieces and increasingly to inform your strategies for reaching them.

To do that requires an understanding of APIs and some elbow grease, but a lot of the hard work can also be taken out by choosing tools wisely.


Where social data gets really interesting for marketers is in the associations between people and brands. Increasingly, individuals are Liking brand and company pages. By mining Facebook’s Open Graph data you can get some amazing insight into what makes them tick – including some eye opening info about how wrong your assumptions can be.

Investing in content isn’t cheap. To make the return on investment (ROI) work you must almost guarantee that an idea will work. In practice this is incredibly difficult to do but with access to the right data you can severely increase an audiences’ propensity to share and digest whatever it is you want to create for them.

One way to do this is to pull data from key relevant pages across Facebook so that you can see exactly what is being shared.

What Facebook Stores About You

To give you some idea of the depth of data that Facebook holds about us all, it’s worth pointing out the famous example of Austrian student Max Schrems, who requested a copy of all of his data from the Palo Alto, Calif.-based corporation. The result was a 1,200 page PDF that included everything from a field called “last location” to what he alleges was old “deleted” messages, posts and even Poke and IP addresses.

Facebook’s true value is in their data and targeting capabilities. Facebook are combining their huge graph of data on individuals with real world purchase and preference data from a number of suppliers, meaning that soon they’ll know even more about us than ever before.

If you don’t have a suite of tools, it’s possible to better correlate that data manually, even without API access. To do that you can use tools such as Facebook’s Power Editor & Google’s Display Ad Planner, where you can see data based on potential audience sizes for planning purposes.

Analyze the content for the brand in question across all social media pages, and combine this with Google interest sets and other data sources to allow you to identify content opportunities and appropriate weighting. You can then give each brand 1,000 “content points” which are distributed among topics based on their relevance.

This then allows you to identify whether a brand should be focusing on branded content or spending more time on interest led content for example as a first step. The image below shows how this data looks once polished.


In this example the data has told us that for this particular brand (in the alcoholic beverages space) has a social audience that likes “brand content” on page but also engages with celebrity content (see the deeper dive right hand ring for who). This gives you an immediate hook for a content brainstorm, to ensure you have enough of the right kind of celebrity-edge content for the Page.

Google Pitches in

The data dive can then go deeper still and extract information from both social and AdWords’ Display ad group Interest Categories based on specific interests (such as specific celebrities, make up brands, and car makes for instance). This can give you more information about other things your audience is interested in.

To see how Google currently “categorizes” you log in and visit this link. The results are eye opening! This post gives you even more info about other ways you can utilize Google’s ad targeting data to help.

Another great and under utilized tool is the search giant’s Display Network Ad Planner. Its GDN Research Tool, or Audience Builder doesn’t dive massively deeply but provides easy access to data sets around specific interest points. You can work out potential audience sizes based on everything from location and language to specific interests and even sites. It also works as an outreach tool to find the biggest sites in very specific niches.


Again, below is that data in “polished” form for the same brand, showing a huge correlation between them and specific beauty brands. Again, this is great content insight.


Social Data Tools

As well as mining Facebook, you can also look to existing tools to help paint a picture of your audience and provide inspiration for what they are currently sharing and what makes them tick.

You can get carried away with a lot of this, but through trial and error a fairly simple process will provide the insight you need using the following tools:

  • Social Crawlytics: A great tool to give you unique insight into the content strategies of your competitors, telling you straight what has worked and what hasn’t. This tactic is great for a business of any size, but particularly cool for those attempting to take on bigger players in the market without having to simply outspend them. You can use it to pull the most shared URLs into csv from the top three sites in the vertical and aggregate so you can then look for subject matter that is shared consistently at a high level. Add these topics to the content plan.
  • Bottlenose: For those have yet to jump on this tool, do so. It uses a fairly sophisticated algorithm to sort through the cacophony of social noise to surface, by keyword, the most shared content of recent days. As it’s in beta of sorts and its API is still in testing you’re limited right now to what you can pull, but in future this will be a key tool. The ability to extract message analytics and to unify user interactions at scale will provide amazing detail about “live” trends. This is the next level of “big data” for marketers. Being able to quickly form opinions about whether something is “going viral” will allow us to quickly adapt content strategy to make the most of “now” before it happens.
  • Pinalytics: It’s still very early days for those building tools to extract useful analytical data from Pinterest, but Pinalytics is one that does a decent job at present. Having a view on what kinds of content is being repined and shared is important and should help inform your strategy. If it works on Pinterest, then Facebook will act similarly – particularly as the latter has made imagery much more prominent and, therefore, engaging.
  • SocialBakers: SocialBakers free products can help identify seed pages for creating interest sets, particularly in verticals or markets where you are less familiar. The free tools rank pages by size by vertical and by country, and also show a snapshot of the fastest “growers” and those on the decline and can be a useful tool for competitor identification and tracking.

How to Pull it Together

As with all data analysis the key is not drowning in too much of it. Take top fives from each category of data (or tool used) as this keeps it simple.

Bring this data into your idea creation meetings. Then follow a strictly defined process to ensure you brainstorm around as many “pillars” as possible.

Those pillars include looking at everything from the content needs of existing marketing personas and keyword opportunities to semantic phrasing and back again. It results in a data-informed strategy that leaves no stone unturned.



Data-Driven Market Research and Competitive Analysis

whitepaper | Market Research Data-Driven Market Research and Competitive Analysis

Semrush Keyword Difficulty

whitepaper | Analytics Semrush Keyword Difficulty

Searchmetrics Core Web Vitals Study

whitepaper | Analytics Searchmetrics Core Web Vitals Study

The Ultimate Guide to Forum Link Building in 2020

whitepaper | Analytics The Ultimate Guide to Forum Link Building in 2020