How to be effective on social media

We’re all forever tweaking and finessing our social media approaches. But with dwindling attention spans, huge competition for eyeballs and the never-ending changes to algorithms it’s really key that we understand what we should and shouldn’t do to remain effective. 

Since we met at Comms Unplugged last year, we’ve both been thinking about ­– and testing – how we can use a more data informed approach across strategic communications in the public sector.

We know there’s big interest in this area – a recent Webinar we ran with local government consultant Georgia Turner on taking a more data informed approach to Twitter had 150 sign-ups. But we also know that lack of time, understanding, confidence and skills can be a barrier to a wider use of data in what is (generally) the domain of wordsmiths, creatives and Excel-phobics!

A recent comms review and social media deep dive audit brief provided us the opportunity to apply powerful data mining and analysis to help a local council build a fuller and more accurate picture of how it was using social media to deliver on strategy and priorities. We asked the team for their hunches, assumptions and received wisdoms, and set out to find the data needed to test them.

The work we did informed key recommendations and tactical advice ahead of the launch of its major campaigns. 

In this blog, we share five things we’ve learned so far during this and some other data-driven projects.

1. ‘This could have been a photo’ – video isn’t a magic bullet

If you’d agree with the line that video gets higher engagement on social media, it might be time to think again. We’ve tested this theory within a couple of projects recently and it just doesn’t stack up for Twitter and Facebook.

Videos can be costly in terms of money and time, especially when budgets are constrained and teams stretched. We isolated video content, image content, and text only content, and when you look at average engagement rates and also representation in Top 10 and Top 20 highest performing posts, there is nothing to suggest video content gives you a benefit over (good) images. 

So is it time to stop making videos? Not quite. Videos can be a factor in high-performing posts but alone it’s not something that guarantees success, and there are other content elements that seem more important in driving engagement.

We used data analysis to give the council team we worked with recommendations on key factors in successful content – for example, content that prompts an emotion, celebrates ordinary people, or that offers a financial benefit. We’ve hopefully also given them a strong evidence base to push back against internal demands and baseless theories.

(A caveat, though; on Instagram, the data we’ve looked at recently strongly suggests that video performs better than photos).

2. Beware of basing your content approach on outliers

In statistics, an outlier is a data point that differs significantly from other observations. So it might be a social media post that gets especially high engagement – and because it does well, you might remember it, and conclude that success will follow from doing more of the same. And you overlook the other times when this type of content didn’t do so well. 

In a recent project, a member of the comms team offered a theory about a particular type of post ‘often getting higher engagement’ on Twitter. We set out to test this by capturing tweet content and analytics for a three to four month period and isolating just the relevant tweets. Sure enough, there was one post in this dataset that performed especially well, that had been remembered but all the others were all distinctly average and forgotten. 

There can be value in outliers ­– for example, looking at all the elements that make up that successful post – you may find some of those elements are common in other successful content. This gives you something to trial and test, to see if your hunch stacks up.

But you need to be aware of drawing sweeping conclusions based on that one result, and look at ‘success’ in the context of the bigger, data-informed picture. One swallow doesn’t make a summer, as they say.

3. Authentic over designed

Do you spend time designing images for social media? Is it worth the effort?

With a little bit of know-how, we were able pull thumbnail images into a dataset of LinkedIn content and analytics. This allowed us to see the type of posts that got high engagement and study the assets attached to them. We noted a trend for images of people, buildings and locations (and sometimes a mix of these) in the top performing posts. You could also use this to assess the variety and diversity in your images. 

Authenticity appeared to be particularly important for Instagram – as outlined above, data strongly suggests it is video content that performs best. But where images were used, natural and authentic images were seen to achieve higher engagement than those that had clearly been designed for promotional reasons. We believe this is, to some extent, the Instagram algorithm at work.

4. Echo chambers aren’t always as echoey as you think

Do you have employees or elected officials who always like and share content? Do you ever get the feeling you’re in a bit of an echo chamber, ‘preaching to the converted’ and reaching only people you already know? In the cut and thrust of a busy comms week, you may dip briefly into content engagement, and feel that it’s always the usual suspects who get involved with your content.

In a recent deep dive, we were asked to test the theory that LinkedIn posts received more likes from employees than third parties. The data did not support this – we found just 20% of engagement came from people employed or connected with that organisation, based on their job titles. Yes, many of these were regular engagers with most content, but they certainly weren’t the only ones.

The same went for Facebook and Twitter – we were asked to test the theory that ‘most of the engagement on Facebook and Twitter posts are from the same individuals’. Again, the data did not support this. For Twitter, we found that 50 per cent of all engagements came from just 50 accounts. But we also found dozens of one-off engagers who appeared to have no connection to the organisation. We concluded from the data that putting aside the ‘serial engagers’, it remained a valuable channel for getting messages out to the target audience.

Remember that social media algorithms will decide whether to show content more prominently based on how well it’s doing organically, and so a cohort of regular in-house ‘reactors’ and commenters is valuable and can help you reach wider audiences outside your immediate networks. 

5. Don’t keep posting the same thing ­– and don’t get obsessed with length

Are you asked to ‘give X another quick push’ by a team colleague or an internal stakeholder. Stretched for time, do you simply reheat and repeat what went out previously? It can’t hurt, can it?

From our recent work, we saw some strong evidence that repeated content resulted in ever decreasing returns on both Twitter and Facebook. In some cases, the repeated content went from good to zero engagement.

Our suspicion is that in the highly algorithmic world of social media, a platform determines this is ‘old content’ and of low value and therefore doesn’t show it to many or even any people. In short, it becomes a waste of time – and possibly harmful for your future engagement rates and a ‘drag’ on your digital reputation.

In a recent project we also tested the theory that ‘content with less text does better than content with lots.’ With data available on post lengths and engagement rates, we were able to test very directly for a link between the two numbers across Facebook, Twitter and LinkedIn content datasets. The hunch about shorter posts performing better simply had no basis.

So, fascinating stuff. We enjoyed undertaking these comms reviews and the data and insights delivered really will help the teams counter internal opinions and directives destined to not gain the best possible results.

And, just as importantly, these busy comms teams can now confidently focus on activity and approaches they know to work better, freeing up time to innovate, check out other opportunities and make bigger dents in those naggingly long ‘to do’ lists.

Alex Waddington is owner of Whetstone Communications. Darren Caveney is owner and creator of comms2point0 – where this post was first published – and specialist consultancy Creative Communicators Ltd. Read the original post.