Opinions And Ramblings By Adam Kmiec On All Things

When Data Leads To Bad Insights – A Social Customer Service Analysis

I hate linkbait. Scratching your head? Linkbait is when someone crafts a headline that designed to get clicks, views, shares, etc…but the content is often misleading. Well, earlier this week, there was some serious linkbait going around with respect to a recent “study” completed by Socialbakers, that looked at the social media customer service habits of companies.

The stat, often being quoted from the study was, “58% of social media users want a response. But only 22% actually receive one.”  That stat is sometimes presented as “70% of fans are being ignored by companies.” Provocative, right? Did you lean forward? Raise an eyebrow? I’m sure lots of people did.

Even the “study” looks legit. I mean check this out:

Socialbakers Customer Service Study

and then this…

Brand Number of Fans Question Response Rate Number of Questions Answered
Personal Argentina 309,968 89.13% 14,306
Claro 415,026 96.99% 10,266
Acer Indonesia 612,837 95.05% 4,992
LBC Express Inc. 58,176 91.75% 4,083
Comunidad Movistar Argentina 385,451 91.67% 3,301
Volaris 335,410 77.92% 3,297
Safaricom Kenya Official Page 130,201 88.52% 3,255
KLM 1,540,788 94.14% 2,843
Vodafone Nederland 82,926 92.27% 2,758
Claro El Salvador 133,270 96.74% 2,758

Real companies. Accurate fan counts. So clearly…the data is right.

Add in the fact that Socialbakers is a “global social media and digital analytics company with customers in 75 countries representing every continent. Socialbakers helps companies measure the effectiveness of their social marketing campaigns across all major networks, Twitter, Facebook, YouTube, LinkedIn and Google+.” and wow this is looking good.

Are you concerned yet? With all this data…and all these insights, you should be right? Wrong. This is the problem that happens when organizations who have never actually brought social media to the enterprise at scale, start offering advice and “insights” about social media. Having actually been part of teams or leading teams that were focused on bringing social media to the enterprise at scale, let me drop some reality to balance the linkbait.

Most organizations have some type of response protocol for addressing content and conversations created in the social media space. My personal favorite is the flow that Ohio State created:

Social Media Response Flow Chart

It’s detailed and allows for a very quick assessment of what content is actionable. In Ohio State’s model they have categories called “Trolls” and “Ragers.” Posts/content flagged as such do not get responded to. I know this is going to just blow away the social media pseudo experts, but, not every comment or post should be responded to. Wait, look outside your window…did the Earth stop rotating? Yes, it’s true, your organization doesn’t have to respond to everything, nor should it.

Did you notice in the nice colorful chart from Socialbakers, the response rate for the Alcohol category is ONLY 5.2%. That looks really bad right? But, take a closer look. Use insights, not raw data. You know how whenever you visit a website from an alcohol company you have to provide your birthdate? They do that to make sure you’re 21 (in the USA) and/or of age. Do you know why they do that? They do it because they are legally obligated to not communicate/market to people under the legal drinking age. The penalties are extremely stiff. Unfortunately, even “age-gating” pages or profiles in an attempt to only allow drinking age adults to enter is not a perfect science. People can lie. People can skirt the system. The tools/options provided by social networks are generally weak. Then, add in the fact that there are significant number of people who choose to omit critical pieces of info, like their birth date, from their profiles and you can start to see how complex this problem is.

So…combine the flow chart from Ohio State with what I just shared about the legal ramifications of communicating with people under the age of 21 and is it any wonder that the average response rate for the Alcohol category is so low?

I could do this all day and poke holes in just about every single data point in this study. But, here’s what I think would be more productive. Check out the total mentions of “AT&T” on twitter here. Scan the first few screens and keep track of how many of the mentions deserve a response. Feel free to do the same for Bank of America on Facebook here. When you look at them or any other brand’s page, what’s the percentage of comments that deserve a response based on the Ohio State Social Media Communication Flow? 30%? 40% I can tell you this, it’s not 70% and it’s not 100%.

Look, we can all improve. We can all be better. We can all be more effective at turning frowns into smiles. But, let’s also be realistic. Let’s not be so Chicken Little. Let’s not fall for the linkbait and scare tactics. Let’s be pragmatic, not just as marketers, but as consumers. The spirit of what Socialbakers was getting out I can support all day long. The letter, the details and the specifics are just so misguided that I can’t possibly endorse the data or the insights.

Look, it would be great if we could scale social across the enterprise and response to every single person’s tweet, blog post, comment, etc. But, that’s not reality. Of course, most social media consultants, wouldn’t understand that…because they’ve never really been knee deep in a large enterprise, trying to scale social. It’s not easy. If it was…anyone could do it….then again, right now, it seems like everyone thinks they can.