Sense and Online Sensibility

Jonathan Moody — Sat, 13 Oct 2007 04:18:30 GMT

PDF version available here. Abstract: User-generated content (UGC) and social media are constantly growing in number and influence on consumer decisions. Blogs, forums, usenets, audiovisual sharing media, social networks, podcasts microblogging sites and social news sites all offer the opportunity for people to share their experiences of products, services, brands and organizations. Consequently, these media are largely beyond the control of organizations and several large companies have experienced crises that have spread and grown through social media. This article argues that monitoring and understanding social media is essential for many organizations if they are to formulate a strategy for dealing with adverse and supportive comment. Some organizations are taking this one step further by recognizing that when correctly gathered, analyzed and presented, UGC represents a valuable source of insight into understanding and engaging with consumers. Author Information: Jon Moody Market Engineering jon.moody@asomo.net skype: jmoody-asomo ASOMO Service www.asomo.net socialware Garrick House 26-27 Southampton Street London WC2E 7RS United Kingdom tel. (+44) 207 717 8475 www.socialware.eu Paula buys a bottle of shampoo in the local store in a small Arkansas town where she regularly does her shopping. Later that night she washes her hair with it and wakes up the next day with orange hair. After going to the store where she bought the shampoo to complain to the perplexed store clerk, she takes a few photos of her orange hair and posts them on her blog. Like many personal blogs, Paula’s is little more than an online diary and is read regularly by about 4 people. One of these is her sister Anita who lives in New York City and goes onto a health and beauty forum where she posts the story (photos included) of the orange-haired girl. Within 24 hours, the story has spread widely and a manager of a well-known shampoo brand is on the the phone to a communications agency specialised in crisis management. This story is fictitious but internet lore abounds with such crises: explosive mixtures of candy and soft drinks, easy to pick bicycle locks, exploding laptop batteries. Google the search terms Nestlé + powdered milk, Coca Cola + Columbia or Nike + Thailand and among the first results that come up are websites that present key public relations issues for these companies. These websites create information currents of public sentiment that spread through the Internet to form part of the global rumor network that is consumer generated media (blogs, forums, message boards etc.) Hence some companies have woken up to this environment beyond their control and at least try to monitor online sentiment to pick up on potential threats and sometimes even opportunities. A select few have taken it one step further and recognise online sentiment as a source of insight and a channel for engagement with consumers and other stakeholders. The power of online sentiment Less well known than the aforementioned crises but nonetheless remarkable are the small brands which leverage online sentiment to build up an impressive business. Last year, a wine label called Stormhoek in South Africa used a blog to launch their brand in the UK with the help of Hugh Macleod. They sent a booklet called “Wine Blogging as Marketing Disruption” to 85 bloggers, accompanied with a bottle of their wine. These bloggers were alpha influencers key to other customers' awareness, consideration, preference, and purchase. They advocate, rank, sort, evaluate, and ultimately create marketplace adoption. They have a powerful multiplier effect, spreading the word quickly across a broad network when they find something they want others to know about. The bloggers received a free bottle wine and in turn, maybe two, three, or fifty of them blogged about it. They also told their friends about it in pubs, bars, dinner parties and the ripple effect created consumer awareness and favorability towards the brand that would have cost a considerable sum to achieve by traditional marketing channels. But what would this small wine company from South Africa have lost had it not worked? They would have written off the approximately $400 cost of implementing the campaign and tried the traditional approach. Welcome to the world of word of mouth (WOM) or word of mouse. David Godes of Harvard University and others provide a good summary of how myriad definitions of word of mouth have developed over time. Since the traditional view of WOM is too narrow to capture the breath of influences, Godes and his fellow authors prefer the term social interactions (SIs). “This term is intentionally general, capturing as it does, a wide and heterogeneous set of contexts. We define an SI as an action or actions that a) is taken by an individual not actively engaged in selling the product or service and that b) impacts others’ expected utility for that product or service.” To clarify the SI concept, Godes et al set out three essential facets of any social interaction: channel, content and impact. The channel is the medium through which the influence takes place. This might be face to-face discussion, for example. Important dimensions that differentiate various channels include credibility, number of recipients, and whether or not the channel is mediated by a third party. The content is information that is transmitted over this channel. Important dimensions of content include its objectivity and valence. Impact represents the ultimate effect of others’ actions. Growing in number and influence Blogs, forums and other online communities are a channel growing in number and influence and references within them the content. Any figures on the total number of blogs, forums, videos posted on YouTube, photos on Flickr or Facebook members will be out of date by the time you read this Log on to these sites, search on Technoratti or Blogpulse or run an enquiry on the myriad sites that answer these enquiries for up to date figures. Another question altogether is the impact of these new media channels compared with advertising, media coverage and corporate websites. A degree of cynicism is prudent – these are, after all, personal opinions. However, in a recent Ipsos/Mori poll ((http://www.mori.com/polls/2006/blogging.shtml), customer reviews were considerably more trusted than company statements, e-mails and TV advertisements and about the same as newspaper articles. Indeed, Godes et al point to a number of factors that have lead to some observers stating that consumers’ judgements of the usefulness of advertising have declined over time and is now deemed to be lower than than WOM. These relate to demand-side factors: products are becoming increasingly more complex and hence consumers’ information needs also become more complex. On the supply side, technological developments in e-mail, internet-based media and mobile communications have increased the number and forms of decentralized communication. Hence information can be transmitted more quickly to more people increasingly cheaply. However, as Godes and his colleagues point out, this is just potential: the decision on whether to share information still resides with the individual. Furthermore, less compelling information might travel more slowly and be “crowded out” by more exciting developments. Yet it has become much easier to aggregate interpersonal communication. This is particularly true of online discussion forums and strings on blog discussions. Notice the plural in the graph above: Reviews on a recognised review website. If three out of 20 reviews of a hotel are critical, we might still go ahead and book a room there. The more that negative proportion increases, the less inclined we will be to stay there. According to the Ipsos /Mori poll, 39 million Europeans decided not to buy a particular product after reading negative comments about it on the Internet. There is often a direct relationship between a wave of positive or negative online sentiment and sales figures. Studies of those relationships and even between online sentiment about listed companies and their respective share prices would make for an interesting line of research. So, with the myriad sources of information at consumers’ fingertips, why are an increasing number of people visiting specialist forums or leading blogs before buying a mobile phone, taking out a mortgage or booking a holiday? They want to hear what fellow consumers think about a product or service before making the purchase decision. Car buyers can visit the brand website or even read a car magazine review waxing lyrical about car X’s beautiful lines or superb handling. But what about those annoying problems that sometimes emerge a couple of months after taking delivery of the vehicle – doors that jam, uneven steering or scraping windshield wipers? Log onto to a forum on car X and find out about these and other issues from people who already own the automobile. More positively, you might be 85% decided that this is the vehicle for you and seek out positive references to finally convince yourself (and perhaps another decision maker). Fake and be damned So isn’t this a golden opportunity for managers to get online and blow their own trumpets?And why not denigrate the competition while you are about it too? After all, myriad forums are available to everyone or you can even set up your own blog. Many travelers have alleged that the influential Tripadvisor portal is riddled with glowing reviews of hotels written by the hotel owners or managers. How do companies offering social media insight services get round this? These reviews are sometimes glaringly obvious because they are usually short - lack of detail is often a sign of lie. However, some clients want them included in the analysis (often tagged as suspect) because they can still influence perception whether they are are genuine customer reviews or not. Furthermore, social media lore also includes stories of attempts to influence perceptions which have proved to be unethical, disastrous or both. Whole Earth's CEO John Mackey recent bashing of the financials of Wild Oats (under the moniker "harobed" -- his wife's name spelled backwards on Yahoo) to damage Wild Oats stock price before he purchased it. US retailer Wal Mart received a wave of criticism when it attempted to launch its own social network. The Hub, as it was called, was supposed to be a place where high school kids could socialise and make friends. However, visitors became increasingly suspicious when they read some of the comments that were very positive about the brand. Word spread that these and been written by marketing professionals and the site closed down after four months. McDonald’s was also caught using a flog (fake blog). The blog in question appeared to be a written by a security guard who was so fascinated with a marketing promotion, that he wrote about it in his blog. Another blog call The Consumerist exposed the flog as the work of a PR agency. This subject is examined in some detail in a paper entitled Strategic Manipulation of Internet Opinion Forums: Implications for Consumers and Firms by Chrysanthos Dellarocas (Sloan School of Management, Massachusetts Institute of Technology (2004). The most striking result of their investigations was that strategic manipulations can either increase or decrease the information value of online forums to consumers relative to a case where no manipulations takes place. The study discovered circumstances in which the presence of honest consumer opinions induces firms to reveal their own, more precise knowledge of product qualities by manipulating the forums at intensities relative to the actual qualities of the product in question. However, if there are enough honest opinions, then manipulation can prove to be harmful to companies because costs outweighs benefits. Lastly, other channels of customer insight are not exempt from untruths – people can and do lie in online surveys, hotel quality questionnaires, telephone interviews and focus groups. Many to many communication Moreover, online sentiment is just one of several information sources consulted in pre-purchase consumer research. Even the most ardent word-of-mouth evangelist will not deny the influence of other channels on consumer behaviour and purchasing patterns. However, consumer-generated media provides qualitively distinct insight. Media content, company communications, advertisements are all essentially one to many communications. Online sentiment emerges from many to many discussions with points of view expressed, debated or even refuted by a wide range of opinion holders. There are present customers talking to potential customers who are consulting former customers all of them tapping into opinion-leading blogs commenting on sector news flowing seamlessly between blogs, forums, offline conversations, e-mails, podcasts, social networks, news groups and traditional media in audiovisual, print and online formats. Organizations might not be able to control what is said about them, their products, brands and services online. They can, however, at least begin to monitor (for potential threats and opportunities) and even take it one step further and extract valuable customer insight from it. These are people expressing themselves spontaneously and of their own free will, which allows market research departments to discover new issues that were not included in questionnaires. A key value of this research is that is being freely expressed with no issue of question or researcher bias. Research bias can be very subtle such as the way a question is phrased if part of an online survey or asked if part of focus group research. For example, the question: What would you do if all the minibar drinks you drank were omitted from your hotel bill? This question would probably reveal many honest consumers in the presence of a researcher. Discussions in forums and blogs tend to be more candid as they are unprompted by a researcher. Specialist online sentiment monitoring software and analysts passively monitor the internet, never contributing or influencing the discussions. Getting started The start point for any research project is clearly establishing the objectives. Typically key issues to decide are time period, languages, key words or key concepts for searching social media for references This may be ongoing monitoring of brand and customer service perception for say, a hotel chain or monitoring awareness and favorability to a particular product before, during and after a particular campaign. Clients are interested in knowing which aspects of their products or services are discussed and how positively they are perceived. The monitoring company may have a fairly good idea about this from previous experience and can set out to the client in the brief the key concepts and their respective sub concepts that make up online sentiment. In the case of an airline, for example, key concepts might be the following: Tickets, Airport services, Airport staff, Boarding, Flight, On board services, Airplane staff, Baggage Each of those concepts can be broken down into subconcepts. Taking On-board services as an example: Food, Drinks, Service, Air conditioning, Blankets/pillows, Business class, Economy class, Coffee, In-flight entertainment, Lighting, Newspapers, Safety The sentiment monitoring company needs to make it clear that it can only gather, analyse and present what people are discussing online. In some cases, a potential client may propose monitoring sentiment about a product, service or concept which generates very little in the way of discussion. References to a product or a service vary a great deal in length and may contain a number of opinions on sub concepts ranging from one to a hundred or more This means that the analysts need to assign a positive/negative sentiment value at opinion level to truly understand which aspects are valued by consumers in social media. A reference can praise the way check-in staff handled overbooking by giving the passenger an upgrade to first class but then criticise the leg room and the dull food served on board. Language, not country A key issue for multinational clients to grasp is that online sentiment influence works mainly by language not by countries. Hence, opinion in English may be from a native speaker in Pittsburgh, London, Cape Town or Perth. Similarly, many people exchanging views online their second language (more often than not, English). French speaking Belgians participate in French forums, their Flemish-speaking counterparts in Dutch ones. Austrians will read German blogs, Germans will discuss issues in Swiss forums. The applications of online sentiment intelligence cut across several functions within an organization: public relations/communications, marketing, customer relationship management, product/service development, stakeholder relations. The insight can lead to modifying existing products, services, the way they are communicated and promoted, discovering new business opportunities and measuring the success of marketing and communications activities. One size does not fit all A particular project might be analysing sentiment about multiple product lines, in several languages with the intelligence to be used by multiple departments and across several territories.The board only needs topline statistics for the period. The PR manager’s insight needs differ from those of the product development department. Website managers want the feedback about user experiences on their websites. Marketers want immediate feedback on the product launch or to know which strengths to play to in their next camapign. Country managers might only be interested in sentiment in their respective languages. In some cases, feedback can get down to the business unit level – a particular store or hotel receiving topline statistics and verbatim references. At the other end of the scale, those in corporate social responsibility may want to discover how thier company is perceived by stakeholders as well as straight customers – labor organisations, lobby groups, pressure groups, the local community, employees. Text and / or graphics reporting is limited by the fact that different people need different insight. As a result, it is essential that the data is presented in such a way that each person can hone in on what is useful to them, by using interactive viewing software that allows them to apply filters to the data and to drill down to the original references. Everyone may not have or indeed need access to this viewer software, hence those that have the software need to be able to export the data and graphics to conventional formats to produce their own reports. This is also important to be able to integrate online sentiment insight with online surveys, instore questionnaires, focus group findings. Similarly, clients are likely to want to compare the study period in question (two weeks, month or quarter) with previous periods or see how an issue is developing cumulatively. They may even want to focus in on a particular week within a monthly report to see how consumers reacted to a particular campaign or piece of news. As a result, the ability to filter by date is also beneficial. Typical filters normally include language, business line (for example hotel chains in a hotel group), opinion holder profile (customer, former customer, competitor customer, potential customer, employee, opinion-leader-journalist). This range of people sharing experiences online is one of the key strengths of social media sentiment insight in that it allows companies to access opinions from groups of people otherwise difficult to engage with. Former customers may be too angry to want anything to do with the source of their grievances. Potential customers and competitor customers are similarly difficult to access. On the downside, blogging and participating in forums can be anonymous and so much of the information so vital to market researchers and their clients is often not available: sex, age, education, income group etc. Some of this information can be inferred by personal information on a blog (About Me) or from the text in a forum posting (“my teenage daughter thinks the dashboard is neat” – parent, probably in income group X that this car is aimed at, probably within age range 38 – 55). In reference to the multi-functional applications of online sentiment insight, a filter defining the opinion type is often useful. Hence the analyst will categorise opinions according to whether they are declarations (statements about particular concepts), suggestions (“a rear windshield wiper would greatly improve the limited visibility of this car”), questions (“What’s the best way of getting from the station to this hotel?) or calls to action (i.e. recommendations “Don’t stay at this hotel, dirty rooms, rude staff and appalling food” or “I would definitely recommend a weekend stay in this fascinating town). All of these opinions provide valuable insight but the latter three also provide an opportunity to engage with target audiences. Product development can get ideas for new features or even new products from suggestions. Travel forums are full of questions from people thinking about visiting city X or country Y. Tourist authorities are well-placed to answer these questions which are often answered by fellow travelers. Lastly, calls to action identify those people who are truly dissatisfied with the product or service and could form the basis for an outreach initiative in the customer service department to address issues directly with the person affected. The company may even consider participating in the debate so that other people reading it get a balanced view. Conversely, those recommending a product or service could be identified as key influencers to participate in a corporate blog or to collaborate as brand evangelists. Client needs vary from organization to organization. One automotive client using an online sentiment insight service said: “the (reports) help you detect and track the hot issues and to see how the groups of people expressing opinions on the different models develop over time”. Knowing the perceived strengths and weaknesses of each model allows the client to design strategies and redefine messages aimed at accentuating the positive and mitigating the negative. Another client from a well known soft drinks brand indicated that the company is “subject to numerous rumors and urban legends because of it’s brands reknown” Analysing online sentiment helps the client to discover if key messages are reaching their target audiences and to find out where people are talking about the company. Listen, act and engage The importance of online social media insight to an organization depends largely on the quantity of references to concepts such as its products, brands, services or people and the influence of these references on reputation and ultimately sales. For example, a group with several branded clothes store chains extracted significant insight about consumer perception of womens’ products, service and stores. The managers concerned then requested the same for some mens’ clothing chains only to discover that men discussed their clothes shopping experiences far less than women, at least online. In our experience, business to business product and services do not generate much CGM: companies do not go online to rant or rave about their suppliers. Airlines are sources of debate in CGM because we often have a choice of different airlines and seek to share our experiences of them. The airline we fly with usually determines the airports involved in our trip. Hence we see less utility in sharing our aiport experiences beyond factors involving the airline (check-in, customer service etc.) because, like us, our fellow travellers have often little choice in the airports they use. Sentiment in online social media is just one of several elements determining how an organization, product, brand or service is perceived. Organisations should not throw the baby out with the bathwater – robust media monitoring and analysis will still be essential to many organizations as will other forms of customer intelligence – the myriad forms of quantitative and qualitative market research. However, the importance of social media intelligence in the media mix is set to increase, particularly as social networking sites grow and non-textual references (photos, video) increase. There are concepts which do not attract a significant amount of online comment at present but could do so in the future as social media evolves. In September 2007, UK-based confectionary group Cadbury announced the re-launch of its Wispa chocolate bar, defunct since 2003. A camapaign run on social network Facebook eventually had more than 14,000 users demanding the re-introduction of a 1980’s iconic brand. The story about the explosive mixture of a well-known soft drink and a candy brand spread principally through YouTube videos. Some organisations are actively using social media for customer engagement. When JetBlue had to cancel or delay many New York flights over the Valentines weekend earlier this year due to bad weather, Chief Executive David Neeleman appeared on YouTube to reassure customers that the terrible service levels experienced would not be repeated. The company also made the effort to listen to as much customer feedback as possible. Consequently, customers posted positive messages on YouTube after Neeleman’s appearance pledging their support. As one customer commented on YouTube: “Honest communication is always appreciated even when things haven’t always been smooth.” Do you speak digital? Client needs vary from organization to organization With the media hype surrounding web 2.0 (and subsequent web 3.0, 4.0 etc.), many organizations are struggling to get to grips with an area many consider to be beyond control and consequently dangerous, irrelevant or both. Regardless of the type of organization, there is a permanent need to regularly review dedicating resources to listening to key audiences or stakeholders. Once those online conversations reach a certain level, regular monitoring becomes essential in to be less susceptible to suffering a web-spread reputation crisis. Once an organization is listening it can move on to acting upon the valuable intelligence that can be extracted from this dialogue and gain substantial competitive advantages. A good place to start is by reviewing your current agency/supplier roster in market research, direct marketing,advertising, public relations and customer management? What ideas do they have for communicating and engaging with consumers online? Your suppliers need to become very good at online or if you are already working with online agencies, do they have the range of services and the planning discipline that more established “traditional”agencies have? Another recommended port of call is Social Target's Guide to Social Media Analysis: a worldwide guide to the companies who monitor, measure and analyze online social media (www.socialtarget.com).

Consumer Generated Media from Journal of Online Research

Sense and Online Sensibility