Keep them talking: Survey best practices

17th December, 2008 - Posted by paydayperx -

First off, a huge thanks to Max Kalehoff for posting an intriguing question in his MediaPost piece, and then following it up with more discussion via his comments section. Interaction, interaction interaction is the name of the game and Max is uniquely talented not just at promoting his opinion but in promoting discussion, which is a totally different ballgame.

Thumbnail of Dec08 Darden paychecks w/survey at right

Thumbnail of Dec08 Darden paychecks w/survey at right (Click for larger version)

His original question was “what do you think about pop up web surveys?” and the responses, even from his follower demographic that has to skew towards digital aristocracy, were predictably negative. Quite honestly even the stereotypical Nigerian widow who needs my bank info - the lowest of the low in terms of internet business models - is unlikely to ‘anticipate’ or ‘enjoy’ pop ups. And the discussion goes on to lament how difficult it is to get survey data from the audience, whomever it may be comprised of.

What struck me, again, is what a unique business PaydayPERX is, and how we have some advantages in our meatspace moorings when it comes to making forays into the wild jungles of the online universe. One of the things we’ve been able to leverage well is using print space to collect permission-based data from our user base. Currently we’re doing this on Darden Restaurant’s payroll checks - Olive Garden, Red Lobster, etc, you’ve heard of them. In addition to the advertising we place there - special discounts to employees, and great visibility for our advertisers, I have to say - we also have a place to promote our own online surveys.

 

“Yipee,” you might think sarcastically - in addition to being forced to look at ads on your paycheck, now you’re getting hit with a survey prompt too. But the surveys work for the same reason our ads work - there’s something of value in it for everyone in the chain. The attitudinal difference is a major shift, so don’t miss it. We make money off of advertising, and the advertisers are our customers, make no mistake; but when the offers are consistently exceptional, employees are conditioned to not merely tolerate - but to actually ANTICIPATE - the product. Usage goes up, employees are happy, employers are happy, advertisers are happy, and we’re happy.  And the same exact model seems to work splendidly with surveys, and as I bring this back around to the subject at hand, it even holds true on web-based surveys.

  1. Think of survey interaction as a transaction. You seek a dialogue, and that implies not just you getting what you want but the person at the other end getting something too. Completing the survey is a transaction just like completing a sale, where the value of the money to the customer is less than the value of the product/service, and the value of the money to you is greater than the value of the product/service, and so goes the free market. How many people complete a sale where they get nothing? About as many people who fill out pop up surveys just for fun, I’d bet. Who was it that said “anything worth doing is worth getting paid for”? I believe it was the Joker, but regardless, it’s a truism.
  2. Decide what you want from the lead. Certainly their contact info and something to qualify them as a prospect, an email address to opt in to all your fascinating newsletters, bla bla bla - all that is a given. But unless this is the first one you’ve ever done, you may already have this information. You could do a username password thing, but seriously who’s going to bother remembering a password for something like this? Use a cookie, use ajax when they start filling the thing out to look for a match on something personally identifying like a phone number or name + phone number in your database. In other words, fall back on another best practices principle: Let the robots do all the boring work. The same goes for how you collect the info even the first time - Do you really NEED my city and state? Isn’t the ZIP going to be enough? Unless I’m going to be physically mailing you something you don’t need to type out ‘San Bernadino’ and then pick a state code off of a screen-obscuring combo box dropdown. It’s ugly code and just because everyone else does it is no reason for you to do it.
  3. Determine what it’s worth to you. If you’re a retailer you had better already have a very precise dollar figure to assign to this piece of data - how much is a lead worth? How many leads does it take to make a sale, how much is the average sale…with this math you can come up with a dollar amount and if you’re paying on clicks you should be in the same ballpark with this number. The more specific and objective you can be, the better. “Leads are worth everything to us” is a nice but useless sentiment in this case.
  4. Flesh out the other side of the transaction. Come up with a sustainable reward. You want them to give you X data that is worth $Y dollars to you. Hopefully you see where this is going - what can you give them that is worth more to them than the data, but costs you less than $Y? In our case we run two types of surveys - short and quick that don’t even collect name or contact info, just some anonymous demographics, and a long one that collects more, but comes with a bigger reward. The short one poses a clever little trivia question, and to get the answer you have to complete the short survey. Despite purposely not requiring ANY of the fields, we’ve never gotten a single one that wasn’t completely filled out, so we know the reward is worth the trouble and the trouble is worth the reward. The long one requires more effort and enters you into a 1 in 10 chance of winning 2 free movie tickets. Those are good odds, so we get a lot of surveys, but the data is still being transacted at a profit for us because of our transaction analysis and value assignment. The latter are used to attract more advertisers, because they tell us what offers they want to see on their checks.

Can you do this kind of thing as a web pop up? Skepticism is very high whenever a pop-up appears, especially thanks to the ‘win a free* MacBook Pro’ crowd, who are feeding down at the bottom with the Nigerian Widow in my book. You overcome skepticism with sincerity in your dialog, so if I trust you I’ll pay attention and if I don’t, I won’t. I’ll fill out a coke survey popup on coke’s site but I won’t fill out free clipart survey popup on a free clipart site, because it seems suspicious and quite frankly the clipart site is a bad neighborhood in web terms. I’ve got one eye on my spyware blocker and another on my virus protection whenever I visit those kinds of places. The environment matters, and its a metric that needs to be more clearly noted when you’re looking at your deployment. Ultimately I recommend doing the cool stuff on your own site, and using branding and positioning to increase trust with users.

There’s plenty more to be said on the subject of doing surveys well - I haven’t even touched on data collection or query methodology, and really if you don’t have that then there’s no point in doing the survey either. But on the assumption that the database stuff is pretty standard and creativity is needed for the design, hopefully this can provide some fuel for the idea machine in your head.

Now, keep those thought-provoking pieces coming, Max! And PS MediaPost is essential, if you don’t get five or more of their daily newsletters you may as well poke one eye out on purpose.

Tags: , ,

Posted on: December 17, 2008

Filed under: Work

Viewing 1 Comment

 

Trackbacks

(Trackback URL)

close Reblog this comment
Powered by Disqus · Learn more
blog comments powered by Disqus