Take the guesswork out of social media marketing

Learn how you can take a scientific approach to social media

For many brands, social media marketing is a mysterious, unknown quantity. They don’t know how it works or why. As a result many businesses just pump out tweets and posts because “you’ve got to have something on social”. There is a real science to how and why social media works, and you can make it work for you. Discover the science and take the guesswork out of using social media for marketing.

Under the bonnet of social media marketing

To understand how social media marketing works, we must understand how people work. On a biological level, human brains are geared towards developing and managing complex social structures. That’s how we’re able to live in such large, complex groups. Our neocortex, the part of the brain that scientists believe is responsible for social interaction, accounts for 80% of our brain’s weight. A much higher proportion than in most other animals. We’re hardwired to place importance on other people’s emotions in a way that most other animals simply aren’t.

Sinan Aral, an MIT professor and author of the book The Hype Machine, believes that social media exploits this instinct. Platforms like Facebook, Instagram, Twitter, TikTok and Snapchat are made to trigger the same dopamine hit for human interaction. As Aral puts it, social media is like “dropping a lit match into gasoline” for our highly-evolved social brains. This is why we respond excitedly to every like and comment.

Harnessing the hype machine

Social media tunes into the human instinct for interaction. So how can brands use this to improve their social media marketing strategies? In Aral’s opinion, there are two rules for brands and businesses to make the hype machine work for them.

Rule 1: Target networks, not individuals

There are commonly understood to be four phases in the evolution of consumer engagement:

  • 1980s: One message for the whole market
  • 1990s: Different messages for different market segments
  • 2000s: Messages tailored to individuals
  • 2010s: Delivering content to networks of individuals

People are bound together and influenced by their social groups. Statistically speaking, people tend to share their friends’ interests. But do people make friends with those who share their preferences, or do they pick up the preferences of their friends? In truth, it’s a little bit of both. This makes social networks a valuable target for brands, because it provides lots of opportunities to reach a receptive audience.

Firstly, brands can infer a lot about someone just by knowing what people in their network like. If someone’s friends like craft beer, folk music and expensive coffee, then there’s a good chance that they will too. This allows brands to target their marketing towards an audience that’s likely to be interested.

In addition to this, brands can make use of people’s tendency to adopt their friends’ preferences. For example, Sophie takes out a subscription to a craft beer service. The brand might then show a ‘Sophie just bought . . .’  notification to the rest of her social network. This implies that her peers should be interested, because Sophie is. The importance of network targeting goes well beyond selling, too. Take a look at this vaccine rollout video from the NHS. It doesn’t talk about facts or figures, the dangers of forgoing vaccination, or anything else; it shows people getting the vaccine.

On the surface, it’s an extremely simple piece of content. It’s a clever piece of marketing that plays on our inherent bias to what we see others doing. Because we see lots of people are getting the vaccine, our attitude towards vaccination becomes more positive. You may have noticed people who have been vaccinated sporting ‘I’ve had the vaccine’ stickers. This is part of the same idea, influencing the behaviour of people by showing how their social network is behaving.

Rule 2: Take a scientific approach to measuring ROI

The second concept underpinning a modern social media marketing strategy, according to Aral, is a scientific approach to measuring the impact you’re having. Thanks to highly detailed tracking and attribution metrics it’s possible to work out who buys what, how many adverts they’ve seen. And nearly anything else you’d care to know about your consumers. The trouble is sorting the wheat from the chaff; what data is significant, and what isn’t? As Wanamaker’s said, “Half the money I spend on advertising is wasted – the trouble is, I don’t know which half”. With some clever analysis, however, we can make some serious improvements in ROI tracking.

The key metric here is ‘lift’, which Aral defines as the effect your intervention has on the recipient’s behaviour. On his students’ first day, he hands out flyers to each of them as they enter the classroom. When asked what the conversion rate of the flyers is, the students respond with 100%. They’re right, because every single person who received a flyer subsequently attended the class. The problem is, precisely 0% of the students who received a flyer changed their behaviour; they were all coming to the class anyway. So, there’s 0% lift. There’s a difference between correlation and causation, and this is what social media marketers need to understand when measuring ROI.

Handing out flyers advertising a class to students entering the classroom may sound laughable. It’s not all that different to the broad social media marketing strategies that some brands use. If you show as many people as possible your content, you’re probably going to reach a lot of people who were going to buy your product anyway. Money spent reaching people who were going to buy anyway is just as wasted as on people who don’t go on to buy. We need a better social media marketing strategy than this.

Measuring lift in social marketing

Aral emphasises the importance of thorough split and A/B testing. Pit different versions of a marketing strategy against each other and iterate on the best-performing ones. He advocates focusing within each channel at first, before repeating this testing process across all the channels in use. Once your brand finds the best-performing content, it needs to find the best-performing channel, then focus ad spending on this combination.

It’s also important to have a clear idea of which metrics are the best indicators of success. Aral’s description of working with the New York Times on their paywall is a great example. After refining their paywall, the NYT saw the average number of readers drop but the number of subscriptions increase. The subscriptions bring in the revenue, so the NYT was happy with this result.

How to respond to the data

After performing in- and cross-channel testing, your brand should have a clear idea of what works. This data is a treasure trove of information, and you can use it in innovative ways to improve your ad ROI. A great example of this is Aral’s project with Jet.com (sold to Walmart in 2016 for $3.3bn). The online retailer offered shoppers discounts: the more they spent, the steeper the discounts would be. The products were discounted based on where they were in the warehouse. The logic was that products close together in the warehouse were easy to pick and pack. Therefore, it would be cheaper for Jet to send.

This runs counter to all the underlying psychology that we’ve seen so far. People don’t buy random items because they’re cheap, so it doesn’t make sense to offer discounts on, essentially, random items. Aral’s insight was to look at which products were complements and which were substitutes. Products frequently bought together, like pasta and sauce, are complements. If a user browsed several versions of the same product, such as coffee, before settling on one variant, then these products are substitutes. To maximise the appeal of products, Aral recommended discounting complementary products while not discounting substitutes. This makes sense, and is exactly what we’d expect in a physical shop. If someone buys teabags, you ask them if they need any milk or sugar to go with it. You don’t ask them if they’d like to buy a different brand of teabags.

The benefits of a data-driven approach to ROI

When using a scientific, data-driven approach to measuring ROI, brands benefit from a much-improved ability to assess the impact they’re having. Look at how different strategies perform within a channel, then compare across channels. Brands will gain the ability to measure lift in a meaningful way. This provides fertile ground for optimising marketing strategy because brands can focus their ad spend solely on the most effective methods. Proctor & Gamble, for example, slashed their advertising spend by £350 million in 2016, and their online performance improved – how? Because they were able to switch from a broad, frequency-based approach to a more targeted reach-oriented strategy. When brands understand what works for them, using the tools we’ve discussed, they can cut their ad spend while increasing performance.

How you can harness the hype machine

To understand how social media marketing works for you, you’ll need to focus on the key takeaways:

  • Social media marketing is about networks; how one person acts influences their friends. How does your brand use this to its advantage?
  • You can’t succeed if you can’t measure success – so focus on identifying and improving ROI by rigorous split testing, in-channel then cross-channel.

Sinon Aral’s book is a great read, and we recommend that anyone interested in learning about social media marketing checks it out. Discover what social media marketing can do for your brand, the best thing you can do is contact our team of experts today. We’re always happy to talk about how we can help you, so drop us a line anytime [email protected].