InSites Blog

Over the last two days, I attended the SPSS Directions Conference in Rome (IT). As a speaker, I was invited to bring a case study on using text analytics on social media data and eager to learn what others were doing on this domain. Moreover, since SPSS was recently acquired by IBM, I was curious to see what the future would bring. So I headed to Rome…

SPSS is typically associated with number crunching. One theme that however kept on coming back was the importance of qualitative and contextual data. I was truly amazed by the number of speakers on text analytics. With about 80-90% of company data that are unstructured, it is no surprise that this analysis technology is taking up. Especially in the area of customer satisfaction research, more and more people are using open questions in surveys to explain the “why” behind the numbers. The usage of text in predictive analytics is augmenting and so far, text seems to add quite a lot in the predictive power of models.

Many presentations also stressed the importance of the context in data collection. People are less rational than we often assume. In order to predict behavior, we have therefore to make more use of data collected at the heat of the moment. Particularly interesting was the speech of Ravi Dhar, Director of Yale center for customer insights. With some vivid examples, he stressed the importance of integrating context variables in your survey. For example, in a study with P&G (the measurement of the most effective communication for washing powder). The results revealed that on TV, it is better to the cleanness of your laundry whereas on the shopping floor ‘colors’ were a more decisive element.

A final session that drew my attention was organized by Mark Jeffries on soft skills. Many statisticians prefer to stick to the safe environment of their computer. However, we also need to sell projects and present the results of studies in an impactful way. Here are some tips & tricks to make this process more smooth:

• Whenever you get a business card, do not immediately put it away. It is the story of the person in front of you. Instead make a remark about it. Everybody likes to hear about oneself.
• If you want to get a ‘yes’, start already nodding with your head when you are asking the favor. There is a big chance that the other person will mirror your behavior and you will get what you want.
• In networking events, have a look at the feet of the person you are talking to. If his feet point too much in the other direction than where you are standing, he probably finds the conversation boring.
• The power of three: make sure that you always mention 3 reasons if you want to make a point. People have a tendency to believe you more.

It was the first time that I attended the SPSS conference and I must say that I was pleasantly surprised by its content. It seems that quantitative analysis is embracing qualitative data. At InSites Consulting, we believe in the fusion of methods and I can only consider it as a positive signal that companies like SPSS and IBM place this so high on the agenda.

However the industry needs to take even more steps. I wonder why there were no presentations where text analytics was applied on data from qualitative research. It seems that concepts like research communities, online discussion groups, blog research have not yet penetrated in this audience. Secondly, I appeared to be the only speaker who really showed how to embrace text analytics for nethnography. People were enthusiastic about the topic and I received the IBM- SPSS Insight Award 2010 for my submission. I hope this might inspire other researchers for next year’s conference! More information on the award will follow soon.

Scandinavian countries are famous for their innovative and progressive ideas in the organization of their society. Wondering if they would make this reputation also true for market research, I headed to Stockholm where I attended the Scandinavian Marknadsundersökningens Dag.

My favorite presentation of the day was without doubt from Teo Härén, author of several books on creativity. According to Härén, the key to more creativity is breaking your patterns and this was clearly demonstrated in his speech: he decided to do the Q&A session in the middle of his presentation and challenged the public several times to find alternative solutions for a problem. Some concrete examples you can start with today are taking another way when driving to work or change the side of the bed you are used to sleep on.

I also enjoyed the presentation of Jonas Ydén, head of market research at Carlsberg Sweden who gave his vision on the role of market research in an organization. Market research should not only be applied to make tactical decisions on communication, pricing or packaging, but should also monitor uncovered consumer needs and translate them in new business. Because insights can be present in different reports and places throughout the organization, they made an internal knowledge platform to make the market intelligence easily available throughout the whole company. Ydén finished with some guidelines for market research agencies:

• Do not just present the results of the study. Share also other customer knowledge you have with your client
• Do not say that something is an insight if it is not. (An insight at Carlsberg is defined as a single discovery of something enlightening about consumers, shoppers or customers’ underlying needs and motivations that Carlsberg can address to create consumer value and competitive advantage)
• Tell a story during your presentations and please the eye!

Next to the local speakers, some international speakers were placed on the agenda (like myself speaking on social media netnography). Although it is hard to bring multivariate statistics in a entertaining way, I believe that Ray Poynter did a good job in explaining CBC and MaxDiff in a simple way. I was also triggered by the presentation of Anna Thomas who used a technique called creative conversations to investigate sexual health among teenagers in the UK.

The Swedish market research day was definitely interesting. Although I did not discover any radical new ideas, I would definitely call them progressive in terms of market research. In a country that has about the same number of inhabitants as Belgium but a much larger spread, they managed to bring over 100 market researchers together for a full day of presentations. The openness and vividness where the participants were discussing research between sessions clearly showed the involvement of the whole market research community in taking research forward.

Last year, ESOMAR decided to replace its online research conference with a conference solely dedicated to online research. The program included a nice mix of traditional research into quality and reliability of online panels, along with the adoption of social media in market research. Moreover, I also presented a paper on a new methodology called social media nethnography together with Emilie Van den Bergh from RTL Nederland. Hungry for new insights and curious to see what people thought of our contribution, I headed for Chicago…

• C from communities

Research communities were probably the most discussed topic at the conference. Ray Poynter did a great job at illustrating that the industry still has a long way to go. By looking at the respondent experience and evaluation of research participants using this new type of research tool, he demonstrated that online research communities are generally not as enjoyable and fulfilling as focus groups and not as convenient as online surveys. The most used application of online communities was co-creating. Two approaches were presented: on the one hand, Volker Bilgram brought a case study about Swarovski where they challenged research participants to come up with a new design for a watch with diamonds. In a second round, consumers could vote on the different concepts. Thewinning idea was produced as a prototype. A second approach was presented by Darren Lewis and Koen van der Mal: they took a more traditional qualitative approach to co-creation where they stimulated idea generation by making use of projective and creative techniques.

• H from Hybrid researcher

A recurrent theme in many speeches and discussions was the need for education and training in the industry. Boundaries between qualitative and quantitative research are blurring. We need guidelines for analyzing visuals and multimedia in an efficient way. The gap between researchers who adopt social media in their toolbox and those who do not is becoming bigger. I was able to experience the need for more information about our industry myself in the Q&A after our presentation where I was quite overwhelmed by questions about text analytics and web scraping technology. Researchers increasingly need to expand their skill set and knowledge: the future is clearly in the direction of the hybrid research!

• I from Implicit measurement

Market research is starting to realize that not all behavior is rational. This growing awareness is translated in two trends: first of all researchers are looking into emotions. Secondly, biometrics provides us with new information about what we look at (eye tracking) and what engages us (arousal, heart rate, motion and breathing).

• C from codes

A substantial part of the conference was dedicated to ethical codes for online research. ESOMAR, ARF and ISO presented their ethical guidelines. A couple of important points were made: With the increasing adoption of social media tools in our profession, the danger exists that those tools will also be used as advertising channels to research participants. Therefore, a first important distinction between marketing and market research was put forward. Secondly, nowadays technology makes it possible to track consumers by installing cookies and other web tracking devices without the knowledge of research participants. The informed consent principle is in this case in danger. Finally, there are also guidelines under development for panel experience in order to maintain the quality of our data.

• A from Authenticity

Authenticity is a real buzz word and this was not different at this conference. A first context where it popped up was in sampling: there is an increasing interest in sampling through social media. People however often take a different identity online. As a result, consumers who do not fit the profile can enter one’s sample or the quality of the data can be altered to the different identity. According to the people from Facebook and LinkedIn, the way to go is social networks. Because the aim of social networks is to connect with friends, there is no point in having a fake ID. A second context where authenticity comes in to place is in social media netnography. In order to place studies that are based on user-generated content into context, we must know more about the creator of the post. A first attempt for the blogosphere was given by Josephine Hansom. Based on qualitative research, she identified three types of bloggers: people who use blogs to work through their own ideas, people who blog for a large public and people who blog about a niche topic only. In the future, we should not only expand this profiling but also try to identify detectors of the different profiles by looking at the vocabulary they use in posts.

• G from on the GO research

On the last day, a full session was dedicated to mobile research. Mobile phones are nothing other than a new channel to communicate between research participant and researcher. The simplest form is simply text sms: the researcher sends a question and the participant texts the answer back. A second form is flash surveys. The challenge here is to adapt the look,feel and type of questioning to mobile phone devices. The more advanced version of flash surveys are widgets. One last application that seems particularly interesting is interactive voice response: participants are invited to give the answer orally. Since this usage is not bounded to the size of the mobile phone screen and with the evolution in speech capturing techniques, I believe this will be the future!

• O from observation

As I said at the beginning of my post, we presented a paper on social media nethnography. With the aid of a case study on X-Factor, we demonstrated how we can make use of user-generated-content to answer a certain research question (in this case, what do people think of the television program X-Factor?) It is of course difficult to judge your own work objectively but looking at the vast amount of positive reactions and questions about how exactly to conduct social media netnography, I can at least say that we were able to put this new research paradigm on the research map. Although many research companies are incorporating social media tools into their research designs, only a few companies are really doing something with the content. I am therefore especially happy that we were able to trigger other researchers to help fine-tune the methodology of social media netnography. As I said in one of the last slides of my presentation, we have many challenges ahead:

- How to communicate optimally with our clients in a research paradigm where there is often no real research question to answer.
- Further thinking about the sampling procedure, ethical guidelines, data collection and framework development should be encouraged
- How will we deal with the new set of research skills where the boundaries between quantitative and qualitative information are disappearing?
- What about representativity when you have thousands posts but no profile information?

Chicago is often called the windy city and looking back at the conference, it is clear that a fresh breeze has entered the online research landscape! Chicago is also the hometown of Obama and therefore interconnected with the concept of change. The Esomar online research conference 2009 demonstrated that market research will need to change to take research even more forward! Looking at the presentation and discussion of the last couple of days, I am confident to say: yes, we can!

Part two of my take-aways from The Vlerick Informeel Lecture (16 June 2009)

Why people prefer Coca-Cola over Pepsi?

Unconscious thoughts have clearly a big impact on our actions. Neuromarketeers try to study this impact in different ways. On the one hand they make use of psycho physiological measures: they study eye movements and the size of the pupil in the eyes and measure if a certain stimuli evokes a reaction in the skin or voice. The problem with those psycho physiological measures is that they do not reveal any information about emotions.

A second type of measures come from from neuroscience. With the aid of fMRI scans, neuromarketeers look at the activity of the brain structures when seeing marketing stimuli. Depending on the region of the brain that is activated, they can tell what the person unconsciously thinks. Neuromarketing is used for solving all types of research questions: to detect what your most popular brand is, how attractive you find packagings, what you think of the price, why you buy a certain product.

The most common example of neuromarketing is the Coke-Pepsi challenge. In a blind taste test, most of the people prefer the taste of Pepsi above the taste Coca-Cola. Whenever the research participants know what they are drinking, they prefer Coca-Cola. Neuromarketeers have investigated this strange phenomenon and they came to a remarkable conclusion: Pepsi activates indeed the ‘taste’ centre in our brains but Coca-Cola shows a strong activation in our emotional brain. The marketeers from Coca-Cola have clearly done a good job in branding!

Why a cow drinks milk?

Neuromarketing does not always involve the usage of complex machinery. One way in which they study also the unconscious brain is through reaction time tests. By measuring the speeds of reacting to a certain stimuli or putting participants under time pressure, the association with a certain brand, product, advertisement are measured. A small test can make it more clear. Please answer as quickly as possible on the following questions:

• What is the color of snow?
• What is the color of paper?
• What color is a doctor’s lab coat?
• What is the opposite of black?
• What does a cow drink?

The majority of people will have answered ‘milk’ on the last question, not immediately realizing that a cow actually drinks water. How can we explain this common mistake? As you have noticed, the answer to all the previous questions was the word ‘white’. Since we have repeated this word several times, it has unconsciously been activated in our heads. It is however also the case that the words that are associated with ‘white’ are also a little bit activated. In this context, ‘milk’ is thus a word with a greater level of activation. When we finally ask the question about ‘a drink’ and ‘a cow’, two words that are also associated with ‘milk’, the threshold of this word is exceeded and people incorrectly answer ‘milk’.

The neural networks in our brain are clearly a powerful tool for marketers who can use common associations with their brand in their communication. The associative networks in our brains influence our behavior more than we think. Still not convinced? Have a look at this movie that illustrated how brain associations even influence marketeers.

In a previous life, I studied cognitive psychology. I have always been intriged by the functioning of the brain and how it effects human behaviour. I experimented with eyetracking, implicit measurement through reaction times and EEG. During my internship at the neuropsychology department of the UZ Gent Hospital I even conducted an fMRI study in order to learn more about emotions. Given my background, I was therefore immediately enthusiastic when I saw the invitation for the Vlerick Informeel Lecture in cooperation with Baqmar about Neuromarketing.

How sure are you that you have total control over what you are doing? Even if you think you always know what you are doing, neuroscientists have discovered the opposite: there are many thought processes going on at the unconscious level. Not convinced? Check out the following movie where you see people playing with a ball. Count how many passes the white team makes. Ready?

http://www.youtube.com/watch?v=Ahg6qcgoay4

People who carefully watched the movie, will have counted 14 passes. However, many people will probably not have seen that a black bear was also visible in the movie (if you do not believe me, watch the movie again).

Neuroscientists have discovered that the human brain actually consists of three layers: the reptile brain, the emotional brain and the smart brain.

• The reptile brain, the oldest of the three layers, controls the body’s vital functions such as heart rate, breathing, body temperature and balance

• The emotional (or limbic) brain is the container of all our pleasant and unpleasant experiences. Based on our past experiences, it helps us to judge the current situation.

• The smart brain (or neocortex) is the rational brain that is typical for humans. It controls language, memory, abstract thought, imagination, reasoning, decisions and many more ‘intelligent’ functions.

From the three layers in the brain, humans are only ‘aware’ of processes in the neocortex. However, the unconscious brains also have an important influence on our behaviour. In our head there are two routes that determine our decisions: a controlled route in our smart rational brain and a emotional route coming from the emotional brain. The two routes often conflict with each other. Consider the following situation: a boy is secretly in love with one of his friends. In normal circumstances, the neocortex brain controls his behaviour: since he is not sure that his love will be returned, he suppresses his feelings: the smart brains dominates the emotional brain and the controlled route is taken. However, if the boy has however drunk too much alcohol, he might declare his love anyway. Under the influence of alcohol, the smart brain can no longer inhibit the emotional brain which will direct the action. The latin saying ‘in vino veritas’ that says that the wine contains the truth is thus supported by neuroscience!

Over the last year, we have explored text analytics within the ForwaR&D lab from InSites Consulting. Text analytics can be defined as a set of linguistics and/or statistical techniques to extract concepts and patterns out of textual documents. In essence, it can be applied on all types of text. Within InSites Consulting, we have applied it so far on social media, open questions and research communities.

Except for a couple of Esomar papers or academic publications, I have not encountered text analytics too much on market research gatherings. I was therefore eager to share experiences with other users: How do they use text analytics to get new insights? What software do they use? How do they approach the research process? How do they sample? What visualizations are new? What about sentiment analysis? Therefore, I decided to subscribe for the 5th annual text analytics Summit.

So did I found what I was looking for? Here is a summary of my main take-home messages:

• Text analytics have many faces

80% of data are unstructured. It is therefore not surprising that text analytics appeals to many companies. Text mining is applicable on fraud detection and e-discovery (finding evidence for criminal cases). The technique is also often used in development of search engines. A nice example of this is wolfram alpha. In this search engine you do not need to enter keywords but you can directly ask complete questions. The engine at the back end analyses the sentence with aid of text analytics and tries to give an answer. Another application is online advertising: based on the key words you enter in search engines or your profile on for example social networks, you get customized online advertising. Other purposes of text analytics can be found in knowledge management and Business intelligence.

• The voice of the customer

Attempts to integrate text analytics in market research are mainly focused on two areas: Alike InSites Consulting, research agencies try to capture the buzz on social media in order to take spontaneous consumer feedback into account. Capturing the online buzz is however only the tip of the iceberg: customer services have call centers where they collect complaints, questions, etc. from their clients. Many companies hope that text analytics can help them to automate and get additional insights out of those texts.

• Part of speech: looking for the Holy Grail?

Text analytics is much more than only counting the words in text. It also needs to take linguistics into account: the software needs to understand stemming (e.g. spoke is a declension of speak), causality (e.g. tom hits Jan is different from Jan hits Tom), co-references (e.g. the boys are playing sports, they are very tired: they refers to the boys) and intensifiers (e.g. I am so dirty implies an amplification of dirty). Even when this is countered, we need to try to make the computer so intelligent that it can grasp sarcasm and irony.

Another challenge is sentiment analysis where we derive the attitude of customers about a certain brand or product indirectly through the emotions people mention in their texts. First of all, the extracted emotions are often limited to positive and negative. Secondly, from the moment there are several brands mentioned in a text, the computer needs to derive which emotion is related to which brand. Different algorithms try to provide an answer for all those challenges. A hybrid approach from different algorithms delivers also the best results.

It seems that human involvement will remain important to do the final check in case of doubt.

• You are being watched

Text analytics is only the beginning. The next wave of innovation in this area will be in other types of data. Speech mining is already on one’s way: telephones to customer services are recorded and with the aid of speech technology transmitted to texts. The next big thing is video analytics. First cases are done in retail where the traffic of customers through the shop is analyzed based on the image of security camera.

Back in the headquarters in Ghent and looking back at the conference, I think back on what I learnt over the last couple of days: whereas intelligent computer used to be something from science fiction movies, this summit has convinced me that it is becoming a reality. Although the applications of the new analysis technique are wide spread, my ‘research heart’ beats quicker when I think about the infinite possibilities it can bring to market research. We are however still at the early days where many questions remain to be solved: How do you take a good sample of social media sources? What is the profile of people posting online comments? How can we present these huge amounts of information in a managerial way? What are appropriate sample sizes to draw valid conclusions? Within The InSites ForwaR&D lab, we are trying to pin the methodology down. When I left to Boston, I hoped to discuss the methodology more in-depth with other users in the field. And although I saw interesting things in this field, the conference was a bit like a text analytics client project: you often find answers for questions you did not really ask. It was for sure an inspiring conference! I am sure the market research industry will address best practices in text analytics in the future. If somebody would like to already change ideas now, I would be happy to join the discussion.

For more information you can contact Annelies Verhaeghe, R&D Consultant (Annelies.Verhaeghe@insites.eu)

Thursday 5 March 2009. I am on the Eurostar to Brussels after a two day conference on online research organized by the WARC in London and I make up my mind: ‘what did I found striking?’ Although, I had expected a somewhat larger and international public and I had never (before yesterday) seen anybody faint on stage, it occurred to me that this conference was centered on three big themes: the listening economy, social media research and fusion research

Listening economy

In line with our connected research philosophy at InSites Consulting, a lot of the speakers focused on the fact that we need to start listening to our customers instead of always interviewing them. As Adam Phillips explained during his short speech, this thought is not new at all. In the 40-50^th, P&G conducted mass ethnography. At that time,  5 ways of listening were identified:

·         The best way to get real insights is to eavesdrop natural conversations in a pub, the bus etc.

·         You can launch the topic you are interested in outside a research context and see what people are saying about the topic.

·         You can mimic natural conversations as authentically as possible in a research situation: you admit that you are a researcher and you memorize your topic guide/questionnaire and respondent answers.

·         If you go one step further, you also take notes during the conversation.

·         The final type of listening is the structured questionnaire or interview.

The researchers at P&G declared that the more you evolve to the last type of listening, the more one will obtain socially desirable and superficial answers. Therefore the third type of listening was implemented. Because of the memorization, this way of conducting research turned out to be too cost and time intensive: they had no other choice than to stop this type of research at that time.

Over the last decades, we all forgot about actually listening and bombed the consumers with questionnaires. With the rise of social media, we are now back to listening and the question really is if researchers will be able to do better this time?

Social media research

Although market researchers are gradually becoming aware of the importance of listening, the big question is how brands can use new social media in the best way possible. The rise of blogs, social networks, video and picture sharing websites, etc. has first of all provided us with new research tools. On the one side, qualitative research is brought online mainly by setting up bulletin boards or asynchronous discussion group. The usage of online discussion groups and blogs for qualitative research was hardly discussed. On the other hand, panel providers reshape their panel databases to research communities and recruit participants for research through social networks.

Listening to consumers can also be managed by tapping into natural user-generated content. Although, some dashboards were shown for capturing the online buzz, it was clear that a lot of pioneer work still needs to be done in this area: what are the best sources? How should I collect my data, how to analyze such a huge amount of data? What is the profile of people who post information online? What is the profile of the reader?

Fusion research

The importance of fusion was embedded in a lot of speeches. It seems that this fusion needs to happen on three levels in the industry:

·         Fusion of data: as stated before, we do not only need to learn from what consumers tell as in interviews and surveys but also to what they tell us spontaneously. Ideally, we also add other types of data to get a 360° degree: behavioral data (e.g. based on loyalty cards), data collected through neuroscience and cognitive psychology, insights from econometrics (which heuristics do consumers use in their decision process).

·         Fusion of insights: Researchers are no longer responsible for just delivering a good report for one study; we should try to integrate the insights coming from different reports. In line with the slogan of ‘no size fits all’, we carefully need to select the right methodology for each research question and study research questions from different points of views, an idea that is often called triangulation. Every methodology has also its bias and we should also take those different biases into account when trying to see the whole picture.

·         Fusion of skills. The internal organization of a market research agency or department also needs to be adapted to the changing environment. Researchers need to have more skills than just good analysis and reporting techniques. They also need to be good journalists so they are able to communicate about the research on the new social media.  The line between quantitative and qualitative research is nowadays blurred: by using social media for qualitative research, the amount of data that needs to be analyzed has increased. In order to get a grip on this large amount of data, qualitative researchers can get inspiration with their colleagues of the quantitative department who have experience with processing large amounts of data. 

9 March 2009:  Back at the Ghent office. I am going through my notes again from the conference and completing my blog post before it goes online. It is however hard to summarize all information that was shared during the conference in 1 blog post but I have an idea. Over the past couple of days, I have been working on analyzing social media via textmining (some wrongly call it blog tracking while it should be nethnography) on all types of user-generated content. So, I decide to do a little experiment to see what the common elements in all presentations at this conference were: I text mined all the documents from the conference and the following word cloud was the result:

I agree that textmining DID a good job at summarizing the conference. Again, what did I find so striking in this conference? It is clear that social media like communities and social networks are dominantly present, but surveys and internet penetration (very much 1.0 if you ask me) were still a significant theme. No attendee at the conference would probably disagree with the fact that we need to start listening to our customers, but in the majority of the cases, presenters were still speaking about respondents instead of participants (?). Clearly, many are talking about it, but who is putting the deed to the word? With the forwaRD lab of InSites Consulting, we are conducting online safaris with our participants; we follow their life through blogs; we study how we can moderate synchronous and asynchronous discussions via chat and bulletin boards in the most optimal way and are analyzing user-generated-content with textmining tools.  I hope therefore that for next year’s conference we can start discussing our experiences and best practices with new types of research so we can really take online research forward.