Open rates: a review of statistics in email marketing.

Many people talk about “open rates” in email marketing – implying that they are able to judge the number of people who have actually seen an email, and that there is a clearly understood definition of what the word “open” actually means.

HHM is not an expert in this field. However we do send out a lot of emails both through a series of news services that we run for ourselves and some of our clients, and through general lists of emails, so it is important for us to understand what “open rates” actually means in terms of email marketing.

Our research has revealed several things – including some information on what the crooks are doing with emails. That might seem a bit off topic, and I’ll leave that bit until last but it does give a clue as to where the market might go.

In the meanwhile, here’s what we found out about “open” rates.

Depending on the software you use, open rates can mean any of these things:

• The number of emails actually received by people you emailed (even if the email went straight to a delete box through the use of some “message rules”). This will be the number sent minus those that never arrived because the mail box was full, or the system down, or the address discontinued.
  
  
• The number of emails clicked on by readers – but this clicking could mean clicking to delete the item as much as clicking to read it further.
  
  
• The number of emails where the recipient has clicked to reveal blocked images that are on a website that links to the email. The problem here is that our trials and research shows that when we put in such a link many people read the email without clicking, and others just delete it at once – thus the stats can be quite misleading. We feel the latter option (delete at once when the “reveal pics” message is shown) is growing, because of the increased awareness of the sophistication of cyber crooks (see below).

So the situation is unclear, and it is made worse by the fact that some software actually seems to record open rates twice for some emails. For example, an email received might be clicked on once to open it up, but then if one clicked again to reveal the images or to go onto a site linked from the email, that produced a second “open” statistic. So the number of “opens” would include some emails where one recipient counted twice.

It’s all a horrible mess – made worse by the way Outlook Express works. Most people using OE can look at the content of their incoming email without clicking or doing anything else. Typically one window shows the sender and subject line and another shows the email itself. The recipient can read, and then delete. Is this “open” or not? No one can agree, no one knows.

This does not mean that open rates can’t be used at all – but it does mean that you can only compare open rates measured through the same software.   So if Hamilton House says to you, “Your open rate was 25%” and another company says, “we got 35% open rate for you” those two figures can’t be compared, because almost certainly our open rate is measuring something different from their open rate.

For this reason we have now developed a chart which shows average and best open rates and click through rates measured through our system – so that you can get an idea of how your results compares with others using the same software.

The one statistic that does have a consistent meaning is “click through” – obviously defined as the action of clicking on a link within the email to a website. The software can measure this clearly and straightforwardly and always gives us consistent results.

Indeed the fact that this statistic always works throws further doubts on all the other stats. If we can always measure “click through” and can verify the measurement through a set of tests, why do the “open” figures vary so much?

This raises another issue: what is the point of “open rates” at all? If I “open” your email, read a line, think it is rubbish (no offence) and press delete, and you count me as “opening” – so what? I haven’t read much, I haven’t clicked on the link…I don’t like what I read. What have you learned from this statistic?

I can see the point at one level – if 80% of people you sent the email to “opened” it but only 1% clicked on the link, we could say that your subject line was obviously good, but you lost it all in the message. But to do this I would need to be sure that “opening” was a real choice, and wasn’t incorporating within the count messages that were simply delivered.

In reality, because emailing is cheap, it is possible to sort this out in a much more positive and clear way by undertaking a set of experiments based around either

• the number of enquiries or sales you get or
• the number of people who click on your link.

To turn to the other issue that I raised, SC Magazine has published an interesting piece (full details below) about emails from cybercriminals. They point out this scenario…

The user clicks on the link and is directed to a site where a Trojan file automatically downloads onto the user's computer.

The trojan then downloads an additional spyware file that captures sensitive information, such as bank account numbers (known as spy-phishing).

http://www.scmagazineuk.com/Cybercriminals-move-with-the-times/article/112273/

Now what strikes me is that by and large people are getting more and more cautious about email. They will look at emails (and using the Outlook Express example above they will do this without actually clicking on anything), and then decide if they want to read on. 

At the moment I believe most professionals feel that they can recognise a criminal email when it pops up and hits them between the eyes because it is so obvious – the Nigerian scam, the sex pills, the tax refund, Britney Spears pictures, the bank website update (please re-enter you account details and password)…

What we all know is that the writers are sophisticated in their IT skills, but we rely on their lack of other skills – such as knowledge of grammar, an ability to spell... 

But supposing the criminals start to get more sophisticated in terms of their ability to copy legitimate sites and in their use of the English language?

At that point fewer people would be willing to click on a link even if it looked perfectly reasonable and related to their profession. In which case the only way to sell by email will be to give information and offer not only the link to the online shop but also a phone number, a fax number and an address.

In fact, most serious players in the market do this, but a fair number of smaller firms don’t supply such background detail at all, and this could ultimately be their undoing.

In a sense I am predicting a movement in email marketing which moves away from all the high tech “click here” approach of the last few years, and back to treating the email as if it were a bit of direct mail, which is forced to rely on words to excite the readers. The links can still be there for readers willing to take the risk, but over time I suspect fewer will be willing to give a website a try, no matter how well the email is presented.