Hit ’em where they surf

Editor’s note: Leslie Townsend is president of Kinesis Survey Technologies, Austin, Texas.

Our firm coined the term Thumbpower to refer to the increasingly adept and rapid utilization of our opposing digits to type, text, check e-mail, surf the Internet and conduct the myriad transactions that are increasingly migrating to mobile devices. Recent press has played up a Morgan Stanley report predicting that within five years “more users will connect to the Internet over mobile devices than desktop PCs.” The growth in smartphones - now used by nearly one-fourth of the U.S. mobile subscriber population - shows a rapid shift even while in a down economy, so this vision seems realistic.

What are the implications for the market research industry? From the online data collection perspective, discussions have focused upon the user experience of a mobile survey. From a panel perspective, attention has focused upon the challenges of projecting results from a population that uses their mobile devices for survey-taking. It is time to think about the mobile experience in full circle, from the respondent perspective, for control is only an illusion, and in terms of determining the mode of data collection online panelists will choose, the illusion is already a thin one.

Will be incomplete

The “full circle” mobile experience includes the survey, the panel, the community and social media. The mobile survey experience will be incomplete if it consists only of respondents taking surveys on mobile devices, with researchers and panel partners determining who should be invited to take mobile surveys. In a typical panel situation, the experience must be one that is the choice of the respondent. In fact, it already is, because the respondent determines where they receive their e-mail. The issue for researchers is whether (and when) the non-response bias from individuals attempting to use a mobile device becomes an untenable issue.

As a vendor of online survey and panel management solutions, our firm has a great deal of metadata drawn from historical surveys that is available for traffic planning and respondent verification. In a random audit of our own traffic, we found that 2 percent of respondents entering Web surveys entered using a mobile device during the past 12 months. These individuals were almost certainly recruited from the Web when they were on a traditional PC since the vast majority of them came into the survey from an invitation sent by a public access panel.

In order to have a complete mobile experience, the survey respondent must be able to replicate the Web experience as much as possible. This includes:

  • being returned/redirected to the mobile Web site that they came from - if they saw the invitation link while browsing a mobile site;
  • being able to check and redeem their incentives from a mobile portal - if they came from a panel;
  • being able to enter communities and engage with social media on either device.

To prevent non-response bias, it may also be necessary to recruit some panelists who are mobile-centric (panelists who do not necessarily originate from a traditional Web source because they are more likely to be using their mobile device than a computer). They might originate instead from a mobile site, a social networking site, or another source - which are themselves gradually morphing. For this reason, a panelist portal needs to offer respondents a stimulating mobile experience - registration with double opt-in, ability to retrieve lost passwords, the ability to check and redeem incentives and other components that they are accustomed to on the panelist Web site or on their community site. This may create some challenges in authenticating panelists who might have been recruited in one mode versus another - a single authentication process is highly desirable.

Several challenges

All of this creates several challenges for the market research world, which must leave behind the notion of creating an identical user experience for all.

It creates greater complexity in optimizing the Web site experience for panelists. While a mobile portal for panelists does not necessarily need to offer the full functionality of a Web portal, panelists who use both Web and mobile sites are likely to find the mobile site disappointing if it is incomplete in providing a feature that they use frequently.

It adds to programming and testing complexities, as well as development time - and hence has cost implications at a time when most researchers are looking for ways to reduce data collection expenses.

Lengthy profilers may pose a challenge for mobile survey-takers. While the industry trend has been to use multiple, relatively short profilers to keep new panelists engaged, there may be a need to redesign and simplify even further.

If profilers have been designed to be engaging - which often entails the use of interactive question types - they must be redesigned for mobile devices. Most mobile devices will not support Flash, JavaScript or even table structures.

A basic decision must be made regarding what types of devices will be supported, and what exactly “support” consists of. Netbooks, gaming devices and televisions are all utilized for Web access, thus the same arguments for supporting mobile devices must be extended to these devices as well.

There is no doubt that the experience of smartphone users is a much richer one than those using older devices based upon Wireless Markup Language, and the vast majority of respondents surfing the Web are utilizing smartphones. So while Kinesis supports devices of all capabilities for survey-taking, most of our clients have decided to optimize their panelist Web sites and mobile surveys for smartphones rather than all mobile devices - since these represent the preponderance of mobile browsers. This means, as an example, that a registration form can include multiple fields on a single page, and a smartphone user can utilize their touchscreen to navigate from field to field, which is significantly preferable to using navigational keys.

There may be projects that simply do not render at all to the vast majority of mobile devices. A highly qualitative project requiring large amounts of text input is one example. A card-sorting exercise might be another. Therefore, it is both desirable and necessary, at least in the short term, to restrict mobile respondents from entering some surveys and/or survey exercises. This can be achieved in a variety of ways, with the preferred method being automated device detection for incoming respondents.

Invitation modes will need to be reexamined. If SMS is included into the invitation mix, per-messaging fees apply and it will become far more critical to optimize and route panelists appropriately to more closely target quotas. It is much more important to optimize sending times for text messages than e-mail since most phones have an audible alert when the messages arrive.

Mobile recruitment modes

One can take this exercise further. To truly reduce non-response bias from mobile-centric respondents, the panel must also include mobile recruitment modes. This may entail extension of river streaming techniques to mobile Web sites and social media sites, which then entails a change in recruitment strategy, potentially to include downloadable applications (these apps can bias the panel because demographics of phone ownership vary significantly from OS to OS). But undoubtedly surveys may need to interact with mobile applications, gathering data from them, becoming embedded in them and sometimes becoming the app.

This migration is taking place much more rapidly than one might think. The efforts of Paradigm Sample, Port Washington, N.Y., are a prime example of the recent trend in moving towards a dual-mode (Web and wireless) panel. Paradigm acquired one of the first dedicated smartphone panels in the U.S. with its recent integration of Data Innovation Inc.’s panel assets. What made Data Innovation unique was that members of its panel participated exclusively through their smartphone devices - accessing a mobile panelist Web site to register, update account info, take available surveys, etc. While Paradigm clearly sees potential in continuing to market these smartphone panelists, the same technology has allowed it to better communicate with its traditionally Web based panel. Now any of Paradigm’s IdeaShifters panelists can access either a mobile Web or standard Web portal to participate in the panel. The panelist experience is almost identical, allowing Paradigm to stay connected with its members throughout the day while capturing data closer to the point of impact. Kinesis envisions this continuing as the mobile browsing experience merges more closely with that of traditional Web.

New data

Emerging from all of this comes new data that the market research industry will devour: location-based information, arising from the point of purchase and consumption and potentially linked back to panelist profile and survey data in real time - all via the rapidly expanding emersion of mobile technologies in our daily life. And that ought to create a lot more Thumbpower in motion.