October 02, 2006

Does Homepage Design Influence Net Sales?

There is a creative tension that occurs in homepage design. Businesses must constantly balance the need to sell merchandise with the desire to communicate emotional aspects of the brand. Both issues are important. In some companies, the concept of "selling" wins. In other companies, the concept of "branding" wins. In either case, driving profitable sales should be the end result.

Which style drives more business, selling, branding, or both?

I analyzed data from the top 300 businesses in the Internet Retailer Top 500. I wanted to make sure I analyzed businesses that had at least $15,000,000 in annual sales.

Next, I created three segments, segments that loosely describe the selling technique employed by the website.
  • Selling Websites: These websites have a strong focus on featuring a lot of products on the homepage. Frequently, the customer has to scroll down the page to see all product. Many individual items are featured on the homepage, and numerous links make most of the website available to the customer. Examples of "selling" websites include Super Warehouse, Abt Electronics, and Drugstore.com.
  • Hybrid Websites: Websites that balance selling and branding are classified as "hybrid" websites. These sites typically include the entire homepage display on your monitor, without need to scroll down. There is ample white space or empty space, so that the customer is focused on a specific area of the website. Examples of hybrid websites include Eddie Bauer, Crate and Barrel, and Collections Etc.
  • Branding Websites: These sites embrace the customer by evoking emotion. The homepage is not about directly selling merchandise. Instead, the website creates a mood, a feeling, something that differentiates the website from all the others. Examples of "branding" websites include Patagonia, Coach and Tiffany.
After placing each website into one of the three segments, I further segmented websites into those in the top 125, and those from 126-300 (ranked by sales). These six segments (selling, hybrid and branding by big sites and smaller sites) were analyzed for conversion rate, average order size, and net sales per visitor.

First, let's review the results among the 125 top selling websites.
  • Conversion Rate:
    • Selling Sites = 4.55%
    • Hybrid Sites = 6.08%
    • Branding Sites = 5.16%
  • Average Order Size:
    • Selling Sites = $219
    • Hybrid Sites = $204
    • Branding Sites = $125
  • Net Sales per Visitor:
    • Selling Sites = $9.96
    • Hybrid Sites = $12.40
    • Branding Sites = $6.45
Next, we observe the results for websites 126 to 300.
  • Conversion Rate:
    • Selling Sites = 2.86%
    • Hybrid Sites = 3.56%
    • Branding Sites = 2.34%
  • Average Order Size:
    • Selling Sites = $178
    • Hybrid Sites = $187
    • Branding Sites = $120
  • Net Sales per Visitor:
    • Selling Sites = $5.09
    • Hybrid Sites = $6.66
    • Branding Sites = $2.81
In the spirit of full disclosure, net sales per visitor is arrived at by multiplying average conversion rates by the total average order size. This is a calculation. A different, though directionally similar result, occurs when taking the total average of net sales per visitor.

What Does The Analysis Tell Us?

The analysis clearly indicates that companies employing a "hybrid" strategy have the most productive metrics. In fact, hybrid sites are about twice as productive as branding sites. Branding sites tend to have low average order sizes. Further study needs to be done to understand if the small average order size is due to the merchandise assortment on those sites, or are due to a lack of cross-shopping done by customers due to site usability issues.

It is interesting that "selling" sites have lower conversion rates than hybrid sites. It may be possible that customers are overwhelmed by massive amount of content on selling sites, causing customers to leave before they begin shopping.

It should also be noted that there are numerous examples of each strategy (selling, hybrid and branding) working exceptionally well. There are numerous examples of each strategy delivering poor results. In other words, there are ways to make the branding strategy effective. The data suggest the hybrid strategy is simply more likely to work.

Lastly, I defend the branding sites by saying that most of the big-ticket sites, those selling computers, participate in a selling or hybrid strategy, driving up the average order size. Still, when viewing conversion rates, branding sites lag behind hybrid sites. Conversion rates at selling sites are frequently hurt by large-ticket items like computers.

Given all of the caveats, it is still clear that the design of the homepage influences the productivity of each visitor to a website. It appears likely that a hybrid strategy is most likely to maximize the net sales of each visitor to the website. Selling sites may overwhelm visitors, while branding sites may not present enough merchandise to entice consumers.

How does your website stack up, against these metrics?

8 comments:

  1. Anonymous7:55 AM

    Thanks for your comments, I'm glad you enjoyed the discussion.

    As you point out, there are many different ways users arrive at a site. On my site, I have two very different audiences. One arrives at specific permalinks via RSS/Atom feeds, the other audience bookmarks the homepage, and visits via their browser. E-Commerce businesses have the same dynamics. In a perfect world, how the homepage is merchandised and creatively executed should be dependent upon the audience visiting the page.

    ReplyDelete
  2. Anonymous2:59 AM

    Hi Kevin --

    Two observations:

    [1] I believe the IR conversion data are self-reported and not audited. I'm familar with a few cases where retailers are on the list due to their own inflation of numbers. Now, to change your result, one would have to argue these 'errors' aren't random -- that selling / branding / hybrid had as a group different proclivities to report to IR acurately. Don't know if that is the case, but I take IR data with a big grain of salt.

    [2] Your results depend critically on how you assign retailers to the bins, which could be subjective. I'd give additional hard criteria (eg percentage of above-the-fold homepage pixels devote to branding messages and images, vs. percentage of pixels devote to selling messages) on how you made the B/H/S decision, as well as provide your three lists.

    Neat stuff, pulling observations from very thin potentially very noisy data.

    Alan
    rkgblog

    ReplyDelete
  3. Alan, you couldn't be more correct --- the data is frequently self-reported and errant. I made the assumption that everybody over-states their results at approximately the same rate, regardless of selling/hybrid/branding strategy.

    On the second point, I agree with you, I should use facts that are more quantifiable than just my opinion. Doing that work would be worthy of writing an entire book about the topic ... hmmmm!

    ReplyDelete
  4. Your post is pretty illuminating and practical,I completely agree with the above comment. Thanks for the sharing.Will you share more about it? I just want to know more.You are very talented.I guess, this site would be my farorite one. Thank you again!!!

    ReplyDelete
  5. WOW.. Its a great information ... great article... keep working on it.. thanks for it.

    ReplyDelete
  6. I should use facts that are more quantifiable than just my opinion. Doing that work would be worthy of writing an entire book about the topic.

    ReplyDelete
  7. There is a creative tension that occurs in homepage design. Businesses must constantly balance the need to sell merchandise with the desire to communicate emotional aspects of the brand.
    Revitaderm Cream
    Perfect Radiance

    ReplyDelete
  8. The contest is free and open for everyone.

    ReplyDelete

Note: Only a member of this blog may post a comment.

Summer Schedule / Document Your Progress This Summer, Ok?

Most summers, I cut back on how often I publish. This summer will be no different ... and you may be subjected to a few videos instead of bl...