Facebook and Linkedin to Offer Market Research?
It seems like the natural second monetizing scheme many dot-coms think of. First Advertising then Market Research. That was the case for me back in 2000 at the beginning of the dotcom boom. I was hired by Snowball.com (now IGN.com) to help start a market research vertical Iquestics.com. It was to be a new profit center, and we had about 1 year to do it before the VC money ran out (that was my own financial analysis which was off only by about a month). Snowball.com had 14 million 13-30 year old unique visitors per month then. It seemed to make sense to add market research on as a second revenue source after advertising. While we were able to get a few clients like AE, Visa, Microsoft, it wasn’t enough, fast enough. However, marketing research takes time to implement correctly.
Last week someone from Linkedin asked to set up a call with me to discuss market research on Linkedin. They were interested in speaking to a few of the better connected consumer insights professionals in the Linkedin group “Next Gen Marketing Research“. It seems they may very soon be making a “market research” (polling) function available much like what Facebook has been offering now for nearly a year.
For those not aware of Facebook polls it allows users to pay per response to ask a single question. You are limited to 5 answers. There are no other question types. You can filter sample by interests, gender, age, or location. Even these few filters are limited and you can not apply multiple filters. For instance, it is technically not possible to select females and age 13-17 at the same time. Also choices like location are limited to individual states or smaller countries (Luxembourg, Sweden). You can not select USA for instance.
Therefore this polling tool is not useful for actual consumer insights projects I work on. Though I have found it useful to flush out some ideas for proposals, or prior to giving a talk/speech on youth research (GenX2Z).
I think it will be interesting to see if Facebook and Linkedin will be able to provide more robust consumer insights tools to us in the future. For now though Anderson Analytics and other research firms will need to continue to build their own panels which contain users of these systems. There are a few ways to do this including advertising, or working with application developers (on Facebook).
Has anyone else used the Facebook polling? Do you think Linkedin and Facebook will be able to provide useful market research products to us? If so, what would you first most like to see available?
-Tom
14 responses so far ↓
1 Ha Vo // May 4, 2008 at 5:29 pm
Tom,
I do think that it should be possible to create more structural tools for communities like LinkedIn and Facebook to offer market research possibilities. The amount of people and their profiles are interesting and probably reliablee enough.
In my opinion getting paid might indeed work like any other type of reward system that is interesting. For instance the http://www.e-rewards.com concept works together with corporates that have a member reward systems. I think that might also work, because these systems already appeal to clients as client understand that they can save slowly for something very nice.
Paying small amounts for contributions can deliver a person a few cents to a few bucks, not too appealing I think.
I have never used the polling tool before, but I’ll be sure to check it out.
Cheers, Ha
2 Steve // May 5, 2008 at 7:43 am
Hi Tom;
Yes, I believe linkedin has tremendous potential as an excellent sample solution. I see it play a B2B roll for market researchers. If I’m not mistaken linkedin has over 20 million professionals representing about 15o industries. I would be very interested to tap into this source for my clients. Please keep me informed on any new developments.
Thank you for your update!
Steve
3 JSV // May 5, 2008 at 8:52 am
Thanks so much for posting this Tom. Extremely interesting, would love to hear more.
JJ
4 Tom H. C. Anderson // May 5, 2008 at 9:02 am
Thanks. I offered to be part of beta testing. Will let you know if I find out more.
Tom
5 Chris // May 5, 2008 at 9:15 am
One of the challenges from the purchasing side is how representative the sample is going to be. You would need to show some solid validation studies across a couple of demographics (except perhaps, executive recruiters) to make me believe that the sample I am tapping on LinkedIn is going to be generalizable.
6 Oliver Picher // May 5, 2008 at 9:30 am
Tom,
Social media could provide researchers with more information about respondents than ever before possible. Facebook or LinkedIn profiles, for example, could provide background on a person far beyond their responses to a survey questionaire, and the “social map” also provides a context to a survey response that was never before available.
On the other hand, the people on today’s social media sites are a self-selected population, and I wonder whether it would be possible to build a representative sample of the general population by polling users on Facebook or LinkedIn.
Oliver
7 Brian // May 5, 2008 at 9:48 am
I agree with Chris and sampling is getting more and more difficult as people move away from landline telephones, and more towards all sorts of mobile and mobile web solutions.
It will depend heavily on the type and purpose of the research being conducted. In general consumer market research, our clients are asking for more and more “needle-haystack” respondents. I joke around by calling it “extreme target marketing” but that’s the sort of requests we’re getting nowadays.
But getting back to the question, there will continue to be questions around the validation of the respondent with any sort of convenience samples. Are they who they say they are? Admittedly there’s a form of validation as your # of contacts grows since others are validating you and your profile, but with all of the noise right now around Online Data Quality and consumer market research, the barriers are going to be set pretty high to ensure that the population surveyed, is valid and somehow projectionable.
8 Tom H. C. Anderson // May 5, 2008 at 9:55 am
Oliver, I agree. And I believe Link Analysis, and perhaps more importantly Text Analytics/Text Mining are the keys for unlocking that value. But in terms of what LI and Facebook are ready to offer to market researchers I think we are getting a bit ahead here.
In terms of sample representativeness SNS could solve some problems. For instance, if a person has >= 50 contacts as a minimum threshold of being on the panel, it is highly unlikely that this such a person would have duplicate accounts. That is because the effort to set up such an account with 50 links would be far greater than the incentive benefit achieved by being able to take a survey twice.
There are several other benefits as well of course.
9 Kevin // May 5, 2008 at 10:14 am
Given the demograpic and sociographic make-up of the LinkedIn “community,” I wouldn’t see research efforts in this space yielding anything generalizable to a broader population. At least not at the moment.
As someone offered earlier, I would see LinkedIn potentially providing access to specialized, difficult-to-access subgroups in the B2B community, specifically people in marketing-related disciplines.
One of the concerns I would have around any kind of attempt to generalize findings of a LinkedIn poll to the business community as a whole is representativeness of the proportions of specific verticals on LinkedIn. I would really want to see the proportional distribution of members included in each silo (i.e. consulting, marketing and advertising, market research, etc. etc.) that make up the universe of people on LinkedIn. For instance, I would think that people in marketing-related disciplines are perhaps over-represented on LinkedIn (relative to the broader professional community as a whole), while those in other verticals like medicine, education, and others are likely under-represented. Consumer online panels, sourced from double opt-in solicitations often go to great lengths to show their comparability to the general population as a whole, but also the flexibility to access specific demographic subgroups if necessary. In the B2B category, I might think LinkedIn would want to do the same if their aim is to produce generalizable results. If their point of differentation is solely access to a highly networked group of people in specific disciplines, certainly not as necessary to show projectability.
Similarly, while this is very tactical, I would think that some standards need to be established so that people are able to classify themselves appropriately, in the correct vertical, on LinkedIn. For example, if someone wished to conduct a LinkedIn poll and wanted to choose some select disciplines within which to conduct this poll (as I imagine they would), they would need to be confident they would be polling the correct people, based on what I might term “discipline selects.” There is too much ambiguity presently in how members classify themselves to allow for accurately “selecting” the right people (i.e. commissioning LinkedIn to poll a cross-section of people in ‘Market Research’ for instance). As an example, I have a small network of people, principally in marketing, advertising, and research. Some of my research contacts who are all doing similar work classify themselves as being in “market research,” but others also working in a similar space classify themselves as being in “management consulting,” “strategy,” “research,” and “marketing and advertising.” Either standards are required for classification or people need to be able to choose multiple discipline “selects” when conducting a poll…and have the option for obtaining guidance when doing so.
10 Tom H. C. Anderson // May 5, 2008 at 10:19 am
Paul Neto posted an interesting analysis of why Linkedin and Facebook would offer only polls here:
http://blog.paulneto.com/2008/05/linkedin-to-start-market-research.html#links
He points out that a panel is much more difficult to maintain.
However, I would argue it all depends on the long-term perspective and importance these SNS companies put on a potential Consumer Insights revenue stream. They do have some internal MR and DM staff already. To start a sample business or even a full service MR arm would take only a couple million and about 2 years, if done correctly. While MR is usually not a get rich quick business, it is a good steady income stream. Not sure if they would have the time/$/patience to follow through correctly. Obviously Snowball.com in my original example, unfortunately did not.
-Tom
11 Per-Eric // May 5, 2008 at 12:23 pm
Hello,
Nice question and topic.
I know a couple of youth communities that offer good market research services. But this s most most certainly not generating any revenue streams close to that generated from the ad sales.
There is obviously the issue of generalization there. But if you ignore that for a second and look at the possibilities to find various professionals’ in specific regions – LinkedIn would be great. I cannot really see surveys asking what brand of tennis racquets or toothpaste we like to be suitable. As LinkedIn is a business network, surveys need to focus on topics that are related to this and are interesting for the members. For a healthcare professional in Spain – I guess the questions would need to be related to healthcare/pharma and asked in Spanish.
However, I think that surveys should not be user-created. Surveys should be managed by a LinkedIn team that effectively can work with filters/segments in their database and arrange for survey sending’s. Any bad survey posted through LinkedIn would harm their brand – nobody else’s. The issue of respondent experience and survey quality is a key. Also there is a need to guarantee that respondents are unique and real. A good technical set up is necessary.
12 Mark // May 5, 2008 at 3:44 pm
I think a Linkedin panel would be terrific for recruiting for B2B qualitative studies: particularly for in-depth interviews, bulletin board focus groups, or other online qualitative methods.
For polling, I agree with everyone else - the sample would need to be random enough to be predictive. With specific verticals - hard to predict what the population might be. Guess it would depend on the study objectives and target goals.
Overall, I’m sure I’d find a use for a LInkedin panel probably sooner rather than later. At least it would be yet another option to the existing B2B panel providers.
My 2¢
13 Jenine // May 7, 2008 at 9:25 am
I would also worry about representative sampling for consumer studies. However given the information available for community members, LinkedIn may be able to provide some terrific B2B services for hard to find samples.
Jenine
14 Sigrid // May 7, 2008 at 10:15 am
Conducting market research via linkedin seems to be a great opportunity in various aspects, supposed a flexible research tool software can be provided:
- anonymous research
- customer specific research
- cross-cultural research
- easy to use (basic-level) statistic output
- interactive research
- business orientated
- results can be easily provided to those who joined the research with some log in etc.
In case linkedin starts such a project, more detailed information should be provided to all users!
Sigrid
Leave a Comment