Big data is the thing of the moment. There are posts upon posts extolling the benefits of social media data and the ability to laser focus your marketing to the most relevant audience. But for many brands, the data itself can seem out of reach. It’s one thing to talk about how Facebook can utilise all the data they have in their open graph, but it’s another to understand how to access and analyse that info for yourself – and, understandably, they don’t necessarily make it easy to access all that user info. But there are ways.
With that in mind, here are a couple of methods that anyone can use to squeeze relevant data out of their own Twitter and Facebook communities. It takes some work, but what you reveal could change the way you think about your marketing and the best ways to reach your target audiences. And what’s more, while time itself is an expense, these methods are available totally free of monetary cost.
Analyse Twitter Follower Profiles
There’s a heap of great Twitter analysis tools out there (Followerwonk, Twitonomy, SocialBro and Hashtagify are my personal favourites) and all of them provide great insight, excellent info that you can utilise in your marketing strategy. But one thing they can’t do is show you what, specifically, your followers are tweeting about. Several of them have reports that show you the most common words in Twitter bios or locations your followers, but analysing the actual tweets of users, the content of their 140-character missives, is a little more complex. To do that, it takes a bit more work, some more detailed extraction to pull out all the relevant info.
The best place to start is to work out who your most engaged community members are – Twitonomy gives you a listing of the 10 people you’ve most re-tweeted, most replied to and most mentioned in your tweets, which is a good indication of who the most engaged members are amongst your Twitter community are.
While it is possible to analyse every single one of your followers, you’ll need to process each profile individually, so you’ll probably want to start off small, with a specific group of users whom you know are engaged with your message – like those most replied to. Once you’ve identified the profiles you want to know more about, note their Twitter handles and head over to All My Tweets.
All My Tweets is an app that displays up to 3,200 of any users’ most recent tweets on one page of text. You can hide replies and re-tweets from the list, which narrows the data down to the actual messages that user has created, giving you a clear indication of what they’ve personally made note of. Enter in the first profile on your list, let it load up (can take a couple of minutes), then select all and copy the text over to a Word file – then repeat the process with the next user and paste their tweets on the end of the previous. Once you’ve extracted all the tweet data from the profiles you want to analyse, go to Wordle and paste in the entire document.
Wordle creates word clouds of any text you enter. There’s no character limit, so you can enter in the entire document of tweets from your chosen users then create a word cloud to see the most common things they’ve discussed (you can also view the full list of terms from the drop-down menu). The more data you can enter, the better, though I’d recommend choosing who you analyse judiciously – if they’re not actually engaging or interested in your messaging, they’ll skew your data. While entering the info for each profile does take time, this process enables you to create large datasets which you can analyse to find trends and common interests among your Twitter communities.
Tip: When you cut the text from All My Tweets, you’ll also get the dates of each tweet, so your word cloud will show the first three letters of each month as prominent mentions (e.g. you can see ‘Nov and ‘Oct’ prominently feature in my example above). To fix this, you can do a find and replace in Word to get rid of any mention of these month abbreviations – but make sure you put a space at the end of each abbreviation you want to remove, otherwise you’ll end up with mentions of things like ‘keting’ instead of ‘marketing’.
Use Facebook Graph Search to Analyse Common Interests
Facebook’s Graph Search is a powerful tool for gaining insights into Facebook communities. There are so many ways you can use Graph Search that it’s almost overwhelming – given so many options, where do you begin? What do you look for that’s going to provide the most value for your data profiling purposes?
A great way to use Graph Search is to search for interests of the people who follow your brand page. You could also look up pages liked by people who like your page, but the difference between ‘likes’ and ‘interests’ is significant. People have clicked ‘Like’ on pages for various reasons over time – to enter contests, to follow a campaign, etc. Interests are things that people have specifically noted that they are actually interested in, which makes them more indicative of a person’s preferences than page likes in most cases. Both metrics can provide valuable info, but interests is more indicative of audience preference, in my opinion.
The catch with Graph Search is that the results are not displayed in the order you might expect. The aim of Graph Search is to help you connect with people in your extended communities, so the results of the searches you conduct are based on affinity, which is measured by various factors built into the algorithm. This means the first result on the list won’t necessarily be the one that the majority of your page fans are interested in, it will be the one with the most people you’re connected to, in some way, linked to it.
Facebook also doesn’t show you how many of your fans are interested in the topic – see how the line on each interest says ‘People who like *** like this’ – not ’63 people who like *** like this’? There are a couple of ways to extract that level of detail – the easy way and the hard(er) way. And unfortunately, the amount of likes your page has will dictate which option you take.
The Easy Way
The most straightforward option, as outlined in this excellent post from Simon Penson on the Moz blog, is to utilise the data available in Facebook’s ads manager to determine how many of your page’s fans are interested in each topic. This method requires a few steps, and I doubt I’d do them the justice that Simon has in his post, but effectively if you select your page in the ‘Interests’ section of ads manager, you’ll get an estimated reach figure, indicative of your audience. For example, let’s say you work for Coca Cola – you’d choose ‘Coca-Cola’ as an interest, which would show the potential reach to fans of your brand.
If you then select another interest (choosing from those listed in your initial graph search data), whilst also keeping your page listed, the reach figure will be refined down to people interested in your page AND the chosen interest, enabling you to get some idea of the how popular that topic is amongst your fanbase. It’s a logical and well defined strategy, and well worth a read. The problem with this process, however, is that it only works if the page you’re analysing has a lot of likes. Facebook ads once had a ‘Precise Interests’ option, with which you could target any page you chose, but it was amalgamated into the ‘Broad Category’ option and renamed ‘Interests’, which is what you see now, and in that process, smaller pages were removed from the interests list.
There’s no official explanation on which pages will or won’t show up as interests, but in my investigations, unless a page has around 10k likes (or a heap of engagement), it’s not going to be listed.
The way around this is to use the connection targeting features. In the ‘Connections’ field, choose ‘Only people connected to [your brand] or choose your brand in the ‘Include people who are connected to’ field to narrow down to your audience.
Then, when you choose an interest, it will narrow down to people who are connected to your page and like that topic, giving you an idea of how popular it is amongst your connections – though again, if your Like counts are low, as is the case with many smaller businesses, there may not be enough data to go on and the potential reach won’t show up. So, the hard way it is…
The Hard(er) Way
There other option, while more time consuming, is actually a lot simpler – and the only way to go if you don’t have a high enough like count. This is the ‘lo-fi’ route – no tools, no spreadsheets, you just click into each interest listed on your ‘favourite interests of people who like…’ list and manually count how many people are listed. Of course, this takes time, but you can get a gauge pretty quickly on the interests with a lot of followers and those with a few by scrolling through, and you can move through them fast. And as noted, for pages with fewer than 2k likes, this is the easiest way to get an idea of the most common interests among your page fans. Appealing to those interests in your marketing efforts will increase your chances of likes, and as we all know, likes are the currency of reach in the new Facebook world – if you wanna’ have any hope of getting your message seen, you have to appeal to what your fans are interested in.
Is it worth the effort?
While you’ve probably got a fair idea of what common interests you’d expect amongst your fans, you may be surprised with what you’ll find with a bit of investigation. For example, the brand I’ve analysed here is a company that makes baby clothing and accessories – for them, the most common interests among their Facebook fans are:
There’s no way I’d have picked photography would be a top interest in this group – fashion, yes, crocheting (some of their products are crocheted), but in the results, those top two were a fair way ahead of the others. This is great insight, as the business can now consider how they might appeal to those interests. Using better images or running photo competitions might be one way to work with those interests, but that’s probably too literal – you also need to consider why people like photography. Photography is a largely aspirational interest, it takes people away or re-awakens memories of travels and other times. That ties into the second interest, Bali – approaching those two interests from this perspective, I’m willing to bet a lush-looking photo shoot down by the beach would be a big winner among this audience.
Insights vs Time
These methods are a great way to gather information about your Twitter and Facebook communities and, as noted, all of these tools are freely available to anyone. You can also use these methods to analyse your competitors, if you were so inclined, giving you a better understanding of how to reach their audience and appeal to wider groups outside of your own. The downside of this is that these processes take time. All in all, it shouldn’t take more than a few hours to gather the required info, dependent on the size of your dataset, and that effort could pay off with big engagement benefits, if you can target your findings in the right way.