How To Write Email Subject Lines That Get Results
Do emoji increase open rates? Is shorter better? What about #?
The short answer is: Maybe… But that’s not a real answer.
What does science say?
Here at Phrasee, we have proven that there is a science to creating email subject lines. And to save you the heavy lifting, we’ve done the analysis to prove it.
We took a random selection of Phrasee tests (around 700 million emails) sent by commerce and retail brands based primarily in the US and UK. Then we ran billions of Markov chain Monte Carlo simulations to statistically predict which phrases work the best.
We looked at specific elements in subject lines to determine how they affect open rates, clicks and conversion rates.
We won’t go into details about the scoring, but each element was given a normalized, weighted score which aggregated the overall effect on the response metric.
In this experiment, the higher the score the better. (Check out the full report here to see the details of how we did this.)
Here are the highlights:
Trying to figure this all out and implement it in a systematic way can we almost impossible, but here’s our top tips for making sense of all of the above:
DO test your subject lines. Anyone who says they can predict results by looking at past subject lines without testing their assumptions is categorically and scientifically wrong.
DO look through the results and think about how you can apply them to your brand. Are there some phrases you’ve not considered using? Some you use too much? Some emotional sentiments that you’ve not considered?
DO consider all the statistics, including the quartile spreads. Some words and phrases work some of the time, for some people… how’s that for a wishy-washy outcome! This is the challenge of subject line testing. What works for you one week won’t the next… and what works for one brand may not work for another.
Phrasee is the world’s preferred subject line optimisation platform. To read the full report on how to write subject lines that sell, check out our website.