While marketers tend to measure their e-mail programs using such basic metrics as clicks, opens, and sometimes even conversions, one expert says most of them have got it all wrong.
“There are metrics that people should be looking at that they’re completely ignoring,” said Andrew Robinson, director of international services for marketing services provider Lyris. “The real screamer is short visits to a Web site from an e-mail campaign. This is one we look at all the time that people in the industry aren’t looking at. Once people click through, if they spend less than two seconds on the site and then move away, then that’s a short visit.”
Robinson added: “We’ve seen e-mail campaigns that have great open and click-through rates but 60% short-visit rates. This means the e-mail’s doing its job but what’s on the Web site is completely wrong and is not meeting expectations that have been set up by the e-mail.”
Robinson said this phenomenon tends to happen at large companies where there is a disconnect between the people crafting the e-mails and the people managing the Web site and landing pages.
Robinson added that companies should set up their Web analytics programs so they can measure their e-mail marketing program’s performance based on business goals.
“You want to be testing for particular goals,” he said. “That means people complete certain actions such as a purchase, or download of a whitepaper or download of an application or whatever goals you have. You need those goals reported on your e-mail campaign.”
One problem is that at many firms, e-mail marketing and Web analytics are considered two separate entities.
Also, he said: “People don’t spend a lot of time on Web analytics because it’s so difficult to wrap your head around it.”
Part of the problem is also who tends to be in charge of marketing, Robinson added.
“There are too many English majors in marketing and not enough math majors,” he said. “You need to understand how to make statistically legitimate insights from masses of data to get anything useful from Web analytics. The real problem with Web analytics is you’ve got a load of amateurs drawing conclusions from reports that they really don’t understand the validity of.”
Robinson’s recommendation: “If you’ve got the cash, hire a statistician. If you haven’t, get some consulting from a statistician. Get a statistician in for three to five days and really hammer out what is going to be useful to track. Set up reports to track that, understand the methodology behind it and understand the levels of variance and potential misleading results you can get so you’re not over-claiming victory when you shouldn’t be and you’re not over-worried when you shouldn’t be.”
