1. Clear Seas Research
  2. SIS International Research
  3. Research Live
  4. RN_In_App_Effectiveness_GBook_480_60

Counterpoint: Data From Social Media Can Turbo Charge Insights Professionals

Last week Ray Poynter wrote a well-constructed post wrestling with the question “Why Has Social Media Analytics Met With Limited Success In Market Research?”. Having spent 11 years analyzing social data for some of the biggest Fortune 500 clients, I wanted to point out (in the friendliest manner) why some of his assumptions are wrong.

turbo

 

By David Rabjohns

Last week Ray Poynter wrote a well-constructed post wrestling with the question Why Has Social Media Analytics Met With Limited Success In Market Research?. I respect Ray a lot; he is one of the smartest and most prolific bloggers in the world of new market research. I also cannot disagree with his central premise that Social Media Analytics [has] met with limited success in market research. However, having spent 11 years analyzing social data for some of the biggest Fortune 500 clients, I wanted to point out (in the friendliest manner) why some of his assumptions are wrong.

In the article Ray claims, “Its weakness is that it can’t answer most of the questions that market researchers’ clients ask”. We have a slightly different perspective, and here it is laid out in a point-by-point response.

• Most people do not comment in social media.

Most people also do not attend focus groups or answer surveys. The question should be: does a representative and relevant sample of people comment? Given that the people commenting tend to be the most passionate in the category (often heavy users) and those asking questions tend to be trying to make buying decisions, you could argue that it is a very relevant sample for many business needs (especially given the flaws and biases apparent in participation in traditional asking research).

• Most of the comments in social media are not about our clients’ brands and services

Exactly! That is the best thing about listening research. It lets you listen to what people actually want to talk about and share. Most comments are about people’s lives. In the food and beverage category for example, only 5% of conversations mention a brand. The opportunity is to understand the other 95% which, in our experience, is where the most powerful communication and innovation insights are found. It’s here that you can best learn what consumers really care about, identify their unmet needs, and attach your brand to those passions.

• The comments do not typically cover the whole range of experiences (they tend to focus on the good and the bad). This leaves great holes in the information gathered.

Actually, we find that the good and bad account for around 35-40% of conversations. People also spend lots of time talking about consumption circumstances, the purchase process, and the research that went into their decisions to buy or pass on a product. By virtue of their broadcast nature, Facebook comments and Twitter activity do skew to pithy statements focusing on the good and the bad. However, adding the unconstrained and anonymous discussions people have with each other on forums, message boards, and blogs paints a much richer picture of the product experience.

• It is very hard to attribute the comments to specific groups, for example to countries, regions, to users versus non-users – not to mention little things like age and gender. You are correct that it is often hard to overlay traditional demographics on the data (it can be done to varying degrees, but it is neither easy nor precise). However there are many other, more practical, ways to segment the data, typically in terms of psychographic characteristics such as motivation. For example, moms cooking dinner for the kids love cheese (it gets them to eat the broccoli), while the same mom trying to lose weight has a very different attitude toward cheese. The beauty of digital/social is that you can understand the same mom in different mindsets. In fact, the motivational and other psychographic segmentation that social media supports may even be more useful than demographics because it exposes a universe of micro segmentation opportunities that demographics do not.

• The dynamic nature of social media means that it is very hard to compare two campaigns or activities, for example this year versus last year.

It is true that recruitment and measurement looks very different when using social media to research consumers, but it is still possible to compare response to campaigns. There is a science to measurement in traditional research which cannot be applied to social media data. However, if measurement is a science, learning is an art that is well-supported by online consumer conversations. I have seen many campaign comparisons – for example the launch of Halo in 2007 vs the launch in 2012 – in which it was not only apparent how the new campaign was doing vs the benchmark of the old, but easy to learn from consumer-to-consumer interactions. Consumer conversation analysis may not be the only method brands use to measure campaign results, but it can be a valuable one.

• The number of people using social media is changing, how they are using it is changing, and the phenomenal growth in the use of social media by marketers, PR, sales, etc. is changing the balance of conversations. Without consistency, the accuracy of social media measurements is limited.

OK something we agree on 100%. The quality of the data used for input and how it is organized and cleaned is crucial. Social data sources are as different as mall intercepts (Twitter) and friendship focus groups (online communities such as forums and message boards). The industry needs to analyze the right data in the right way for the job at-hand and ensure that the data is clean and unsullied by marketing messages (just as you always needed good quality respondents to yield good research in the old world).

• Most automated sentiment analysis is considered by insight clients and market researchers to either be poor or useless.

Agreed again. Most (not all) market research companies are not using the right tools or algorithms today, and many of our clients say they have abandoned the sentiment analysis function in their in-house dashboards because it is so inaccurate. Northwestern University did a study in 2006 that found that sentiment bears little relation to sales and share. However there are metrics that do.

• This means good social media usage requires people, which tends to make it more expensive and slower, often prohibitively expensive and often too slow.

All great insights need people. I think the question is: how can technology accelerate the speed to insights, not how can it replace the people. The key is to find tools that can turbo charge great people, not replace them.

• Social media deals with the world as it is, brands can’t use it to test ads, to test new products and services, or almost any future plan.

Social media reflects the interests of the world at scale. If you want to understand what people are sharing with each other in the social age or what new opportunities for innovation and communication exist, listening research is a great complement to asking research. Companies like Procter & Gamble, Samsung and AT&T are thriving because they have worked out how to use it for competitive advantage.

Having spent 15 years living in the world of traditional market research (as a client and an agency planner) and another 11 years in the world of listening analytics, I can say that both have pros and cons. However, the biggest redeeming feature of social data is its ability to reveal important questions that we never thought to ask and the ability to quickly test hypotheses in the data and refine or discard them. Online conversations leave a permanent trail that allows us to explore consumers’ worlds in constantly changing ways without having to take on the expense of re-fielding the study.

My sense is that the real reason that social media has not been more embraced by market research companies is the same reason that Netflix was not embraced by Blockbuster; it represents a dramatic change in technological and human capabilities that traditional research companies simply do not yet have.

For market research companies to cross the chasm to the “age of the customer” they will need to:

1. Invest more time really understanding and gaining access to different kinds of online data, and the ability to know how best to use them.

2. Stop relying on dashboards to provide deep insights (this is like doing all your research using SurveyMonkey).

3. Start thinking of “big social data” as “oil” that can expand marketers’ thinking, increase the speed to insights, uncover growth opportunities (like micro-segments and innovation), and improve the overall effectiveness of brand strategy.

4. Realize that the “age of the customer” is not going away. Tools do exist today to find powerful insights from consumers’ online conversations. However, this technology is too often confused with social activity dashboards, and data decisions are driven by cost, not value.

Some of the most innovative and thoughtful brands’ market researchers derive great value from social media. However, they work with specialists rather than traditional market research companies or in-house dashboards. Eventually, their sophistication will become the norm, but until it does, most attempts to use social media for deep consumer insights will produce disappointing results.

Share
You can leave a response, or trackback from your own site.

4 Responses to “Counterpoint: Data From Social Media Can Turbo Charge Insights Professionals”

  1. Jeff Reynolds says:

    May 2nd, 2014 at 11:36 am

    I too respect Ray a great deal. I also agree with a bunch of the critiques as to why Social Media analytics will not substitute for “traditional” market research. But I would urge all of us to look at what social media insights CAN do for the insights industry. I’d like to suggest three very specific benefits and applications. 1. Speed. Business and our clients are moving at lightening speeds. Clients are tired of 12 or even 6 week turnarounds of insights. We use social media to turn around insights in days. 2. History. Our tools allow us to pull history back 5 years on a question posed today. Current MR relies only on recall. 3. Trends. We use our tools to watch influence on low base rate trends. If there is a competitor or consumer behavior phenomena that is of VERY low incidence (<1%) MR can't deal with it. But social media can watch it closely, and when it grows, the red flag goes up and we figure out the pattern of influence.

    Lots for our industry to learn and figure out with this new source of data, so let's keep discussing it!

    Jeff Reynolds is the President of LRW, Lieberman Research Worldwide.

  2. Mary Aviles (@connect4mary) says:

    May 2nd, 2014 at 8:30 pm

    We are signing from the same song sheet! We posted along very similar lines earlier this week: http://baumanresearch.com/qualitative-mrx-one-best-current-applications-smr/ I can honestly say the unique contributions of our #SMR efforts make our overall results so much better. So great to hear that others are experiencing similar things.

  3. Carlo Erminero says:

    May 4th, 2014 at 10:28 am

    Again on SMR vs traditional research. Having carefully red the post of David Rabjohns challenging the position expressed by Ray Poynter, I would say that I’m fully in agreement with David. That’s not unusual. The surprise comes from the fact that I was in agreement with Ray either, after having red his post. Possible?

    Yes, it is. No contradiction. Ray is a brilliant son of a well respected tradition in research: the trdition of random sampling. For more then one century the theory of random sampling brought great successes to research practices and outputs. And it still does. Everybody in our profession clearly knows that a sample of 100, if representative, is more reliable than a sample of one million. Difficult to believe for a layman, but we know it is true. But three assumptions were behind: (a) a proper definition of the universe should be available; (b) all measurement were independent; (c) every single measurement, every single answer, had the same value, the same weight of another one.
    Now we have to face another world, the world of Social Media, where no one of the three assumptions holds any more. Therefore researchers has to prove that true knowledge could be reached under different conditions. Dificult task: a new paradigm has to be invented. But don’t be discouraged, we are halfway, already. Let’s start with assumption (c), and reverse it: not all respondents are equal. In a knowledge society, every unit (every respondent in the old world) has a weigt that is proportional to his competence and to his power of influence (Plato, Lenin, Pareto, Rogers, among many others, would agree). If you accept this idea, the problems posed by assumptions (a) and (b) easily disappear. But one thing is clear: you are doing research to predict the future, not to describe the present. A strong limitation, of course. But acceptable in many cases.

    Thank you both, Ray and David

    Carlo

  4. Contrapunto : Los datos de los medios sociales pueden potenciar los insights Profesionales | Futuro Labs says:

    May 8th, 2014 at 2:23 pm

    […] Lea la versión completa aquí […]

Leave a Reply

*

%d bloggers like this: