As part of our ongoing mission to help everyone send better emails, we recently brainstormed a series of questions that we could investigate in hopes of finding interesting or helpful insights for you. Here’s a selection of what we came up with:
- Does putting dynamic data in the subject line make a measurable difference?
- What is the optimal length for a subject line?
- What words, if any correlate with a higher open rate?
- If action seems required, is the email more likely to be opened?
- What are senders with the highest open rates doing differently?
We also wanted to be able to classify and clarify both the data that we were testing as well as the answers we were getting. To that end, we went through and sorted each of the templates in the dataset into one of nine categories:
| Category | Keywords, examples |
|---|---|
| welcome | verify email address, welcome |
| reminder | reminder, you still haven’t, trial expiring |
| billing | invoice, billing, payment |
| status_update | confirmed, shipped, received, printing |
| promotional | newsletter, redeem, sale, six tips to help you, coupon |
| social | user to user, invite, reply, followers, shared |
| password | password reset |
| garbage | this looks like a test email and should be deleted |
| unknown | category unclear |
After a bit of janitorial work, our investigation began. Shortly thereafter, it seemed to come to a bit of an impasse. The answers to our questions kept turning out less… definitive than we had hoped. So, not to be discouraged, we zoomed back out a bit. We had gone through the trouble of categorizing all these templates, and we had send and open data for them, so we just threw it up on the board, here’s what it looked like:
Being the astute reader that you are, you may have noticed that, of the original nine categories, only six made the final cut. Aside from the “garbage” and “unknown” categories (excluded for obvious reasons) we also left out promotional emails. There were a couple of reasons for this, but mostly it all comes back to the fact that it really ended up being a catch-all for unsolicited emails. This, coupled with the fact that many of them were blasted out to thousands of users meant that, not only were the weights of different templates wildly different within the category, the category as a whole also massively dwarfed the rest by quite a sizable margin.
So what can we take away from all this? On the surface, it makes sense, people don’t really like being marketed to and broad, untargeted emails are going to be largely ignored. I’m not restating this because I think you don’t know it, but to frame the more important point: it’s not that your users aren’t listening, they just reflexively filter pure marketing content.
Look again at these categories
Chances are you’re sending at least one of these types of emails to every one of your users throughout your relationship with them so why not hit them with the promotional content while they’re paying attention? I promise this wasn’t the conclusion we set out to come to, but it is rather convenient that this is something that sendwithus does quite well.
We’re just starting to tap into the power of the data we’re collecting, and we’d love to find new and interesting ways to put it to work helping you send better emails, so if you have any quandaries that we might be able to shed some light on, let us know in the comments, on twitter, or shoot us an email.