An opt-in email list is precious commodity for marketers. By virtue of having explicitly opted in, recipients of a digitally-delivered newsletter are among the likeliest group to fulfill a goal on your website, and emails sent to subscribers typically result in a high number of conversions. Yet fewer companies than one would expect undertake efforts to improve their conversion rates from email.
Newsletters, though, lend themselves to the collection of metrics by which you can boost conversion rates. Whether you’re sending out emails monthly, weekly or even daily (there are situations where this is okay – really!) the relentless regularity of mail delivery allows you to make incremental improvements that make a huge difference over the course of time.
This regularity also makes it possible to conduct tests even when you have very little technical support at your disposal. Even in the absence of robust A/B tests, you can swap out and measure the performance of different elements over time, in what amounts to testing in slow motion. Yes, the results won’t be as statistically reliable as a true split test and yes, this technique doesn’t lend itself to testing specific offers, but by collecting and analyzing data you’ll definitely be able to make changes to newsletters that result in a higher number of conversions (though, as I’ll discuss, setting up a split test might be easier than you think).
Whether you’re simply going to collect data for analysis or embark on full-blown testing, the first step is tagging your email-embedded links so you’re able to see email metrics in your analytics reports.
Tagging Email Links for Google Analytics
For those using Google Analytics, Google offers a simple way of tagging links by appending parameters to site URLs. Most email delivery services can do this automatically, but you can add these parameters to any link using Google’s URL Builder.
The URL Builder itself lists the different fields and what they’re used for, as well as an additional help article, so I won’t try to replicate the general advice Google gives about specific fields. For tagging emails specifically, take note of the following.
- Campaign Source (utm_source) This should be used to specify the specific newsletter you’re tagging. Use different labels for different newsletters, but use the same base text when tagging newsletters by date so you’re able to group these together later in Google Analytics (biglist-2011-08, biglist-2011-09). When split testing use the same principle (biglist-2011-08-a, biglist 2008-b).
- Campaign Medium (utm_medium) Unless you have a valid reason for high-level segmentation of emails, use “email” for this field.
- Campaign Content (utm_content) As Google suggests, use this to differentiate links that point to same URL. This can be used to distinguish links that appear two or more times in an email (topofpage, footer) or elements that you’re split testing (bluebutton, greenbutton).
- Campaign Name (utm_campaign) The name of the campaign. If you’re running a promotion that involves traffic from other sources, use the same campaign name for all traffic sources so you’ll be able to look at overall campaign performance in your analytics.
The graphic above shows typical tagging applied to a link in the email, and a campaign report in Google Analytics displaying the medium and ad content fields generated by this tagging. In this case the optional “campaign content” field (displayed as “ad content” in reports) has been assigned the value “submit” for a call-to-action button of the same name. An alternate button going to the same location with the call-to-action “Tell me more!” has been assigned the value “tellmemore”. In this way the performance of the two buttons can be compared, either in an A/B test or over the course of time as they are used in different emails.
Be sure to tag all the links in your email. If you fail to do, some traffic and subsequent conversions won’t be attributed to the originating email.
With properly tagged emails, you’ll be able to start taking action on the analytics you observe. As a bonus, you’ll see a reduction in the number of direct visits and those ascribed to web-based email sites, providing you with a clearer picture of your overall sources of traffic.
Improving Email Conversion Rates Without Split Testing
Just by having your email links properly tagged, you’ll be able to make improvements to subsequent emails without testing.
For example, let’s say you have a regular newsletter that lists last-minute cruise specials and you’d like to know what the most effective wording is for the call-to-action text that links to a page of cruise specials on your site. Without split testing, and even if you don’t specifically tag the call-to-action link with utm_content button (though that will make your life easier), you can still observe the performance of different text over time.
Again, you’d have better statistics from A/B testing, and it would be nice to use the metric of emails opened rather than emails sent, but what I’m trying to illustrate is how you can make informed decisions with relatively little data. For my money the call-to-action “More Last-Minute Specials” is a winner. And of course if you have a similar call-to-action on your website, but can’t or won’t test it, you can carry over what you’ve learned from email to your website.
Once you start properly tagging your email links, you’ll be able to judge the relative effectiveness of different emails and the different links embedded in them.
- Poor-performing, highly-placed links Are there product links or other offers near the top of your email that are seeing few clicks or conversions? Consider modifying, replacing or removing these.
- High-performing, lowly-placed links Are there links near the bottom of your email that are performing well? Consider moving these nearer to the top for even better exposure.
- Engagement Which email topics or types of promotions are most effective in drawing users onto your site for multi-page visits? Which links are generating the highest bounces – is an email call-to-action making promises that the linked landing page can’t keep?
Improving Email Conversion Rates With Simple Split Testing
If you have the ability to split your newsletter recipient list into two (which is normally fairly straightforward), you have the ability to split test email. By using the utm_source parameter to differentiate the different email version you’ll be able to compare the performance of the those versions.
In the report below, for example, one can readily see the difference in the overall conversion rate for two versions of a newsletter that link to a lead generation site.
For ecommerce sites, properly tagged email links and a properly set-up conversion funnel allow you to judge the overall value of a email, rather than only its conversion rate. This is useful in testing different offers, where the offer that results in the highest conversion rate may not actually be the one that generates the most revenue.
This simple framework allows you to test the relative performance of different email versions against almost any site metric: bounce rate, time on site, conversion rate, average order value, pages per visit and so on. Because tagging email links only allows the collection of site metrics, this sort of split testing does not directly facilitate testing email subject lines to see which are most effective in getting users to open an email. However, relative traffic to the site can allow you to make an informed guess as to which of two subject lines resulted in more opens, particularly if the difference is traffic is substantial.
Many email delivery services support auto-tagging of links, and may even work in conjunction with testing platforms to do remarkable things like promote “winning” emails automatically to later recipients based on the behavior of the first recipients. However, if you use Google Analytics and are capable of copying and pasting a link, there’s nothing preventing you from starting to collect email data right now that will help you create higher-performing email campaigns in the future.
Opinions expressed in the article are those of the guest author and not necessarily Marketing Land.