Social Media Measurement, Social Media ROI

How to Uncover Meaningful Social Data


Call it the Analytics Effect: People get so hooked on cheap-and-abundant data from sources like Google Analytics that they ignore useful data from other sources.

Interesting as a concept, but does it really exist?

This question occurred to me after reading Jay Baer’s post revealing that social media sites have an an e-commerce conversion rate of .71 percent — a pale shadow of the conversion rates for email and search.

One commenter on Jay’s post mentioned surveys revealing that social media conversations play a significant role in people’s buying decisions, but social media interaction simply does not show up in conversion rates because a social media site is almost never the last click before somebody buys something.

That post got me thinking: If we have research demonstrating that social media is helping people make purchasing decisions, why are we still having these furious debates over the ROI on social media?

And that prompted me to speculate:

Is the abundance of cheap, easy-to-find data discouraging us from conducting the more complex, expensive — and more meaningful — research that would establish the true value of social media in the commerce equation?

Jay was so intrigued by this possibility that he asked me to explore it further in this post. Honestly, I do not know if the Analytics Effect objectively exists, but I have reason to suspect it might.

After Googling “measuring effectiveness of social media”, I read at least a dozen articles outlining all the sharing-and-sentiment indicators produced by posts, clicks, tweets, and so forth. These articles uniformly advised social media marketers to track these data points, but there seemed to be something missing — namely the fact that the metrics which work wonders in other channels are irrelevant to social.

With search, email, and pay-per-click, purchases establish consumer intent. On Facebook or Twitter, people might mention a brand in passing or even make a direct recommendation, but they almost never buy anything. It’s hard to establish show-me-the-money metrics if there are no transactions to paint a picture of consumer intent.

The surest way to figure out what motivates people to buy is to pointedly ask them how they feel. The best way to do that is with survey research, but it’s not easy or cheap: You need a reliable sample, perfectly crafted questions, agonizing decisions about how to conduct the survey, and substantial investments of time and money.

It’s so much more expedient to review the real-time click stats and A/B-test a few more headlines.

I’ve been been a true believer in social media user since my first discussion on AOL in 1990. I started my first blog in 1996 and joined the early waves of adopters for Facebook and Twitter. I’ve been at this so long that it’s hard to believe that we haven’t mastered the ROI of social media.

Seems like it ought to be settled law by now. Yes, we hear about the occasional survey by a big research firm pointing to how consumers use social media, but the mainstream of thought among social media marketers seems to be joined at the hip with data that doesn’t really click with social, from a conversion standpoint at least.

Social media marketers who want more buy-in from the executive suites might have to face up to the fact that social will never enjoy the tasty analytics served up on other channels. Social campaigns can build awareness, engagement, and consumer trust, and social media platforms can help people share trustworthy product recommendations. But there isn’t an easy way to stick a click between two guys discussing their golf club preferences.

To be clear, I have no illusions that analytics data is actually cheap or easy to capitalize on. It always requires smart people who command high pay. But it does seem like marketers have become so habituated to analytics platforms that the’ve become a bit like bears stealing picnic baskets: they go for what’s in front of them and ignore the more nutritious meals that require a lot more energy to acquire.

Facebook Comments


  1. Dave Link says

    LOVE the post. There is a wealth of data that can be attributed back through social media analytics platforms, but the real treasure trove comes from carefully combing conversations and the qualitative data gleaned from social. True, you can drive some incentivized conversions through mobile coupons or social flash sales, but those are flash-in-the-pan solutions.

    Social media’s real power has been – and for the foreseeable future seems to be – in primarily building awareness and branding while accepting that direct sales from social will be a secondary or even tertiary goal.

  2. ndawkins says

    Thanks for thinking a bit deeper on this topic (something that is sorely needed). I’d just like to add that survey research isn’t the only way to go deeper in terms of analysis. Case in point: We collected data in our product (Social Snap, now defunct as an SaaS system) on who was interacting with a client’s posts. We then compared that data with data in the CRM system to find out how valuable Facebook connections turned out to be. Turns out they were a LOT more valuable than customers who weren’t connected on FB. And that was only the tip of the iceberg — We got a ton of useful information as an outcome of that analysis. My point simply is that in the wild, wild west of social media, you never know what data points are going to be useful, so I wouldn’t recommend ignoring available data because its big, ugly and difficult to deal with. Survey research has its place but it isn’t a panacea (nothing is when it comes to social media measurement).

  3. says

    Fantastic post, Tom! I completely agree that many brands have struggled with defining the ROI of social because they don’t fully understand the role that it plays in their customer’s decision-making process. With very few exceptions, social isn’t a direct response channel, so attaching direct response metrics to an awareness or advocacy channel is doomed from the start. The concept of mis-aligning metrics to channel intent is not a new one, however. Just ask any direct marketer the value of display advertising…until they do an attribution model that shows how much their PPC campaign has been HELPED by layering in a concurrent display campaign:) Different story. Instead of looking for the easy answers, try to find the ones that matter. For instance, rather than direct sales from your Facebook page this months vs. last month, find out how many customers are also Facebook fans, and compare their lifetime value to non-fans. What? There’s not an easy answer?!! I know. Do the work. Trust your gut more than just the numbers you can see, and the world will look a lot different. Thanks for elevating the conversation, Tom! It’s one many digital marketers need to embrace.

  4. Rodger says

    This a great beginning. Surveys are important. So are focus groups. And let’s not forget interviewing as a research method. In one case study I have in my files, the 4-H did this kind of person-to-person research and uncovered some anecdotal information that radically changed its vision and how the organization positioned itself. To build on Tom’s work, and the work that came before that, I would recommend a broad overview of the kinds of social research methods organizations can use to mine richer data about their customers. For example, the best survey questions come from collecting information from interviews and focus groups. If the population in question is large enough, information that comes from interviews and focus groups could be tested against a larger sample to determine it value across a public.

  5. says

    Thanks for all the comments, folks. Survey research was just the first topic that came to mind for me, so it’s good see suggestions which dig deeper than that.

  6. Graciousstore says

    What is the problem with Google analytic? what information is lacking it it? Does the cost of a toll make it much better tool than the cheaper or even free tool?

    • Peter Odryna says

      Google analytics doesn’t do social. A necessary tool in the tool belt, but others are needed.

Leave a Reply

Your email address will not be published. Required fields are marked *