Marketing News: 10 Minutes with John Ouren


Published on

A Marketing News interview with John Ouren, Senior Vice President, TrueSample

Published in: Business, Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Marketing News: 10 Minutes with John Ouren

  1. 1. 05.15.08 A PUBLICATION OF THE AMERICAN MARKETING ASSOCIATION THERE’S NO QUESTION that the online research industry—that multibillion-dollar behemoth—provides marketers with cheap, abundant and readily accessible quantitative and qualitative data from a growing number of online panels. But there is a question as to how reliable that data is. MarketTools Inc., a San Francisco-based online research firm, introduced a solution called TrueSample in April. Using the same technol- ogy that works to prevent identity theft and credit card fraud, TrueSample helps vet online survey respondents, allowing marketers to ensure thatMINUTES WITH… those survey takers are not only authentic but also engaged. Venerable marketing giants including Procter & Gamble and General Mills have already signed on.JOHN OUREN, EXECUTIVE VICE PRESIDENT John Ouren, MarketTools’ executive vice president and general manager of panels andAND GENERAL MANAGER OF PANELS AND communities, recently spoke with Marketing News about TrueSample and how his company isCOMMUNITIES FOR MARKETTOOLS INC. working to glean higher-quality and more trust- worthy data from online research.BY ELISABETH A. SULLIVAN//STAFF WRITER
  2. 2. marketingnews  05.15.08Q: Obviously one of the key issues in online Q: How do you measure someone’s level of [MarketTools’ online consumer panel] … theyresearch is the survey takers’ authenticity and engagement in a survey? basically provide us [with] information aboutthe data’s reliability. What role does True- themselves … very high-level demographicSample play in ensuring that the research A: There are a number of different techniques information: their date of birth, their gender,comes from the source it’s purported to have which we are automating into our platform, the and their name and address. And we basicallycome from? first of which is the time it takes a survey partic- require that the information should be correct, ipant to complete the overall survey and various should be able to be verified by some indepen-A: The industry … is going from sort of a different sensitive portions of the survey, and dent, third-party data source. … We believechild phase into an adult phase. The child how does that compare versus the mean? How that independently verifiable informationphase was: ‘Isn’t it neat? We use these tech- does that compare versus expectation? And we should be validated at the point of entry intoniques, but we’re not going to bet our business will throw out outliers. … a panel. So if we can’t validate that informa-on it.’ … [Now] the only research some firms Another element is how they’re actually tion, we do not allow that particular partici-do is online. And it’s a meaningful portion of interacting with particular types of questions; pant to do client-facing surveys. We may allowthe share of what even the firms that still do for example, a matrix question. There are differ- them to do surveys for our own experimenta-most of their research offline [are doing]. ent ways of answering those questions, which tion and other reasons. We may accept them It’s going through what I would consider might indicate disengagement: ‘straight-lining,’ into the panel, but we will not allow somebodyto be an adolescent phase, and an element in other words, picking the same thing all the that we can’t validate as being somebody thatof that is some concerns related to trust. … way down the matrix; … [and] patterns, which is real and non-duplicate on the panel fromBecause it’s online, and because I can’t see people naturally do when they’re really not the get-go. In the database world where I comethat person and I don’t have necessarily all reading, but they’re just kind of answering and from, it would be considered to be basic listof their detailed data … I have some ques- trying not to appear like they’re straight-lining. hygiene. In other words, you make sure yourtions about whether that person is authen- And also a third category is inconsistency file is valid, you make sure your file has been—tic. I have some concerns as to whether or in responding to the same information being to the greatest degree you can—screened ofnot I’m interviewing the same person over asked twice, what they call in the industry track non-duplicates using data.“Without biometrics and retinal scans and fingerprints ...we take every other approach to, at least, validating that the information that the person is providing is consistent with a real individual.”and over again, i.e. are they unique? I have questions. … Inconsistent results may indicate a Q: How do you feel about the role of incen-some questions related to whether or not lack of engagement. … tives in online research? Do you think theythe survey person is actually engaging in the These are things that are done in the have the potential to mar data because peoplesurvey. You know, are they distracted? Are industry and they’re part of best practices, are just filling out surveys to get the freethey actually malicious and they’re using a but they’re not done in a consistent, rigorous sample or the gift card?bot … to take massive quantities of surveys way by which you score the panelists basedin an automated fashion? upon their reaction, and you can measure the We felt it was really important to apply degree to which they are exhibiting unattract- A: We do research, primarily community-data and technology, which has really been ive or suspicious behavior relative to an aver- based research, without incentives whereestablished in other industries, to those three age panel. we really have relatively small populationsquestions to be able to assert unambiguously to do more qualitative work. … Quantita-that each individual person that responds to Q: How can you prove the survey taker’s tive research requires numbers to the degreea survey from our panel or a partner’s panel authenticity? that recruiting populations that create statisti-… is actually real. They are who they say they cally meaningful results in various small sub-are. They are non-duplicates ... each individ- A: There are always limitations, just like there segments really requires aggregating very largeual response comes from a unique respondent. would be on a phone call. Without biomet- panels. Nobody has broken the code on beingAnd … their participation within the survey is rics and retinal scans and fingerprints—which able to do that type of research without someconsistent with behavior that we would regard I think would fail almost any privacy screen level of incentive, virtual or otherwise. …as valid or ‘non-suspicious.’ It doesn’t show a anyway and are really unreasonable—we take To not have an incentive at all I think islack of engagement or distraction. It doesn’t every other approach to, at least, validating ideal. And in the case where you’re dealing withshow a level of professionalism, which would that the information that the person is provid- small communities who are really engaged,be concerning, as somebody [could be] taking ing is consistent with a real individual. and they’re providing their opinion in qualita-surveys in a factory-like fashion. When somebody signs up for ZoomPanel tive research and are operating in a forum and
  3. 3. marketingnews  05.15.08 community-type basis, I think it is very possi- advertising world, and there are a lot of adver- upon the technology that we can bring to ble to run those types of situations—and we’ve tising analogies that say that certain mediums bear on the market, we have an approach that seen it and we’ve done it over and over again— are appropriate for certain types of advertis- allows any researcher to make sure that he without incentives. But for quantitative research ing. And radio is a much bigger business than is vetting his particular panel in an aggres- and building large panels, to do very low-inci- it was 80 years ago, television’s a bigger busi- sive fashion. … The way I see this is that there dence study work or consistently representative ness; does it lose share of wallet? Definitely. So could be multiple approaches, but there’s a populations even in consumer work, we think my sense of online research is it’s still in the lot of power in really aggregating around a that those incentives need to happen. early stages of cannibalizing and gaining share common, objective, independent third party. The question is, can you make a living on of wallet as these innovative technologies get incentives? And if you’re not careful and you’re to be applied to different methods. Q: Because there’s strength in numbers, have not really enforcing these types of standards And it’s not uniform. … There are other you encountered anyone who is hesitant to that I’ve described on a rigorous basis, you types of research where the technology may be apply TrueSample because it will weed out open your panel up to systematic fraud. And applied to further move share towards online; too many of the panelists that the researcher that’s what I’m more concerned with. It isn’t, there may be certain methods, which forever worked so hard to accumulate? ‘Well, do these incentives encourage people are best explored using offline techniques. I to bias the research?’ And we’ve done enough sort of see the example of the mix of adver- A: I have a high degree of empathy because research on research to indicate that’s not the tising across different mediums as technolo- we spend a lot of money on recruiting. And at case. What we are concerned about is that gies have emerged as a really good indicator of the end of the day, we’ve applied all of these we [don’t] go out into the marketplace, and what’s likely to happen to research. techniques to our panel, and the layered effect on behalf of a client, we buy a sample from of that is that we’re going to lose 10 to 15% a panel where fraud has been allowed to run Q: Has the research industry begun of our panel. And we’ve done it to a couple of rampant. … With TrueSample and the certifi- to demand certification or authentication other panels out there, and they’re going to be cation process, we’re going to do everything we tools like TrueSample to ensure losing similar amounts. can to make sure that somebody isn’t making a that the research they’re doing online The cost is entirely economic, though. … living off of incentives. is meaningful? Those panelists that you spent $10 or $5 or $20 acquiring, you’re going to have to not get the Q: With the certification that you can do A: We do think that a solution needs to long-term value out of those people. But if you now, do you think online panels will start be objective and rigorous. And one of the can apply these techniques going forward and to take the place of in-person panels? Will concerns I have of, sort of, a highly balkanized integrate them into your recruiting process, the online version eventually supplant the approach where everybody is saying, ‘Well, I’m then you can avoid the expense to begin with. in-person version? doing the same thing,’ is who’s the indepen- … We don’t pay for people that we can’t vali- dent, objective third party that’s vetting this? date, and there’s nothing that prevents other A: That question is what I call, ‘Will video We’re not the biggest player in sample. firms from doing the same thing. So to me it’s a kill the radio star?’ And the answer is it never There are very large players out there that are one-time hit. … And if it’s something that the happens. We still have radio. It’s a 1911-whatever focused around building huge panels. What sample business has to do to make sure that the technology, and it has its place in advertising. we’re focused on by delivering this effort is to online research industry continues to grow, it’s I’m from the direct marketing world and say, ‘On behalf of the researcher, and based a small price to pay. mama_reprintline 9/29/06 4:31 PM Page 1 B/W Reprints/Self_Covers/AMA Reprinted with permission from the American Marketing Association