January 17, 2010

Dear Catalog CEOs: Improved Conversion

Dear Catalog CEOs:

A quick read through the marketing literature suggests that there are few things that are more important in the online world than improving conversion rate.

I will concede this. It is important to improve conversion rate among first time visitors, or among visitors who have never purchased before.

Now, let's think about our existing customers, those who have previously purchased from our business. Does conversion rate matter?

If you look at simple web analytics metrics, you'd be inclined to think that conversion rate means everything. If the customer doesn't convert, you lose the sale.

If you look at customer metrics, measured over the course of a year, you'll see a completely different story.

In the past decade, conversion rates have declined. And yet, on average, existing customers repurchase at the same annual rate, ordering a similar number of times per year, buying a similar number of items per order, and paying a similar price per item.

In the past few years, marketing experts have done outstanding work developing tools and techniques that greatly improve conversion --- we read about the techniques every week in trade journals. And yet, on average, existing customers repurchase at the same annual rate, ordering a similar number of times per year, buying a similar number of items per order, and paying a similar price per item.

Among existing customers, conversion rate is seldom an optimal metric. Consider this situation:
  • In 2008, 1,000 existing customers visited your website an average of twelve times per year. 300 existing customers purchased, purchasing two times each. Total purchases = 300 * 2 = 600. Total visits = 1,000 * 12 = 12,000. Conversion Rate = 5.0%.

Now, in 2009, you create a marketing program where you have a daily web special, one where customers visit to learn about discounts on various popular items. Your website metrics change:

  • In 2009, 1,000 existing customers visited your website an average of thirty times per year. 320 existing customers purchased, purchasing two times each. Total purchases = 320 * 2 = 640. Total visits = 1,000 * 30 = 30,000. Conversion Rate = 2.1%.

As a merchant/marketer, you have done everything right. You were successful, getting 7% more existing customers to purchase than last year.

According to conversion rate metrics, you are a failure. A thousand different individuals offer solutions to solve your conversion issues.

I want my best customers to visit my website every single day. I don't care if they convert or not, so long as on an annual basis, retention rates increase, orders per buyer increases, items per order increases, and price per item purchased increases.

If you need help analyzing whether conversion rate changes are truly an issue, please contact me.

4 comments:

  1. Kevin, the math changes when you factor in the princely sums companies pay to drive visitors (back) to the site.

    Existing customers *still* search on Google and Bing to come back to the site. Investing in optimizing the conversion rate increases the ROI of that spend.

    ReplyDelete
  2. Sure, that does happen, thank you!

    ReplyDelete
  3. Does this mean I get a point on this round? :)

    ReplyDelete
  4. Great post, Kevin. I've long been frustrated by conversion rate as a metric for websites because all too often it goes in the opposite direction of revenue. One of the reasons for this split is the fact that we use our sites for purposes well beyond the buy button, but conversion rate as we measure it today only measure one action (purchase today) against all of those purposes. As a result, conversion rate is not a very good proxy for sales and we shouldn't treat it as such. I wrote about this in a post on my blog a while ago at http://www.retailshakennotstirred.com/retail-shaken-not-stirred/2009/07/true-conversion-the-onbase-percentage-of-web-analytics.html.

    I also agree with your point that the end goal is more sales. I also see Raquel's point, but there it's about specifically measuring the conversion of the visitors from that campaign rather than measuring an overall conversion rate.

    ReplyDelete