Open rates and CTR: think before you act

Latest posts | Feed | | By Mark Brownlow

Thanks to the power of electronics, we get a lot of data back everytime we send out an email campaign. Clicks, opens etc.

Much of this data gets dumped in the alleyway of neglect. More enlightened marketers, however, put it to good use to track the results of their efforts and learn about what works and what doesn't.

A big problem in this process is the poor attention given to understanding the cause of a particular result. If success is attributed to the wrong cause, then the future refinements you make to your emails might actually do more harm than good.

Let me explain with a couple of examples:

1. Open rates example

Imagine you email your list with two types of email. The first contains a coupon code for 10% off all items in a specific product group. The second contains links to specific items in that group priced at an even bigger discount.

The first generates average open rates of 45%. The second average open rates of around 20%. Clearly your list are more interested in a wider coupon than specific discounts on a more limited selection.

Now imagine the coupon code in the first email is always an image. And the product links in the second email are text. Now imagine that your recipients actually have no preference between a wider coupon and specific discounts.

What's really happening is that a large number of recipients using clients and webmail services that block images are activating image downloads so they can see the coupon. Something they don't have to do with the second email.

Because opens are measured when a tracking image is downloaded (see here for details), what you get is an entirely artificial boost to open rates for the first kind of email.

In this case, it would be wrong to conclude greater interest in a coupon simply because the open rate is higher. You would need to look at the profits generated by each type of email to make a realistic comparison.

A similar theory can be applied to open rates garnered by image-rich emails. Are the open rates high because of recipient interest or because you force people to download images to see the contents of the email?

2. Clickthrough rates example

Imagine those two emails again. This time, the coupon email has a text link to a coupon page at your website: "Click here to get your 10% discount coupon for shoes."

The alternative email pushing specific discounted shoes has a link that says: "Click here to see which shoes qualify for a 50% discount."

Let's assume again that recipients actually have no preference for either approach. The sales generated would be the same for each. Yet the second email will undoubtedly get more clicks.

The coupon link has all the information a customer needs to decide if its worth taking the next step. The discount link does not: recipients need to follow the link to get the information needed to ae that decision (exactly what shoes are discounted?)

The difference in the number of clicks reflects the disparity in information provided, not a true difference in interest between the two types of offers.

Clickers on the coupon link are pre-qualified. If you looked at the number of clicks without considering the quality of each and the subsequent actions taken, you would, again, draw a false conclusion.

What do these kinds of interpretation problems tell us? To make the right decisions based on campaign reports, we need to:

Understand what each metric really means and how it is measured
Understand the various factors that might contribute to a change in that metric
Understand the need to look at other metrics associated with a particular email
Avoid basing decisions on one metric in isolation, unless you're sure of its importance and the causes of any changes in that metric

[This post brought to you by Campaigner Email Marketing]
Permalink | March 07, 2008 | 2 comment(s)
Get posts like this: as an RSS feed | biweekly email | via Twitter

2 Comments:

Mark, another thing to consider with click-throughs is that open rate can greatly affect the click-through rate if you are measuring clicks / delivered.

In order to compare apples to apples you need to look at clicks / opens, something many people neglect.
By Adam @ Bronto, on 12 March, 2008

Thanks Adam, yes, that's a critical point.
By Mark Brownlow - Email Marketing Reports, on 13 March, 2008

Comments closed during migration to a new blog platform in early May