Lie #1: “Improving conversion starts with color testing your buttons. Button testing is THE most important thing because red/orange/green/rainbow/polka-dotted/zebra-spotted buttons perform x times better than any other color buttons.” You know what? Red/orange/green/rainbow/polka-dotted/zebra-spotted buttons can perform x times better than any other color buttons and in the big scheme of things, there are hundreds (yes, hundreds) of tests that you could be doing that are more fruitful. You can test the size of the buttons (the bigger the better), the number of buttons (the more the merrier) and the language on the buttons – they’ll all make a difference.  But is that your biggest swing?  If you want to hit it out of the park, is an orange button versus a red button really going to do it?  Probably not.  You’d be better served by testing new product(s) OR a new offer OR something – anything — else that makes you feel queasy in your tummy (keeping-you-up-at-night-worried-you-might-not-have-enough-stock-to-support-it uneasy.)

Truth #1: GO BIG OR GO HOME. You’ve got limited time. No, I’m not predicting your eventual death (I have a 900 # for that.) What I am saying is that you need to identify the home runs and swing for the fences when it comes to web testing. These teensy things that “improve conversion 613%” but result in only a handful of orders?  Not impressive.  And probably not worth your time either.

Lie #2:  “When you buy a good search package, you can just set it and forget it.”  Things like “good search package” and “responsive design” are completely overrated. Yes, they can both be fabulous but neither of them will solve all your problems overnight. (Adding an outside search package is a big expense for a lot of folks and with spending a significant amount of money comes a tendency to think that all your issues will magically disappear. Sadly, that doesn’t happen.)

Truth #2: You’ve got to work your search. Start by looking at the top 10% of your products and services.  Are they well represented?  Is the copy useful and compelling?  Does it tell a story?  Does it have all the information/details your users need to make a purchase?  Is your copy friendly for online users or did you just plunk in your catalog/offline copy? (Incidentally, that’s one of the biggest mistakes legacy marketers make.)

You’ll also want to look at what people are searching for – start with the top 100 searches to make it manageable.  Look at what they’re finding – what they’re not – and what they’re clicking on.  Tweaking/working those will all help you make your search better. For example, a client of ours noticed a bunch of their major competitor’s item numbers over and over in their unsuccessful searches.  Their “learning” search tool had redirected a couple (of the many) searches to similar products but not the ones that the client would have chosen.  Also, the search package was showing “no results” for most of the items. To fix it?  They added the culprit item numbers to the “right” product pages (with solid success, I might add.)

Lie #3: “_________________ works because so-and-so tested it x years ago and the results were posted {here.}”  Look, I’m all for best practices – I know some people hate them but I think they’re useful. However, one poorly executed test to a market that is NOT yours to a sample size that is NOT significant?  Yeah, I’m not a fan. Not. At. All.

Truth #3: You’ve got to figure out what works best for you and your company.  Best practices are good places to start.  Make sure they are practiceS, with an S, however.  Not one company doing one thing that may or may not have been executed properly.

Lie #4: “It doesn’t matter what kind of user testing you do, as long as you do it.”  I’ve made more mistakes than the average bear when it comes to user testing. Like eleventy bazillion times more. One of the biggest mistakes?  Thinking some random ass tester from East Podunk, Mississippi who knows nothing about my product, nothing about my market, and nothing about what I’m trying to do, would solve my site’s problems.

Yes, I have used several of the remote testing sites.  Yes, I like some of them more than others.  Yes, I think they can be useful.  And yes, I think that a lot of times it’s like giving a lollipop to a child who just fell from the top of a jungle gym TO THE ASPHALT.  It may stop their crying but it may not prevent them from having a concussion.

Truth #4: If you’re using folks to test your site, use a good balance of prospects, customers, competitors’ customers AND people who don’t know you from Adam.  You should NOT use only people you’ve found on the street, your kid’s preschool, or from your great grandmother’s senior home.  Yes, you can use some but people who actually have propensity to buy your product will behave differently and their needs should be represented.

Lie #5:  All mobile transactions are created equal.  This comes up a lot at mobile conferences.  I know because I am often the morning keynote and then, for the rest of the day, speaker after speaker who follows me, wastes 5-10 of their allotted minutes, pleading their case about why I’m wrong and/or an evil demon spawn. (There are many reasons, this is NOT one of them.)

I get it.  I really do.  Mobile conversion sucks – and I mean REALLY sucks, and if you separate the tablet traffic from the phone traffic, it’s even more demoralizing.  So, the easiest way to make it look better/sexier than it really is, is to say “oh, who can define what a phone is?  Is a Nexus 7 a phone or a tablet?” and then greedily rub your hands together like that Mr. Burns guy from the Simpsons as you watch the attendees nod like bobbleheads.  Newsflash Mobile Speakers of the Planet: if it’s not the primary device you hold up to your ear as you take or make calls, we can safely call it a tablet at this point.

Truth #5: It’s smart business to separate your phone traffic from your tablet traffic.  Why?  Because it’s likely that your tablet conversion is fantastic (it’s often 2-3x desktop conversion these days) and your phone traffic is pitiful.   That’s not what’s important though.  What’s important is that you get an honest idea of what’s working in each medium so you can make the appropriate tweaks/fixes/overhauls.

Have something you’d like to add?  Send it to info@amyafrica.com or tweet it to @amyafrica.  If it’s good, I’ll include it in one of my next posts.