Tag: ecommerce

Abandoning your conversion rate worries

When you’ve worked on e-commerce sites for a while, you get to respect and fear the words “conversion rate”. Particularly during a peak trading period (er, for travel, that’s January!)

Well, this year I’m throwing away my fears and worries. Here’s why:

1. Today’s e-commerce sites are only a teeny tiny bit about “conversion”

In 2004 we did a survey of >10,000 people to the First Choice holidays website. Of the people who didn’t book (unfortunately this was a large proportion…)

  • About 70% said that they came to the site to “get prices and availability”
  • Another 20% said they were “just browsing”
  • 5% said they had already booked and were looking for info about that booking
  • A small percentage said they weren’t looking for a holiday at all

So out of the 10,000 respondents visiting our site that month in 2005 there was only a very limited proportion who were going to book on that day anyway. This is what Avinash calls your “conversion opportunity pie“.

If people have no intention of buying it’s pretty unlikely you’ll turn them around, particularly on a high ticket item like a holiday, which requires other people to “buy in” before they can transact. Analytics and conversion rates will not tell you about visitor intent.

So for all of these people (and like I mentioned, it’s a very high proportion), no point about obsessing on conversion rate unless you’re measuring precisely who came to buy on that day.

2. Analytics data doesn’t tell you where your website is really failing

When you look at your analytics data for a purchase funnel, there is no way that is going to tell you how to fix what is fundamentally wrong with your website. Worse drop out rates on a page are not usually related to that page’s design, content or functionality at all in my experience.

Usually the answer will lie in the structure of the experience you’re providing for people. They will step out of the funnel to find other information that is missing within it. Sometimes that will lead them to (gasp!) other websites. (Websites with ads for similar products from competitors, too!)

The problem with funnels is that they won’t tell you what information people were looking for that was missing, or what the customer did next on the external website. So the funnel you’re using won’t be tracking the real customer journey because half of it won’t be on your website anyway.

And even if your website does have all the information needed, you can bet that the funnel you’re using to analyse behaviour is massively over-simplified for what is actually happening.

All of that adds up to: no need to worry about conversion funnels, either.

3. Conversion rate is not a measure of satisfaction

Scenario: conversion rates go up on one day and then down the next. How do you understand why?

Perhaps you’re shifting some dead stock so there’s some amazing deals on, or you’ve got an exclusive discount code. Or you had a great ad in the newspaper that day about some promotion you’re running.

Each of those tactics can provide people with the momentum they need to avoid all the design obstacles you have put in the way of buying from you.

Conversion rates will not tell you how well you did at serving each customer’s needs, and conversion rates are telling you nothing about what to do next about your website. Great!

Another challenge for 2011: can we move away from obsessing on conversion rate as a KPI and move towards customer satisfaction and other more qualitative measures? If satisfaction measures are positives (and your business model is correct) then you can bet the business numbers will look good too.