Email marketing statistics traps
Collecting and analyzing metrics is very important to generate whatever statistics. The question is what exact metrics you should analyze? Here is the sample list of metrics you may need, but remember, that your own list of metrics should depend on your own objectives.
- Average dollars per email sent or delivered
- Average order size
- Bounce rate
- Click to open rate (# of unique clicks/# of unique opens)
- Click-through rate
- Conversion rate (number of actions/unique click throughs)
- Delivery rate (emails sent – bounces)
- Net subscribers (# subscribers + new subscribers) – (bounces + unsubscribes)
- Number of or percent spam complaints
- Number of orders, transactions, downloads or actions
- Open rate
- Percent orders, transactions, downloads or actions of emails sent or delivered
- Percent unique clicks on a specific recurring link(s)
- Referral rate (“send-to-a-friend”)
- Subscriber retention (# subscribers – bounces – unsubscribes/# subscribers)
- Total revenue
- Unsubscribe rate
- Web site actions (number of visits to a specific Web page or pages)
Besides your internal metrics you can also use external metrics, such as “industry averages”. These data can show you where there are any problems with your campaign. But you should remember that industry averages is not an absolute, this is just the numbers you should tend to.
The results you get from tests should be examined carefully as well. Sometimes you can get more information that is less impressive or that provide you with more concrete facts if you try other ways of scanning the metrics. A good idea is to check “filter” reports. For example you could verify the results by top domains: if results differ too much, that might be the problem with delivery issues, including the use of specific spam filters. Following the HTML to text-based versions of your e-messages ratio may lead you to a better determining the open rate statistics. Use of demographic “filter” will definitely help you improve your subscriber-oriented targeting policy. You can also use a lot of such “filters” to make your statistics more useful and comprehensive.
Collecting different kind of data you may ask yourself what number of message recipients should be involved in your investigation to get true results: should it be the whole emailing list or just part of it. The practice shows that to get approximate calculations it is enough to get 30-50 responses, but 100 responses will provide you with almost 95% accuracy count. To determine the size of list you’ll need to make preliminary calculation depends on your CTR average. Presupposing you obtain a list of 10,000 emails. Your average CTR is 8%. Think about how many responses you need (let’s say 100). With an 8% click-through rate, 1300 emails sent would result in 104 responses. If your click through rate is higher, the number of emails you need to send to test would be lower.
One more method of improving final statistics and a better aiming the campaign is to test variables, which can be design, format, subject line content, specific time/ day of the week, etc. The better way of applying this method is to use the technique of “A/B splits” (see the terms) and after comparing the gathered results make necessary changes that will forward you toward your end goal.
Friday, August 27th, 2010 : Email Marketing : No Comments