Adventures in Email Marketing

Today I was midpoint in a sluggish afternoon meeting when something was projected on the wall that made me jerk awake and wonder: why are my delivereds so low? Why bounces so high? What did we do right? What did we do wrong?

In one way email marketing is a godsend, it’s hard and cold data, which we can measure and count to our hearts delight. Yet in another way, it’s also too much data. It’s got lots of ins and outs and variances.

Take for instance delivered (or sent?). And bounce (or soft, unless you use hard…) And then opened (well with images off that number is… interesting). Then clicks, pretty clear, if counted distinctly, as in one-persons-click-per-day. Then clicks-to-open (CTO), or maybe you like clicks-from-delivered (CTR). You get my idea. But the funny thing, is that these are the easy metrics.

How about:
- If there are high opens but low clicks, did the subject line oversell?
- If there are low opens but high clicks, did the subject line undersell?
- If the click through is low, really any content- design, offer, expectation- could be the cause
- Are the leads qualified? Are there known or unknown variances between cells and segmentation? The undiscovered segmentations that magically hit on the offer.
- Timing and contact frequency. Did you have suppression rules on the campaign, de-prioritizing it in favor of some other campaign? Did that starve your list?
- More timing- is this the third email in three days? Could you have exhausted your interested readers?
- Have webmail sites deployed some new foil – such as Yahoo recently has hidden the “view images” button from me (need to research) which I expect to see in lowered opens for newer Yahoo subscribers.

What to do?

I end up routinely asking these questions of metrics, when something on the projected wall of a meeting is odd or unusual:
- Don’t just show the percentages, show the real numbers, We know volume skews ratios, so show the volume delivered, and then the CTO or Open Rate.
- Apples to apples. Be aware of your cell and campaign segmentation before comparing the metrics. Campaigns that target more active, engaged targets will have far better metrics than less engaged, larger “blast” campaigns.
- Be aware of contact frequency of campaigns in tight time frames
- Calculate a few difference ratios, or insignificant factors, to give yourself an idea of the inaccuracy. My favorite is the non-open click, which is the percentage of customers that generally click through an email without loading images (which count as opens). This skews the CTO metric- which I like- but it’s just good to know in general how much. I also keep track of the sent vs. delivered, for a few types of campaigns (transactional & promotional). I don’t need to know for every cell or segmentation, just in general.

More reading on metrics, and their variability:
Improving Email Open Rates by the folks at MailChimp
Email marketing statistics: six misinterpretations and Seven Tips for Interpreting Your Email Marketing Reports by Mark Brownlow
Benchmarketing Email Response Metrics by Tamara Gielen
Obsessed with Open Rates? Stop it; Focus on Feedback Loop by Ken Magill
Unified Multichannel Metric by Kevin Hillstrom
Ask The Expert: How does my open rate compare to my peers? by DJ Waldow