By Michalis Michael
A few years back I was regularly using this quote from former Global Insights Head of P&G Joan Lewis:
“Survey research will decline dramatically in importance by 2020. Social media listening will replace much of it, adding new dimensions”
According to a 2016 report by Markets and Markets, the global social media analytics market was US$ 1.6B in 2015. It’s expected to grow to US$ 5.4B by 2020. We have a couple of years to go, and in my view, the total market size for social listening and analytics in 2020 will be over US$ 7B. Whatever the case, Joan’s prediction was off by a few years but not wrong in my opinion.
Open-minded customer insights practitioners have been seeking ways to integrate social analytics into their insight generation process, however easy or difficult it may have been for them at the start. I personally believe that unsolicited consumer feedback and behaviour tracking will eventually replace tracking surveys entirely.
Reasons Social Listening Hasn’t Taken Off in Market Research Yet
Trust
One of the main challenges insights managers face is accuracy. These professionals have probably tried some traditional social media monitoring tools in the past. When they realised the sentiment precision* of the analysis they were receiving was nowhere near appropriate for insights (in most cases <60%), their trust was broken. Accuracy = trust in this case. Resultantly, they’re now reluctant to use anything branded as ‘social listening’. Even if it comes from legitimate market research specialists who can conduct social listening & analytics with precision high enough (>80%) to be suitable for consumer insights.
*accuracy in big data is measured with the following distinct metrics: Precision, Recall and F Score
Choice
Over 1,000 social media monitoring tools exist. How do we expect buyers to make an educated choice and believe in the potential of the tool or solution if we refuse to accept this methodology that makes our lives easier? We must be careful. We should probably come together and set some boundaries to protect the image and value that our industry offers, from aspiring tech companies that don’t care about the accuracy of their data.
Consistency
Numerous buyers and use cases for Social Listening within an organisation complicate matters further. Large organisations who tend to be the main buyers of MR want to have a “one stop shop” for various departments: PR, Customer Service, Product Development, Digital Marketing, and Customer Insights. The problem with this is that insights professionals are forced to use the tool their marketing department is already using, rather than searching for an accurate, more appropriate alternative.
Knowledge
Some die-hard purists stuck in their old ways still ask the question “how representative is the sample”. They’re unable to understand that we are not dealing with a sample and we are not dealing with participants!
There’s also a misconception that only negative people post online, thus making social analytics biased and unrepresentative. This is far from the truth:
- a) for most product categories most of posts are neutral in sentiment
- b) there are significant variations in sentiment for different categories
- c) even if only negative (or positive) opinions existed online, the rest of the consumers – not posting, just reading – are still impacted by the opinion of the few. Consumer sentiment expressed on social media impacts sales!
Education
As in many other cases where innovation was introduced into an industry, we need to figure out how to accelerate the education of clients to use it effectively. The only way to do this and convince sceptics is by demonstrating that we can discover relevant and unique insights that enhance or replace surveys and traditional focus groups. We need to prove that social listening data can be integrated with survey and behavioural data to produce an outcome that is by far superior than the sum of its parts.
Even though we are heading in the right direction, our adoption pace is too slow. All signs show that a slow acceleration has started with big multinational clients leading the way. The only hope for Joan’s prediction to come out right is to experience exponential growth of social listening in MR during the next 2 years. Such a possibility is still in the cards. I now have a new favourite quote I use in conferences and client meetings. It was told to me personally by Tom Emmers, former Senior Global Director CMI of Heineken:
“I believe that today, given the rapid advancement of capabilities, we should only resort to surveys if we cannot find the answer in social listening or behavioural tracking data”.
By Michalis Michael, CEO, Digital MR
2 comments
“One of the main challenges insights managers face is accuracy… Accuracy = trust in this case. ”
That’s an odd view, when you consider the alternative in a panel survey is a pre-designed and prescriptive questionnaire delivered to a group of particularly weird people who *want* to complete surveys. “Accuracy” is very much a relative term, and there’s little to say that social listening is less ‘accurate’ than survey research for what organisations need. You’re right that sentiment analysis is poor, but many clients aren’t bothered, or work around that.
In fact, your question contains the answer. Social listening has taken off in leaps and bounds and is the must-have tool for communications and media agencies. The chance to hear unvarnished views from the public is key to any campaign planning process (regardless of the quality of discourse). It hasn’t taken off *in MR* because they are seen as different disciplines; a view only reinforced by the lack of interest among most vendors in how social listening is used, or in supporting its use in their own work.
The question is not “why isn’t it used”, but perhaps “why is the research world so resistant to considering it a relevant source?”.
One problem of social listening is the low number of social media contributors. For most of the MR client companies / brands, the number of active contributors (posting comments, blogs, tweets, etc.) is lower than 5% of the respective target group. (Example. If a company has 2M clients+prospects and 100.000 of them are engaged in social mentions, the social media contributors are representing 5% of the target group.) And these 5% are NOT “representative” for the target group: Usually they are particularly engaged customers or people with professional intentions like employees, competitors, consultants, journalists and the like.
Further, it is not true that ” the rest of the consumers – not posting, just reading – are still impacted by the opinion of the few”. Of course users are reading product reviews in webshops and app stores. But how strong is the impact of such readings on their final purchase decisions? And how do reviews, recommendations, social posts etc. affect all those products and services which are sold outside webshops (in the offline world)? Not all people are googling products before the buy it. And not all consumers follow blindly the recommendations they get.
(I know – there are many surveys telling us that the “majority of consumers are relying on reviews and recommendations”. But such findings gathered by surveys/polls do not measure the real behaviour of consumers but just their opinions.)
Finally, I agree that social listening is a highly interesting methodology, but there are still a lot of unanswered questions about the validity and the relevance of social mentions.
We need further evidence.