This post should be written by one of those SAAS platforms that sell social listening services. Alas, they’re always trying to overpromise and underdeliver on the goods, which is why many of their clients get disillusioned by the effectiveness of this technology.
In this post, I hope to do a better job and explain why they do provide an important service, from the point of view of a business client.
What is social listening?
Every day, millions of posts, comments and articles are published online. Through APIs and web crawlers, services like Buzzsumo and Meltwater are able to pick up these content, filter them by keywords that you define, and deliver useful insights like share of voice, consumer sentiment and highlight trending issues.
But first, a quick refresher if these terms aren’t familiar.
What you can expect from social listening reports
Share of voice
The amount of mentions that are collected for a particular keyword, containing your brand and that keyword. You then compare the volume of mentions that your competitor has collected on this keyword as well, and map them out as a percentage.
Example – If you sell stock images, you can set up 4 competitors topics: “Your company + asian faces”, “Shutterstock + asian faces”, “Getty Images + asian faces” and “Adobe Stock + asian faces”. Then for each competitor topics, calculate the percentages, indicating whose company is more widely associated with having a better stock library of asian faces.
In this age of Artificial Intelligence, many SAAS platforms claim to have close to 80% or higher accuracy in understanding human conversations.
Hence, they can read millions of mentions, categorise each by a negative, positive or neutral sentiment, and eventually tell you how the majority of people are feeling when they mention your brand name, your competitor, or a topic that you are interested about.
Besides comparing yourself to your competitors by looking for brand mentions, you can also use social listening to look for general topics that may interest your organisation.
If you’re a pharmaceutical company, you could use an unassuming keyword like “symptoms” or “vaccine” to indicate if people are actively talking about a potential pandemic that is sprouting in a specific part of the world, or if there is a shortfall in medical supplies somewhere that your business can fill.
You would have noticed that I marked the phrase “that you define” in bold earlier. This is where social listening breaks down for organisations who don’t really put in effort into their keyword structure.
First things first, we need to agree that the client (that is you and me) are responsible for the keywords you want to listen to. The salesperson from the SAAS company are often more than happy to explain that they’re going to be holding your hand and guiding you to achieving remarkable things on the platform, but no. They’re likely to be more clueless than the newest intern in your organisation, and if given the deal, will likely get an after-sales representative to do a lazy job of putting up basic keywords based on best guesses and what they find on your website.
What happens next? You start collecting mentions (or buzz, depending on what the SAAS calls it), and it becomes a running number. You get to go in to see the specific content, and you’ll find weird posts where the keyword mentioned is irrelevant or misunderstood.
After a while, you start to doubt the data collected, and want to have nothing to do with it anymore. But since you have an annual subscription, you are still stuck with it. So you get them to churn out reports every day and every month and every quarter, sending it to management who will probably open it once, take a glance and delete them.
Come the next cycle, you might drop them and social listening didn’t work for you. Well, that’s because you used it wrong.
How to use social listening effectively
Find specific categories to search for
If you’re a pharmaceutical, the easy way would be to search for “your company + flu vaccine”, to see how well it is being taken up by consumers. But a more precise keyword might be “your company + your flu vaccine’s brand name”, along with a similar search for your competitors’ brand name and their respective flu vaccines’ brand names.
Sure, it is a lot more work to create keyword lists of different brand names for companies and vaccine products (plus imagine, you sell more than just flu vaccines, right?) But the effort is worth it – after all, when a consumer writes about medicine, they usually use terms like asprin or panadol, not paracetamol, right?
Rubbish in, rubbish out – There’re a lot of junk mentions when you just put in broad keywords. So for example, if I were to put in “flu vaccines”, I might get people talking about the availability of flu vaccines, or academic articles about flu vaccines, or people sharing about their experience in getting one recently.
In this case, having a list of negative keywords is necessary to keep your mentions clean. You could tell the social listening platform to exclude all mentions that have the keyword “doctor’s appointment” or “doctor’s office” or “research article”. This might help keep the mentions limited to the ones talking about overall availability in a pharmacy, and give you an idea if there is a shortage in supplies somewhere.
Now as you can imagine, this is an intensive effort to always keep cleaning the data and making sure that you are filtering out bad mentions when you can. It’s impossible to have a set of mentions that is 100% accurate – but that is not what social listening is about.
Instead, treat this as an iterative process, and never think that your keyword list is finalised. If your organisation sees value and is serious about expanding your scope for social listening, it always helps to have a community manager just looking at such data and fine-tuning it as a full time job.
There are a lot more nuances to share about setting up a robust social listening regime, and it all starts from proper planning and not letting the SAAS vendor do everything for you. Before typing in your first keyword, map them out in a chart on paper, and break broad categories down into the smallest meaningful ones – and only then should you start typing things down into individual keyword lists.
You’ll thank me later;-)