Practical Ecommerce

Don’t Follow when Redesigning your Ecommerce Site

If you’re planning on redesigning an ecommerce site, don’t follow the experience of

The well-known retailer — known for its focus on product design and its communications — launched a new site that attracted the ire of web designers, one of whom called the redesign the “train wreck redesign of the year.”

So, what did Target change that caused such a fuss?

Well, here’s the original home page design.

Target's old home page.

Target’s old home page.

And here’s the new design.

Target's new home page.

Target’s new home page.

The navigation bar has been redesigned, including new main link labels — nothing major there.

But, the rest of the page’s layout — i.e., its wireframe — is entirely different: There are now more products on the home page, for instance, replacing the simple design with a grid of products.

There’s also the infamous shadowbox design element that seems to have most offended the critics.

But was the site redesign a disaster because the design pundits didn’t like it? What if the shadowbox effect actually helps boost Target’s conversion rate by making the different sections stand out?

Sadly, we may never know, since it’s not clear if Target split tested these new elements. That’s what makes this redesign a potential disaster.

In response to the criticism, a Target spokesman said, “User testing prior to launch, and actual site traffic and sales performance since, have been very positive. We plan additional changes to our design and features, and will track guest response and make adjustments along the way.”

Since the spokesman refers to sales performance since launch — versus A/B split testing before launch, like usability testing — we’re led to believe Target didn’t do comprehensive A/B testing of any other new elements before launch.

Yes, usability testing has its place, but it’s not a substitute for A/B testing. Traffic and overall sales performance also are not reliable gauges of the impact of the new design. There are too many other effects, such as seasonality, stock levels, and competitive activity.

The aesthetics of drop shadows aside, the biggest problem with the Target redesign is the apparent risk it took, and not just with its brand.

It made so many changes at once, it will be impossible for the company to really understand which changes helped its conversions and which hurt. Even if, as its spokesperson says, overall sales are up, some of the changes are probably keeping Target from even more sales.

What if, for example, the new search box design lifted the conversion rate by 12 percent (by making it easier to find things on the site) but the layout caused a conversion drop of 5 percent?

The net conversion rate lift with just these two changes would be 7 percent, but Target wouldn’t know which areas made which impact. And, that’s only two of the thousands of changes it made.

Effects of different website design changes

Effects of different website design changes

Learning from Target’s Mistakes

So what lessons can you learn when it’s time to redesign your ecommerce site?

Rather than making such dramatic changes all at once, and then adjusting afterward, make a series of tested changes (which can each be dramatic too). Think evolution instead of revolution. Because, let’s face it, revolutions are more dangerous than simple drop shadows, and pose a high risk of failure.

There is a less risky approach than a complete website redesign — one that maximizes your conversion-rate improvement by isolating the impact of individual changes. I call this approach “Evolutionary Site Redesign,” or ESR.

Instead of a complete change like Target did, with ESR the new design, layout, and content elements are tested continuously, changing your website design and layout toward the best-performing site it can be.

In an ESR redesign, you prioritize your most important pages, templates, and site-wide elements. Then, test new layouts, design, and content with controlled split testing. You get all the benefits of a new site design without the risks.

Since ESR uses proven conversion rate optimization principles to redesign your site, you end up with a new site look-and-feel and conversion-rate lift at the same time. With ESR, new interface methods and approaches can be tested to see if they actually improve business results.

If Target had run A/B tests on some of the changes individually, it may have found that certain changes improved conversion rates, while other hurt. Even if the shadows aren’t an improvement over the original, the new layout probably overshadows (pun intended) the decrease by showing more products clearly.

With ESR, Target could have tested the most important changes separately, and understood which changes would boost sales. That means a renewed website, without the risk.

Target could run the following isolated tests.

  • New navigation design, including logo placement, search box design, background color. I’ve worked with clients that have profited from big lifts in conversion through controlled tests in this area. But not all changes are winners.
  • Page layout wireframe. Adding more products to the home page could boost conversions if there are more entry points to drive action. Or, as the critics suggest, it could hurt conversions by making Target look down market.
  • Site-wide design elements of box shadows, typeface treatment, and colors. Design changes can have unexpected outcomes. If you make changes that affect the entire site, even if they seem like obvious improvements, it’s always better to test than assume.
New design changes on Target's home page

New design changes on Target’s home page.

Then, once the best design has been confirmed through careful testing, Target could continue to test, for example, (a) which product categories to display for each shopper segment, (b) which offers work best for various seasons, (c) how many products and categories to display for the best conversion rates, (d) the design and pricing treatment for products, and (e) the home page positioning headline and value proposition.

Yes, it may seem like an endless cycle, as it should be. But, by prioritizing your ecommerce conversion optimization tests — see “Use the PIE Method to Prioritize Ecommerce Tests,” one of my previous articles — you can systematically improve your conversion rate, and build better returns at each step.

Prefer Dramatic Redesign Changes?

Taking the Evolutionary Site Redesign approach doesn’t mean you can’t make bold changes — even as bold and dramatic as some of Target’s changes. It does mean implementing these changes in a controlled, measured way, making sure that their impact is tested and proven to be beneficial to your conversion rate.

Even if you’re looking at refining or redefining your brand and the entire look-and-feel of your website, ESR gives you the confidence that the changes are for the better and have a proven positive effect on your business.

Even if your redesign doesn’t win the hearts and minds of all the web designers, by taking an ESR approach you will win where it counts: with higher conversion rates and profit from your visitors.

After all, your customers are the most important audience.

Chris Goward

Chris Goward

Bio   •   RSS Feed


Sign up for our email newsletter

  1. Pete Harrington February 7, 2014 Reply

    I guess you could plot a control chart of sales or rather conversion rates for a year and the see if the new changes really did any harm or good.

    Without proper statistical analysis, your examination of their changes is as much guess work as you assume their testing process was. Bottom line, only the data would tell the story. Designers getting annoyed is nothing new.

  2. Joe Buchmann February 7, 2014 Reply

    I agree with Pete.

    It looks like Target is doing some testing after all. I’m getting the new content layout served with the old header. Some of you are getting the full ‘new’ experience. It appears there are at least a few variations out there being served up to consumers.

    I think the critics were a little to quick to jump onto the hater bandwagon. After all, Target is an extremely major retailer. I’m inclined to believe that they aren’t making decisions lightly, or disregarding analytical data. I’m sure we’ll see their site change around a bit in the near future until they settle on the layout that works best for them.

  3. Mike Belasco February 8, 2014 Reply

    Looks like they are testing with site spect

  4. Chris Goward February 10, 2014 Reply

    I agree too, Pete. Yes, there’s conjecture here about whether they tested the new site before launching it.

    From what we can tell, they did a flip-the-switch overhaul redesign and are now iterating backwards to find out whether some of the changes they made (like the garish dropshadows) caused more harm than good.

    The shadows I’m seeing on their site now are much more subtle than when I wrote this article a few weeks ago.

    That approach can work and I’m glad to see them testing now! My point was and is still that, if they didn’t test to achieve their new design, they probably could have saved themselves some grief by A/B testing their way to the new design.

    Again, there’s a large portion of conjecture since we don’t have insider knowledge of their approach. But, seeing that their current testing seems to be reversing some of the initial design decisions, it’s even more valid.

  5. Ian Hough February 10, 2014 Reply

    What part of “User testing prior to launch” don’t you understand?

    The drop shadow on the empty box wasn’t as strange as the empty box itself, in my opinion, on a homepage that sees some serious daily traffic and should be optimizing all available real estate (though not necessarily by displaying tens or hundreds of different goods).

    Does anyone know how deep the redesign went? As in, was just the homepage affected, or were internal pages subject to the same makeover and, if so, did this also result in a lot more goods being displayed on each page?

    I love the PIE testing article, by the way.

    • Chris Goward February 10, 2014 Reply

      Ian, I agree with you about the empty box. That is very odd.

      I’m not sure what you mean when you say ‘What part of “User testing prior to launch” don’t you understand?’

      User testing is not A/B testing and is not a reliable indicator of success. There are 7 reasons to not use qualitative user testing alone as actionable data, which I outline in some detail here:

      • Ian Hough February 12, 2014 Reply

        Hi Chris, thanks for that. I read that piece a few weeks ago and liked it then. I was researching the value of focus groups at the time, and would love to hear how you rate them for UX testing websites.

        You’re right, I confused user testing with A/B testing. But you still think it has *some* value, right?

        • Chris Goward February 12, 2014

          Yes, they do have some value. We use user testing to generate hypotheses for A/B tests. Generally, they’re best for identifying Clarity problems within predefined scenarios, but they don’t reveal creative new solutions or persuasional insights.

  6. ozEworks February 11, 2014 Reply

    You are attacking the live site so it is immaterial whether they A/B tested or not. This is the design they chose and they indicate it is working for them.

    What designers think about a design versus how the design converts are often two different things. A site is not a picture, it is a tool. If it converts more customers than the prior design then they have made a step forward.

    IMO, Target has worked on its search and you can now find products more easily. Search is not “design” in your sense but functional changes are often more important than change to page layouts.

    And, by the way, page layout is not the same term as wireframe. Wireframes are a tool to depict a page layout. There is more than one tool you can use to do that.

    • Chris Goward February 11, 2014 Reply

      @ozEworks If you read the article, you’ll realize you’re vehemently agreeing with me about what designers think. Yes, if it converts more, who cares what anyone thinks. That’s the point of the article! (Read the article again)

      However, I disagree that “it is immaterial whether they A/B tested or not.” If they haven’t A/B tested the new design, they won’t know and can’t reliably “indicate it is working for them.”

      To repeat, neither I nor you know whether they tested the new design before launch. In fact, they may have, and there’s evidence that they now are testing it, which is awesome. It may be producing a conversion rate lift right now!

      And, to repeat another point, I’m not attacking the site per se; I personally think the new design is an improvement in several ways. I’m critiquing the flip-the-switch redesign *process* that it seemed they used. If they didn’t do it that old way, I’m happy to be shown incorrect on that point. The article is about the process, not this particular website.

      The purpose of the article is to use a recent example of an epic redesign to show a better way using evidence-based Evolutionary Site Redesign.