Category Page Template Test
| VERSION A | |
How site visitors voted:
VERSION A (65%)
VERSION B (35%)
|
Actual Test Results:
(Click back to see Versions A & B) The Version A template resulted in 7.74% more purchases; plus, it lifted overall site conversions by 22.18%, and average order value by 7.06%. On the downside, it also had a whopping 19.45% higher bounce rate than Version B. Based on expert advice from WiderFunnel Marketing Optimization, Elastic Path Software, Inc. conducted this A/B test on the Official Vancouver 2010 Olympic Store using Google Website Optimizer. Overall Version A had a more cluttered design, including a column of “Most Popular” items, product-color info, and links to shipping, privacy, and return policies. (Sort wording was also tweaked a bit.) As you probably know, best practices in Web design generally say simpler pages do better. But, in this case the simpler page was the loser… except for bounce rate. We wonder if this means busier pages can scare off less-motivated visitors even while they may increase your best prospects’ conversions? Ultimately, persona-based design may be the answer. |
Product discovery is one of the big challenges for e-comm stores. Having a popular products section provides social direction for shoppers and its not suprising this increased results. I would not attribute the lift to a more cluttered design, rather its a result of helping the shopper find “the best” products- as voted on by our peers.
Trial and Error.
I do wonder sometimes if it is impossible to actually get worthwhile advice from a consultant. Maybe consultants should market the idea that they can help you more efficiently do TRIAL and ERROR. THAT makes sense.
Cuase. I know for a fact that Crappy looking marketing can do tons better than ads developed by ‘marketing professionals’.
Go figure.
“Best practices in Web design”? That simpler pages do better is just plain wrong. It was wrong when people were saying the same thing about direct mail and direct advertising design when I started that…20 years ago.
It would be great know if the 8 main items (A) are easier for users to register than 10 items on version (B). An A/B test with quantity could be informative.
Nice test but i don’t think the page won because of the clutter – surely it was the addition of suggesting most popular products? This could be fixed by having less products on the left and designing with an addition of a list in mind.
@Kent – always be careful hiring a consultant that promises you certain growth or increase in revenue/traffic. I see our work as sherpas
@jkalil – each variation had equal amount of products displayed. In screenshots it would seem like it’s different, but the page was continuing. You can read more on http://www.getelastic.com/product-list-ab-test/
@Abi – While we didn’t have specific data, I would agree with you as well.
It might just be semantics but the terms ‘simpler’, ‘busier’, ‘cluttered’ perhaps need clarifying.
The ‘cluttered’ page to me appeared more interesting. The ‘simpler’ page was less interesting and it felt something was missing as it didn’t balance visually whereas the ‘cluttered’ page was more visually appealing.
It’s 2-col style layout held the attention better than the other and the filtering terminology clearer with ‘View by’ rather than ‘Narrow by’ which felt limiting and off-putting.
I wouldn’t be all that disappointed with a 19% higher bounce rate although it does sound high!
However, if the bounce rate of other page was 40% it represents just 7+% more bounces and if 60% just 11%. I would be more interested in some testing to reduce the headline rate.
With a lift in purchases, conversions and order value it seems to me a pretty good outcome!
I agree with Scott and Abi. More clutter isn’t a response booster. But I guessed this right, I think, because I figured the social proof provided by the ‘most popular products’ would prompt people who can’t think very well for themselves (like people slavishly follow fashions) to choose what to buy …
I suspect that version A won because of the Most Popular Products column.
It would be interesting to see how version A would do against an identical page with the Most Popular Products on the left side of the page.
Heatmaps usually show the left side of a page to be hotter real estate than the right side.
Heck – I find much more counterintuitive stuff all the time – if you think you can guess the result, particularly on multi-variate, get ready to be proved wrong (quite a lot of the time). If your KPI was bounce rate, this test result would be reversed. If the goal is conversions, that’s why you create test beliefs that you then try out – you are trying to get more conversions! Some of the time your hypotheses will be correct when running tests and this is where the insight lies. It isn’t anything to do with whether ‘simple is better’. In most cases, removing clutter and distractions helps the goals I work on but not always. There is no rule, there is testing.
I personally strongly preferred Version A without expanding the page as B was too flat. I agree with those above about the visual balance produced by the favorites column and the ability to see other products at a glance. That being said, I’d never take my intuition – test that ROI!
I wouldn’t call A cluttered, but perhaps that word was chosen to lure us in . . . . I’m just saying, Anne and her team seem pretty sharp. I fell for it.
I agree with the top product list being a big contributor, but Version A also has things no one has pointed out yet, such as color swatches immediately visible and a product count. I believe there was a study (by Gap.com I think)regarding color swatches a couple years back and how impactful they are to sales. Granted, the customer might be able to get that info after clicking into a product, but if one can avoid extra clicking and get more info at a glance, the experience is likely better. No?
Scott and others are spot on. ‘Most popular’ is a common sorting order (see: Amazon). It adds authority and helps overcome anxiety.
I think the idea that simple pages work best is definitely a dying one in some respects. We now micro manage our time by the second, we can loose interest in a web page in half a second if it looks like the information we need is more than two clicks away..there’s so many other sites to choose from, we move on.
I think page A won because there were more products, so higher chance of something catching viewers interests: Having ‘Most Popular’ means you get to see the best products at a glance, this is a huge timesaver in internet seconds.
@Scott – you are right about color swatches. We had one more variation that was different layout, but also had color swatches. It was performing better as well, but not as good as the winner. I would say that “Most Popular” gave additional winning edge.
Definitely like the simplicity of B better and makes sense for the bounce rate but the Most Popular is a must especially for these types of concumers. I wonder if having a better initial navigation (Girls, Guys, Most Popular) with the landing page being Most Popular would have been a better option? This would have eliminated the clutter problem and showcased the most popular which increased sales.
All good comments and agree on the subjectivity of “clutter”… There is organization design that gives hierarchy to the information allowing a quick scan to select or deselect input.
Definitely the “Popular” social aspect is why A won. We humans have a fear of making choices that others would ridicule.
Aside from the “popularity” aspect of the Version A column, if something on the main page didn’t quite meet the needs of the buyer, perhaps they were able to find something in the right-side column that did meet what they were looking for. It’s about giving customers choices.