Many brands spend money on social media without knowing whether offering a coupon or connecting with customers will be more effective. Research by Harikesh S. Nair, professor of marketing at Stanford Graduate School of Business, shows brands how to engage in a meaningful way.
YOUR BRAND NEEDS SOCIAL ENGAGEMENT
Social media is increasingly an important channel of importance to many brands. One reason is the massive reach provided by social media and the other reason is that users self-select into being connected to the brand on very large platforms like Facebook.
This is very helpful because this is a set of interested users that the brand can engage in very easily. The second advantage of that is by engaging with those users, you obtain free virality because those users can share your messages with their friends and their networks and that content gets engagement for free.
The third advantage is that those users that your connected users share your content with, are also likely to be more interested in you as a brand, because of what a sociologist referred to as homophily or birds of a feather flock together, because these users are connected to each other socially, they are more likely to like you as a brand.
WHAT DID YOUR STUDY AIM TO DO?
What we were trying to understand is what causes good engagement and what is the effect of that engagement in terms of brand power and value for the brand and what kind of content gets good engagement and in what way?
We were very lucky to be able to partner with a data analytics company which serves as a consulting company to approximately a thousand very large firms that do advertising and messaging on Facebook. We were able to track for a period of about six months to eight months every post.
WHAT DID YOU FIND?
We analyzed about 100,000 messages and then we built a natural language processing algorithm by which we classified various attributes of that content.
When we looked at what kind of content produced meaningful engagement in terms of shares, likes and commentary, we find that the inclusion of brand personality content in a message produces a massive amount of lift or improvement in engagement.
When we look at the inclusion of performance content for example, information about prices or deals or coupons, we find that it produces very little engagement. As you get more engagement, newsfeed or algorithms will give you a higher ranking and then you get exposure to more consumers. In that sense, reach begets more reach, more eyeballs will get you more eyeballs.
One of the things that we found in the data, is that a brand seems to be doing either one kind of messaging or the other, but often not both. For example, you will see one company doing a lot of performance advertising but very little brand personality advertising. We do find that performance advertising is pretty good for driving part to conversion, but because of the advantages of brand personality content, which generates engagement and contributes to long run brand capital, and at the same time begets more engagement through the role of algorithms, we think that it's actually very helpful to include some kind of brand personality content in your social media strategy.
So mixing things up is really good and a “portfolio approach” of mixing in content is something that is quite helpful.
Nike is a very nice example of a company that manages to engage with its audience, athletes and others in a very meaningful way, while at the same time making sure that you have access to discounts and other kinds of performance advertising that is required for immediate conversion.
BETTER DATA WILL BRING BETTER MARKETING
Marketing has now become a very data-driven field and as part of that evolution, we can now collect actual data on what consumers and human beings actually do, which is wisely superior to stated intentions data. And I think that is the holy grail of marketing because a lot of expenditures on marketing tends to be targeted at people who either don't respond to it or would have bought anyway, in which case you should think of it as a cost not a benefit.
Just because I bought something after I see the ad doesn't mean that the ad caused me to buy it. So, trying to understand attribution and causality, remains an open challenge.
So, I think that it’s a very exciting time where computational social science will be very important for the community, but at the same time there is a huge need for better education and better training of both scientists as well as managers to see how we can put statistics and data science correctly and in an efficient way, to improve the field.