This is a little bit like a print firehose because the content was so rich and my fingers are so slow. The overall impression was, as in past Research Events, enormous value delivered by top professionals to a fascinated audience. Whew!
Not surprisingly, the Facebook presentation was the best attended. In fact, people were sitting on the carpeting and filling all the walls. Fantastic presentations all around… Here are the first three!
Dawn Lacallade, Senior Strategist from ComBlu opened the sessions with a great overview of leveraging SM for “Research and Insights.” Stepping in at short notice to replace a speaker who could not attend, Dawn did a fantastic and dynamic job of presenting the case for communities, how to do them right and make them pay for research.
Community demand often comes from “We gotta have one too!” with little idea what it is supposed to accomplish or exactly how it should be implemented.
Reason enough: online presence and demontrated expertise and fluency are the new measures of competence for individuals (try getting a job in research managment without an effective online presence) and for companies as well.
Case study from Dawn’s many years at Dell… Idea Storm tended to surface the easiest ideas to understand. To glean the most useful and innovative ideas Dell brought in “builders” who interacted intensely with customers whose ideas seemed to show special promise. Communities are not expert systems. Getting the most out of them demands system experts and a much more engaged approach.
Case study: Dell’s battery recall – makings of a gigantic PR nightmare. Negatives from Radian6 were almost off the charts. Videos of laptops with fire shooting out of them on a conference table.
Dell came out with a huge recall, but behind the scenes the recall leaked out before the website for registering was ready. Chat about the recall started on the Dell community on Sunday when their website not ready for entries until Tuesday. Picking up the community chatter, Dell was able to put on extra resources to have the recall and all customer service initiatives ready for first thing on Monday. Exploiting their community, Dell got ahead of the tsunami. Dell converted the crisis to positive coverage about their responsiveness.
Communities are intensive and require commitment from the company, the employees, the moderators, and the customers (who respond when they know they are heard).
Scoring community performance: brand value from engagement with customers, business value from tracking standard established metrics, community health from level of engagement from customers (can be assessed by reactions to events and reactions of community to issues affecting other members.
Linda Ashbrook, Taco Bell – Leveraging Social Media to enhance traditional research …
Taco Bell figure they know their customers already after 40 years of intense consumer research… But using Facebook has opened a new window into their hearts and minds.
Using Facebook, TB went to where their customers were (on Facebook) rather than trying to pull them into their venues for focus groups. They use research on SM for anything they want to introduce, for competitive assessments, to engage with customers locally on FB, and to invite them into their offices for tours and deep exchanges.
By her admission (and no surprise) they started out knowing just about nothing about communities, Facebook, and using it for research. They invented as they went. They are by their own assessment about 2/3 of the way toward the level of expertise they expect of themselves.
All current applications are qualitative. They have internal moderators who talk to individuals as well as the group.
On one project, they sent the panel out to try the $2 meal deal at TB, then go spend $2 at another fast food place: what do we offer by comparison to competition? What’s the real value (or presumably shortfall). Engagement provides excitement about the experiment and very rich exchanges and insights around a rather basic product and service test.
Copy development …. show early stage commercials and refine according to feedback in very short time frames.
Panels live for approximately 3 months, each with a focus (such as mothers about kids meals). When the mission is over, thanks and let them go back to their lives.
Taco Bell is amazed by the frankness of the exchanges and details shared. The ethnographic value surprised and is paying immense dividends.
A project among mothers about take-home meals using FB morphed from simple concept reactions to rich ethnography on how the American family feeds itself. Discovered that the Cleaver household does not exist any longer, but the desire is still there for the connection dinner can provide.
Price-Value Pipeline Lab on $1 value meals (case study). Recruited young guys with big appetites and small wallets. Heavy fast food users. Price sensitive…. Incentives are about what they pay for a focus group but get daily participation over the course of 3 months. Used for up front product development…
Combined online and direct in-person involvement for product tasting and refinement. Local panels are invited into the TB HQ for taste testing and in-person exchanges. Engagement is enhanced as applications for the panel are broadened.
Taco Bell also used Twitter to evaluate national launches…. By launch time they have spent immense resources to understand the product and probable customer reactions. To evaluate a disappointment used Twitter to capture weeks of tweets on reactions to the Black Jack Taco as opposed to the Volcano Taco… Basically, they did not understand the product concept and resisted.
Over the past year most initiatives have been successful.
Started with a field agency recruiting on Facebook. Now recruiting directly on FB, working with a SM expert. Costs are heading toward 20% of the costs of using a field recruiting agency. The SM expert continues to work with TB helping them understand how to use FB effectively.
Screening is fantastic: finding a 40 year old mom in Tupelo? Can do easy.
They pay incentives of $100 – $300 for 3 months on a panel…. Panels are no more than 20 people… Have two in-house moderators … Different dynamic moderating on FB an easier to bring on line…. not as intense or difficult to learn as full real-time groups.
Twice a week send out teasers for upper management with a link to explore. Helps to get buy-in from less-savvy managers.
Moderators maintain some confidentiality and do not allow friending or personal page interaction.
Demand internally is off the charts. Everyone who is a potential internal client wants a FB community panel dedicated to their issues.
Local panels who come into the HQ for physical food tests…
The have a Taco Bell Board of Experts of past panel participants.
The ethnographics have been deeper and richer than anything they previously have encountered.
Still evolving their criteria for traditional focus groups vs. panel projectes. Also looking for adding some kind of quantitative assessments. Could the FB panels replace their other panels in part or even in total? Returns are not in, but the possibility is certainly there.
Analysis can be a bear…. you must keep up with it every day because the information is voluminous. Check feedback with development to see if consumers are experiencing something known by management or perhaps something they have not anticipated or need to begin monitoring.
Recruiting on FB can be done through ads, but using a SM expert who makes large buys gets better placement and much better responses.
Looking for more efficient ways to manage the panels… perhaps moving toward dedicated MROCs? Currently they send out checks and feel there must be a better way. Also interested in software support for analytics.
Starting only one year ago, Taco Bell has found revolutionary value through taking qualitative onto Facebook. They are interested in further experimentation while continuing to reap the surprising and immense benefits.
Sean Bruich, Facebook…
Are SMs changing the rules of MR?
Like the other presenters this morning, Sean is a great presenter whose intelligence and commitment to the topic are electric. Extremely impressive intellectual grounding for every observation delivered in rapid fire style. If you can get recordings of all of these presentations, I strongly recommend them.
Three rules that describe real world research constraints.
First rule: really hard to find the right participants for your study.
Second: Even if you find the perfect person, odds of their wanting to participate are low… response rates.
Third: a successful project is no guarantee that the success can be repeated and results can be compared (paraphrase that may not do full justice to the point).
Facebook can help with all three:
Scale: bigger is better. 65% of 500M users come back every day. Global reach. SM is not just a place where people are but a window into real life beyond merely FB participation.
Engagement: job is to make the world more open to share and connected. FB has constructed a system. Constraints on a system can improve end product.
Research polls unit: one or two question survey appearing on users home page. People like answering the questions. They don’t have to go anywhere or fill in complex fields, often can see their responses against all other responders.
Response rates are 100s of times higher than to off-site surveys.
Happier respondents give you better data. Higher response rates are more representative. You are getting responses from people who would not ordinarily participate in a survey: painless and easy and fun. Shifting the burden from respondent to researcher has a big payoff.
FB started tracking Obama’s approval rating since March. Correlation between FB and Gallup or Rasmussen is higher than correlation between the two to each other.
Suggests that FB is becoming representative not just of the FB universe but of the world. Further alignment between FB and Gallup. Also have been able to predict the Consumer Confidence Index accurately over the past year. Responses come back from “millions in minutes.”
When May CCI came out it was up even though there was a lot of bad news. Sampling had been taken in first two weeks, when bad news happened in second two weeks. Days matter in research, and small, fast, and huge has a major positive impact.
Extending across multiple markets…. just because you have a great question, finding a comparable sample in another market is a bear. FB users interact with the site in similar ways. While answers they give may differ, they can be reached in the same way and can scale very easily. Example: tracking product launch over many global markets. Translations are the only additional issue.
FB already has information on its users and does not have to ask standard and boring demographics. FB can also present the demographics with confident reliability because the users are incentivized to be honest about themselves when relating to their friends as well as your questions. Data can be cut in many interesting ways.
On election day FB could click the “I voted” link to encourage people to vote. They could see how many people clicked and could also parse the clickers by political affiliation and other demographics. Showed a 3% advantage in favor of Republican participation (approximately what results showed). Also looked at the friends of those who voted and clicked. Knowing that a friend voted highly correlates with wither you voted.
Neilsen Brand Lift collaboration with FB….
FB has a strong record of who saw ads. Users tend to be really the users rather than just someone else using a browser. In traditional ad tracking dropping cookies, the most active site users tend to get the cookies and more active people are in the universe which can skew the results.
On FB there is a control group with ads that are disabled vs people who saw ads. They get feedback from both and measure the differences between the groups who differ only in exposure to the ad (purchase interest, for instance). Normally you want feedback on ad copy within a 48 hour window of exposure. Here you get feedback almost instantly.
In answer to a question from the floor: They do not do or allow any tracking of FB users visits to external sites.
FB research takes advantage of the fact that people are interested in engaging, that they get feedback, that it is trouble-free. Policy is that a couple of questions do not a survey make… three or four go over the edge.
I asked Sean to answer his own question. Are SMs changing the rules of marketing research? He said, “Maybe.” Maybe rules are not really changing but SM makes it possible to observe them and come away with better results.
But “better results” doesn’t really seem adequate to describe a universe of 500,000,000 respondents, many of whom are going to answer your questions. Getting millions of responses in minutes with the ability to do simultaneous segmentations? Maybe a few rules are getting bent?
I feel very lucky to have been able to blog on these sessions. Can’t wait for the next ones and ready to walk around and sample others.