How to debug low open rates
Every week or so we get a customer sending us a message to ask us about their open rates and why they are still low after switching to Bento.
When debugging these tickets our job is focused on:
- Providing you with amazing tools to debug your deliverability
- Send from email servers with good reputation
- Give advice on how to debug bad delivery
Your job is to:
- Debug these issues one by one with patience
- Improve the quality of your subscribers being added
- Improve the quality of content your are sending them
At Bento, we have seen upwards of 100%(!) better open rates (vs. competitors) when we've been able to debug successfully together. These results have generally lead to a huge increase in customers engaging (through replies, tracked pageviews, etc) or revenue.
We see most engagement problems can be broken down into two categories: reputation or content issues.
In this article, we're going to specifically focus on the reputation and network issues because they are the easiest to debug.
We don't recommend starting with your content because you'll get stuck in an endless cycle of optimizing things that don't matter in the large scale of things like removing an emoji from your subject line.
The frame of mind you want to be in when going through these steps is: "how can I develop a good relationship with the email providers like Gmail and Yahoo? What would they like to see from me?"
Are your emails actually coming from you?
When you send out your emails you have the choice to sign them to tell email providers like Gmail or Yahoo that the email is actually from you.
In Japan, this is sort of like how we are able to sign everything with our stamps as no two stamps are exactly the same.
Email servers can do this too!
If we couldn't do this then phishing would become rampant as anyone could sign up to Bento or other providers and just pretend to be you. Not good!
Signing your emails protects you from phishing attempts and may help improve deliverability.
If you have a small list, say a few hundred people, you may not need to sign your emails just yet but at scale you certainly will. When you're sending upwards of 50,000 emails per month Gmail and others will want to make sure you're not sending out dangerous phishing scams.
So if you haven't already, go read our guide for setting up "Sender Authentication".
Are you sending too many emails at once?
Imagine you're Gmail and you see 500,000 emails coming your way at once. Do you think you'd handle those requests differently than 500,000 emails sent over 24 hours at a rate of 20k/hr?
Now, imagine those same emails are sent from a new domain or server that you've never seen before. You'd probably mark them as a risk and assess them differently to other emails coming into the system.
To help manage this, we allow our users to control the rate at which we send their broadcasts. This allows our customers to be kind to the inbox providers.
Search for "Batching" in our docs to read more about how to optimize your emails for this but, in general, when debugging your engagement rates, you will want to go to the slowest batch you can stomach.
This could literally mean sending a new email broadcast over the entire week instead of day.
Once you notice a steady incline in open and engagement rates, which you will, you can steadily ramp it up sending over 4 days then 2 days and then back to sending a broadcast in 1 day.
We've seen serious customers send as slow as 50-100 per hour!
These customers were able to effectively take their open rates from single digit to double in the course of a month because they were patient doing it correctly.
Is your HTML too complex?
This one is a little more nuanced but if you've gone through building a broadcast in Bento you know that one of the last steps is a spam check.
That spam check works out if your HTML is too complex or has characters that may flag systems where email is sent through.
This uses our own algorithm and not one used by large inbox providers but it's a nice, helpful guide on things you could change.
Whilst we don't recommend spending too much time here we do recommend split testing simpler layouts and email structures to get the score down.
The most obvious decision is to move from HTML-based emails to plain text.
When you do this you'll see your spam score drop drastically as a result and possibly see an improvement in deliverability.
Again, don't fuss over tiny improvements here. Go for the big macro changes.
Is your email to large of a download?
Short tip here: stop adding so much to your email. Keep it simple, reduce the images you have added, and try to keep emails concise.
If you're uploading massive images which take time to download they will be prioritised differently (and sometimes not delivered at all). Remember that Gmail will check all these images and they probably don't like a lot of bandwidth being consumed by your files.
Are you sending at the wrong time of day?
Something that has surprised us is that time of day isn't a strong factor at all when debugging how to get an open rate from a single digit number to double but it can help get it up a percentage or two.
To properly work out the best time of day send a single broadcast over a period of 7 days (sending hourly).
Study the hourly graph that Bento provides and identify the bumps in opens. This will be the best time window to send in.
Do your subscribers expect your email?
The next thing to consider is if your subscribers expect your email to arrive. If you're a publication, like The Hustle, then subscribers will expect to see your email in their inbox every morning so will look for it.
Most eCommerce stores don't have this regularity or cadence but implementing it can help (and actually help with your reputation due to the consistency).
We will be providing more tips over the coming months but believe with the above you can make a large impact.
If you keep having issues just let us know!