Using social media for market research and new product development: the case of Hallmark:
- Evaluating the use of social media data as a research tool.
- Sharing what we learned and observed from a specific project: what worked and what didn’t.
- Highlighting paradigms of traditional research that are seriously challenged by the use of social media data.
3. Social Media Listening 2000 - launched first consumer community for the purpose of research. 2005 – pilot w/ Umbria to understand the Hallmark brand in the social space. 2007 - pilot with Spiral 16 to understand Influential Bloggers. 2009 - major RFP with 8 vendors – find partner for research and marketing. 2010 - consulted with Ben Smithee; contracted with Collective Intellect for a look into holidays, ornaments.
4. 1 Hard Question“Is there a framework we can use to understand social media data?” 3 “Simple” Questions “Can we use social media sources to get new product ideas?” “Can we use social media data to create marketing interventions ‘in season?’” “Could we use social media data to help understand people and their lives?”
6. Social Media Analysis Space? Discovery Understanding people Uncovering white space Topics + dashboard What you’re trying to accomplish Specific Topic Open Search Ways of approaching the data “Ocean” of Data Small “pond” of data Brand Reputation KPI’s Monitoring Source: Hallmark CU&I, 2010
7. Consumer Leads Focus We Lead Single Stakeholder Multiple Stakeholders Breadth of Learning Hallmark Social Media “Listening” Activities
8. Consumer Initiated Conversation Hallmark Initiated Conversation Single Stakeholder Multiple Stakeholders Breadth of Learning Listening audit showed that activities are many and varied Focus of Conversation
9. 9 Dimensions of listening Feedback is more about the brand and rational Discovery is more about the consumer and emotional Source: CEB
11. Research Landscape Structured Data – it is what we made it PA’s Multivariate Trackers …… Focus Groups Communities Ethnography Observation Qualitative Quantitative Social Media Unstructured Data – it is what it is Where does social media data play in the research landscape?
12. 3 “Simple” Questions “Can we use social media sources to get new product ideas?” “Can we use social media data to create marketing interventions ‘in season?’” “Can we use social media data to help understand people and their lives?”
13. What Did We Conclude? Getting “New” Product ideas is very hard! If you do lots of traditional research and ideation, chances are you’ve heard of thought of everything you will uncover through social media.
14. Some Thoughts on In-Season Marketing Twitter promotional sweepstakes generated major spikes in Hallmark Ornament Mentions Hallmark Gold Crown store visits emerged as a conversational theme - how could that be leveraged? Emotional connection to Hallmark brand ornaments vs. those bought in Wal-Mart or Target – how could that be leveraged? Focus on connecting the troops and their families – what would that look like?
15. What Did We Conclude? Beware of making any marketing suggestions without having Marketing in the design and analysis discussions! Research in social media begins to “rub up against” Marketing and can create friction!
17. What Have We Learned About Social Media? There are new vendors emerging daily in the social media listening space. Most vendors provide the same capabilities – it is hard to find differentiation.
18. What Have We Learned About Social Media? Social media data is messy (spam, advertising, porn, context) and requires significant validation and cleaning – suppliers are getting better and better at this, but it is still an issue. Text analysis capabilities vary from vendor to vendor – you have to know what questions to ask to know what you’re getting. Sentiment analysis is about 75-85% accurate (based on external research) and everybody does it at some level. It is of limited value. It is difficult to classify individual contributors – but not impossible, and it costs more. This capability is still emerging.
19. What Have We Learned About Social Media? Clear and specific listening objectives are essential to getting meaningful data and information from listening efforts. Social media conversations can be very rich and could support many different aspects of our business. Getting meaningful learnings and insights is very time and labor intensive, despite the software tools. There is a learning curve!
22. They may or may not represent your target consumer.
23. They do not necessarily constitute a “probability sample” of the population.
24.
25. Social Media Data is… + Longitudinal and Instance-based – Tradtional MR is usually instance-based feedback/insights, social media research can represent “flow of life.” There is a constant stream, or flow, of social media data continually being created.
26. Social Media Data is… + Self-recording/Archival – Conversations, both public and private, within the social media environment are archived and available for others to consume at will. Conversations with people in traditional research methods are recorded, but are generally only available to permission-based viewers/listeners. This creates an interesting blend of liability and value to researchers and brands.
33. Food for Thought Resources: People, Time, & Money Spectrum of Desired Outcomes
34. Social Media Research White Paper http://www.slideshare.net/CuratingPixels/utilizing-social-media-to-understand-people http://bit.ly/UnderstandingPeople
In late 2000 we launched our first consumer community for the purpose of research. In 2005, we contracted with Umbria (acquired by JD Power) to understand the Hallmark brand in the social space. In 2007 we conducted a 6 month pilot with Spiral 16 to listen to social media with the goal of understanding multiple things including Valentine’s Day, Value, and Influential Bloggers. In 2009 we undertook a major RFP with 8 potential vendors in an effort to utilize this data source for research and marketing. During 2010, we consulted with Ben Smithee of Spych Research to try and get our arms around the rapidly changing social media landscape. We contracted with Collective Intellect for a deep look into holidays and ornaments specifically, as well as measuring our new brand campaign. This was our most ambitious pilot to date.
Is there a Framework that we can use to understand social media data?We found some, and We made some up
In addition to possessing several key attributes in regard to its utilization in market research, social media data has been the impetus for multiple paradigm shifts in the marketing, advertising, communications and market research industries:Social media data is, in general, public and accessible to anybody. This includes all divisions (Research, PR, Marketing, Customer Service, etc.) and other companies. Historically, there has been a clear understanding of who owns particular consumer data. Typically, sales data is owned by the retail department, focus group data is owned by research, email open rates are owned by the marketing group, etc. This is different for social media data because of the data transparency described above. No department or group really “owns” the social media data.As a result of the transparency of the data and lack of clear ownership, what is inbounds or out of bounds for a division is blurred. Even more important, roles that used to be reserved for each division have now blurred as well. Within social media any of the divisions can use the same data. When it comes to engagement, the consumer doesn’t care who they are talking to when they talk to somebody from the company. Everybody is speaking on the behalf of the brand. This creates new challenges for traditional research personnel.The researcher used to control as many aspects of the listening environment as possible. For instance, recruiting based on a criteria, intentional specific questions, specific answer options, etc. Social media data comes uncontrolled, unedited, and unsolicited.The once separate worlds of qualitative and quantitative research are further blurred by the influx of social media data. The potential and possibility for both qualitative and quantitative researchers to utilize social media data outputs is evident and a valuable benefit.Individuals, who create compelling or “sticky” content, have the ability to compete for attention as well as any other media sources can. Who gets heard is no longer a function of size and money spent. Social media data may not reflect the targeted consumer. In addition, for given topic or source, a disproportionate amount of content may be generated from a subset of contributors (see Forrester Social Technographic Ladder). This flies in the face of the traditional researcher’s goal to find a representative sample of the target market. Because of the disproportionate contributions of some authors, the notion of influence becomes an added dimension for researchers to consider. This idea is expanded upon in the Peer Influence Model in Empowered by Josh Bernoff. Because of this, social media data will be more beneficial to exploration and learning applications instead of projecting. As response rates continue to decline for researcher efforts the ways of getting participation from people are changing. The Market Research Executive Board suggests the way to gain participation is now how you can add value to the consumer. This is more than just reimbursing them for their time. It is starting to venture in to the realm of building a relationship with them and adding value to their lives.
In addition to possessing several key attributes in regard to its utilization in market research, social media data has been the impetus for multiple paradigm shifts in the marketing, advertising, communications and market research industries:Social media data is, in general, public and accessible to anybody. This includes all divisions (Research, PR, Marketing, Customer Service, etc.) and other companies. Historically, there has been a clear understanding of who owns particular consumer data. Typically, sales data is owned by the retail department, focus group data is owned by research, email open rates are owned by the marketing group, etc. This is different for social media data because of the data transparency described above. No department or group really “owns” the social media data.As a result of the transparency of the data and lack of clear ownership, what is inbounds or out of bounds for a division is blurred. Even more important, roles that used to be reserved for each division have now blurred as well. Within social media any of the divisions can use the same data. When it comes to engagement, the consumer doesn’t care who they are talking to when they talk to somebody from the company. Everybody is speaking on the behalf of the brand. This creates new challenges for traditional research personnel.The researcher used to control as many aspects of the listening environment as possible. For instance, recruiting based on a criteria, intentional specific questions, specific answer options, etc. Social media data comes uncontrolled, unedited, and unsolicited.The once separate worlds of qualitative and quantitative research are further blurred by the influx of social media data. The potential and possibility for both qualitative and quantitative researchers to utilize social media data outputs is evident and a valuable benefit.Individuals, who create compelling or “sticky” content, have the ability to compete for attention as well as any other media sources can. Who gets heard is no longer a function of size and money spent. Social media data may not reflect the targeted consumer. In addition, for given topic or source, a disproportionate amount of content may be generated from a subset of contributors (see Forrester Social Technographic Ladder). This flies in the face of the traditional researcher’s goal to find a representative sample of the target market. Because of the disproportionate contributions of some authors, the notion of influence becomes an added dimension for researchers to consider. This idea is expanded upon in the Peer Influence Model in Empowered by Josh Bernoff. Because of this, social media data will be more beneficial to exploration and learning applications instead of projecting. As response rates continue to decline for researcher efforts the ways of getting participation from people are changing. The Market Research Executive Board suggests the way to gain participation is now how you can add value to the consumer. This is more than just reimbursing them for their time. It is starting to venture in to the realm of building a relationship with them and adding value to their lives.
In addition to possessing several key attributes in regard to its utilization in market research, social media data has been the impetus for multiple paradigm shifts in the marketing, advertising, communications and market research industries:Social media data is, in general, public and accessible to anybody. This includes all divisions (Research, PR, Marketing, Customer Service, etc.) and other companies. Historically, there has been a clear understanding of who owns particular consumer data. Typically, sales data is owned by the retail department, focus group data is owned by research, email open rates are owned by the marketing group, etc. This is different for social media data because of the data transparency described above. No department or group really “owns” the social media data.As a result of the transparency of the data and lack of clear ownership, what is inbounds or out of bounds for a division is blurred. Even more important, roles that used to be reserved for each division have now blurred as well. Within social media any of the divisions can use the same data. When it comes to engagement, the consumer doesn’t care who they are talking to when they talk to somebody from the company. Everybody is speaking on the behalf of the brand. This creates new challenges for traditional research personnel.The researcher used to control as many aspects of the listening environment as possible. For instance, recruiting based on a criteria, intentional specific questions, specific answer options, etc. Social media data comes uncontrolled, unedited, and unsolicited.The once separate worlds of qualitative and quantitative research are further blurred by the influx of social media data. The potential and possibility for both qualitative and quantitative researchers to utilize social media data outputs is evident and a valuable benefit.Individuals, who create compelling or “sticky” content, have the ability to compete for attention as well as any other media sources can. Who gets heard is no longer a function of size and money spent. Social media data may not reflect the targeted consumer. In addition, for given topic or source, a disproportionate amount of content may be generated from a subset of contributors (see Forrester Social Technographic Ladder). This flies in the face of the traditional researcher’s goal to find a representative sample of the target market. Because of the disproportionate contributions of some authors, the notion of influence becomes an added dimension for researchers to consider. This idea is expanded upon in the Peer Influence Model in Empowered by Josh Bernoff. Because of this, social media data will be more beneficial to exploration and learning applications instead of projecting. As response rates continue to decline for researcher efforts the ways of getting participation from people are changing. The Market Research Executive Board suggests the way to gain participation is now how you can add value to the consumer. This is more than just reimbursing them for their time. It is starting to venture in to the realm of building a relationship with them and adding value to their lives.