Wednesday, April 01, 2015

ANA conference: surveys

What Do Consumers Think? Using Online Surveys To Demonstrate Implied Claims
David G. Mallen, Co-Chair, Advertising Disputes, Loeb & Loeb LLP: NAD now forum of choice for many ad challenges, especially since the standard of proof is different for implied claims. Survey not required but may be useful. 
Kelsey Joyce, Senior Director, Legal Affairs, T-Mobile USA
We deal with competitive ads all the time.  Survey: in 43(a), very helpful; is it worth spending the money on survey for NAD?  Discussed with marketing clients as well as external lawyers/survey expert.  Timing: if this is an ad we really want out of the market—and they all are!—we might not want to take the time to do a survey in order to get the challenge started. 
Hal L. Poret, Senior Vice President, ORC International: possible to put together a survey with 2 weeks’ notice, which can be important w/NAD.  Difference between online and mall survey may allow you to supply a rebuttal survey in short time.
Mallen: what’s candidate for online survey and needs mall intercept?
Poret: what’s the ad and how is it being shown? Online survey may have very small screen.  Could be desktop/laptop.  Even tablet/mobile phone, though you want to stop that if you can. Can it be fairly presented on computer screen?  TV ad w/small print, or graphics/charts that might be harder to read, think carefully about whether showing it on a computer screen would be challenged.
People also move quickly through unsupervised online surveys; want to get through it.  Human interviewer: social pressure to respond; interviewer takes down answers for them = longer, more detailed, thorough answers. If you need people to speak in their own words, online may be more difficult.
You don’t always need a human interviewer—majority of NAD cases allow online; advertisers are often trying to go up to the line between true and false, and thus you almost always need a closed-ended question, and online surveys are ideal for closed-ended questions.
Q: “Nobody knows you’re a dog”—is that an issue?
Poret: it’s not to me; that’s how marketing research works these days. We work with large online panels that recruit lots of people and work to comply with standards. We have techniques to know who we’re inviting—DOB, gender, etc. to know who’s taking the survey.
Q: controls?
Poret: in some ways online surveys lend themselves to what you want as a control—often the most effective thing is altering the original ad to clarify something or make it true. Digital alteration is often most desirable, and presenting it online makes sense.
Joyce: NAD Case No. 5686, T-Mobile challenged Sprint’s campaign for Unlimited My Way monthly service plans. Challenges: Ads w/specific scenarios depicting how consumers can save money imply that consumers will save. Guaranteed Unlimited For Life confuses consumers about whether “for life” applied to the $80 monthly fee or the unlimited talk, text, and data.  “Guaranteed for life and only from Sprint” implied that only Spring had unlimited talk, text, and data. Considered not doing a survey because it seemed misleading on its face. We thought consumers would take away message that the price was part of the fee.  Decided to survey because (1) wanted backup, (2) were challenging another Sprint ad that they thought needed a survey, so taking the time was a nonissue.
Poret: control was clear cut because the issue was combination of “for life” with $80 in close proximity.  Control: unlimited for life, eliminating $80.  Challenge: didn’t show entire webpage, just ad banner. But NAD was satisfied with explanation that nothing else on the page clarified the offer and that this was a standalone ad.
Mallen: issue is net impression, but net impression of what? You may sometimes have to test an entire webpage.  When would that be?
Poret: other content possibly right above or below that bears on that.  Headline, graph, and then a paragraph of text. 
Joyce: we captured the entire page so we could show how the test and control were displayed.  Control: “only Sprint delivers unlimited for life,” without the $80.  The control ad is an ad that we can live with in the marketplace at the end of the day. Create a blueprint for Sprint to fix what we think is the deception.
Poret: NAD skepticism about closed ended questions makes it really, really important to have a good control that shows that closed ended questions on the control didn’t produce the deceptive answer. “Based on the ad, what is guaranteed for life?” Please be as detailed and specific as possible.” After other filter questions, including whether the ad communicated anything about a guarantee for life.  NAD will want that.
Test group: 34.5% said guaranteed $80/life when asked the intro broad questions “what did this ad communicate?” 5.5% in the control group said the same thing. Net 29%. Didn’t even need the closed-ended questions.  If you did, 57%/9%, net 48%.  NAD accepted the survey evidence.
Joyce: other challenged ads were tougher.  $83/year offer depended on buying one particular phone, iPhone 4. Disclaimer was at the bottom but we thought it wasn’t clear; offer was “our most popular free smartphone”—but that wasn’t the most popular phone, smartphone, free phone, or iPhone even at Sprint: it was the most popular free smartphone at Sprint.
Poret: problem was not that something needed to be removed, but that something needed to be added: “when you choose an Apple iPhone 4” was control.  Here we needed closed-ended questions much more because it wasn’t the kind of ambiguity people would resolve on their own.  Online survey works well here because you need the closed-ended questions.  55% in test said that the savings would apply to any phone; went down to 18% with control ad—helped convince NAD that the survey was reliable.
Joyce: we’re more willing to run a pilot survey before the NAD, because that’s not discoverable. But we do think about discoverability even at the NAD; just because we’re not litigating now doesn’t mean we won’t be soon, especially when we’re an advertiser defending the claim.  Follow-on consumer class action lawsuit is often an issue.
I need a control ad that we can live with if we won in the marketplace.  We absolutely every time we challenge an ad, we think about how this will impact our own advertising. 
Poret: Different perspective because it’s not his role to design Sprint’s advertising for them, and there’s no magic answer to the question of how it should be. I’m trying to create something that will allow me to test whether my questions are producing answer X when I know from this ad that they shouldn’t answer X. I have to be satisfied that a reasonable person shouldn’t come away from the control thinking the offer applies to any phone instead of a still confusing version, so I know what I see is just noise.
Q: what about TV ads?
Poret: that would go to what’s in the ad.  I do such surveys frequently, mainly where there are strong takeaways. Sometimes important info is on the screen in the ad that I worry about someone seeing in an online survey.  Maps/charts/graphs/legents/mouseprint. Don’t want to risk people can’t see that in certain scenarios.
Go to court: you don’t know what judge you’re getting, whereas NAD knows me and probably the other survey expert—familiar with expert battles.
Joyce: some judges will accept any survey, while others will never accept one.

No comments:

Post a Comment