Three Reasons You Should Ignore the Experts

Analytics · By On · 5 Comments

Recently, I watched a video where another ana­lyt­ics ven­dor pre­sented their take on which mobile plat­forms and chan­nels retail com­pa­nies should be invest­ing in. What was their con­clu­sion? Focus on the iOS plat­form. They were wrong and here’s why:

Big Mis­take #1—Used Faulty Data

Data used in the analy­sis was not col­lected fol­low­ing mobile best prac­tices. In fact, all devices with­out robust JavaScript sup­port were effec­tively excluded; a mis­take that elim­i­nated from analy­sis a sig­nif­i­cant per­cent­age of the site’s mobile vis­i­tors and increased the appar­ent impor­tance of the iOS platform.

Big Mis­take #2—Ignored Mobile Experiences

Mobile ini­tia­tives were not included in the analy­sis (e.g. mobile-optimized sites, mobile apps, etc.). As I’ve dis­cussed before, pro­vid­ing mobile-optimized expe­ri­ences can pro­duce a very pos­i­tive impact on key met­rics. In fact, these expe­ri­ences can eas­ily dou­ble engage­ment and other key met­rics for mobile users—something com­pa­nies in Japan have known for years. Leav­ing out this very impor­tant lever­age fac­tor pro­duces a faulty bias towards the iPad and by exten­sion iOS.

To illus­trate, I’ve included a seg­ment of actual cus­tomer data (see chart below). You’ll note for this par­tic­u­lar cus­tomer, rev­enue per visit for iPhone vis­i­tors is only aver­age com­pared to other devices on both mobile and stan­dard chan­nels. In addi­tion, the most pro­duc­tive mobile device on the stan­dard site is the iPad both in terms of total rev­enue and rev­enue per visit. When you com­pare iPad per­for­mance on the stan­dard site to the mobile site, you’ll notice rev­enue per visit is much lower on the mobile site—an indi­ca­tor that iPad usage aligns more closely with the pro­file of a desk­top user than that of a mobile user.

After fac­tor­ing out the iPad and includ­ing the impact of opti­mized expe­ri­ences on other mobile devices, iOS mobile rev­enue con­tri­bu­tion comes in around 60% for this cus­tomer. That’s still a big num­ber but cer­tainly not large enough to jus­tify an exclu­sive focus on iOS.

mobile revenue per visit

Big Mis­take #3—Projected Seg­ment onto the Market

Based on a hand­ful of retail cus­tomers, the expert made a rec­om­men­da­tion for the entire mar­ket. In real­ity, behav­iors, usage pat­terns, and busi­ness objec­tives vary widely even between com­pa­nies in the same indus­try. Given these and other dif­fer­ences, using another company’s results to define your own strat­egy is fool­hardy at best.

Where You Should Invest

So which mobile plat­forms and chan­nels should you invest in? The ones that bring the best ROI for your busi­ness! Keep in mind these chan­nels and plat­forms may not be the ones rec­om­mended by the so called “experts”.

  • http://www.gilliganondata.com Tim Wil­son

    Nice post! I’ve run into sev­eral cases where some­one has taken “mobile data” and jumped to a con­clu­sion that seems pretty half-baked. The most com­mon sit­u­a­tion is when some­one sim­ply looks at the data from within Google Ana­lyt­ics and accepts the “Mobile” traf­fic num­bers at face value — ignor­ing the fact that Javascript must be enabled *or* some sort of server-side or sta­tic workaround has to be in place. There seems to be a dearth of infor­ma­tion on the ‘net about “how to get a rea­son­ably accu­rate han­dle on traf­fic to your stan­dard site from mobile devices.” Even in your exam­ple above, no RIM devices cracked the top 10. I’m assum­ing that’s because the check­out process on the stan­dard site is com­pletely unwork­able on a Black­berry? Where does Sitecatalyst’s mobile device track­ing fall on the ease of imple­men­ta­tion and com­pre­hen­sive­ness of devices front rel­a­tive to the other major WA ven­dors? Where is a good write-up?

    • http://blogs.omniture.com/author/ehewett Ed Hewett

      Thanks Tim. You’re right-no Black­berry devices cracked the top ten on the stan­dard site but it isn’t because Black­berry users don’t spend money on mobile as you can see from the mobile site rank­ings. With regards to mobile-centric imple­men­ta­tions, Site­Cat­a­lyst has a straight for­ward approach appro­pri­ate to each of the most com­mon mobile chan­nels (stan­dard web, mobile web, mobile apps, mobile mes­sag­ing, rich media). As needed, the imple­men­ta­tion for each of these approaches can be increased in com­plex­ity; and of course we’re the best out of the major WA ven­dors :) . I’ll send you an overview in PDF form in a sep­a­rate email.

  • http://www.jobsfornaija.com jumoke

    Yeah tim, you are right. i always won­dered why the dis­crep­ancy between my google ana­lyt­ics data and the data from the feed­jit wid­get on my web­site. i guess it means i prob­a­bly have more traf­fic than i think i have.

  • http://msnbc.com Tiffany Doyle

    Can I get a copy of that .pdf overview as well, please? Thanks.

    • http://blogs.omniture.com/author/ehewett Ed Hewett

      Sure thing Tiffany. I just sent this over direct to your email.