Email was the primary marketing channel for Natera’s first-ever D2C (direct-to-consumer) service. Our multi-channel strategy was still in development and, as a result, the sales and marketing teams relied on email to do the heavy lifting.
Internally, the attitude was “more is more.” More promotions, more emails, more frequently. If people weren’t biting, send them a coupon. If the coupon didn’t work, send them a bigger coupon.
Our team designed nurture and ad-hoc campaigns with inspiring imagery, vibrant branding, and concise, snackable content about our product and our value. We shot out the gate, emailing high volumes of our contacts on a frequent basis.Within the first four weeks, overall contact engagement dropped by over 50% (approximately our entire Gmail database), with Gmail open rates disappearing to 0.25%.
This dramatic decline in our open rate indicated we were likely ending up in spam.
We validated this hypothesis through internal testing and insights from HubSpot’s deliverability team.
Over the next eight weeks, we implemented a cocktail of recovery tactics and recuperated our Gmail open rates from 0.25% to 30%. During this process, we overhauled more than our deployment practices.
We identified that we were missing an onboarding process, relevant and helpful content, and a long-term vision for our relationship with the customer.
Now, the organization has shifted its focus to a new goal — helping our community have the healthiest pregnancy possible.
This is a great example of how a "problem" with spam led to a positive outcome. In our case, this meant a directional change in Natera’s marketing strategy.
Here are the steps to recover from and improve bad email deliverability, assuming your IT department and the HubSpot deliverability team have already authenticated your domain.
Suppress all emails that have not previously opened an email from all of your future campaigns. We suppressed contacts to a greater degree than the “disengaged contacts” feature would in HubSpot.
Immediately, when we knew we had a problem, we suppressed anyone with a Gmail id that had never opened an email. I assumed these emails go directly to the spam box, not the primary inbox.
We used list criteria to determine Gmail contacts who never open an email.
You can add property Email Send Date is Known if your database also contains "cold contacts" — people who have never been emailed. This adjusts the segment to only include people who have been sent an email. This list criteria was used as a suppression list on all automatic workflows.
When we began sending email again, we restricted lists to include people who had opened more than two emails. We can assume many people open our first email to scope out who it is. Opening a second or third email is a much stronger indicator of true interest.
It’s a great idea to understand how your database engages with email. Can you answer these questions: How many people open one email? How many people open seven emails or more? This data will help you tell your story. What you do with this information is up to your team.
In order to to compare your segments by engagement, you can either:
A) Build lists by including the email opened criteria and iterating your desired permutations: is equal to 1, 2, 3, 4, 5, 6. . . and compare the list sizes.
B) Or, more efficiently, export your list of contacts and include the emails opened property as a column on this list. Sort this .csv outside HubSpot in Excel.
What’s this? If a contact hasn’t opened any of the past several attempts, send them an email with subject line: “Are we sending you too much email?” We got a 55% open rate from a segment with a historic 0% open rate. This has a few functions (better average open rate, contacts can unsubscribe safely, less likely to mark you as spam).
If after three emails sent in the primary campaign, the contact has not opened more than one email, send to them a break-up email.
Here's the workflow branching:
And here's the email:
This technique worked for us because our database had a high volume of “cold leads,” which could easily be burdened and need a safe mechanism to opt out of email.
You can either replicate the remainder of the workflow below this if/then branch and keep sending the contact email, or terminate the campaign there.
Here's what our "before" emails looked like. This email contains “spammy” language that likely triggered Gmail’s spam filters.
Here's what we were sending after. In this new email, our goal was to encourage engagement with our emails in order to boost our inbox reputation.
Scrub all your email lists for invalid or "risky" emails. We found up to 10% invalid email ids. Use this scrubbed list as a suppression list or re-upload it as an unsubscribe list if preferred. This will also keep your storage costs low by eliminating invalid ids from your contacts list.
After restricting sending to public Gmail domains for six weeks, we did 16 cycles of small blasts over a period of three weeks. We manually sent emails in small batches to the Gmail list. In this experiment, we saw up to 30% open rates, up from 0.25% six-to-eight weeks earlier.
Create a range of relevant content in order to appeal to your user base. Before we created anything new, we brainstormed various categories our customers might be interested in, such as:
We used these to built a content mix in order to keep fresh and interesting to our user base.
When thinking about our content mix, we asked ourselves: What do our customers need to know right now in their lives? Are we talking too much about ourselves?
Significant effort was made to educate internal stakeholders on the technical and non-technical reasons why spam happens. It’s important decision makers understand the basics of how to maintain a good domain reputation and inbound marketing best practices.
Our recovery strategy encouraged us to be more deliberate, thorough, and thoughtful with our email marketing plan. The "down" period of the recovery process created space to investigate, debate, and flesh out the details of our approach with company stakeholders. We finally had the attention, time, and resources to implement a robust plan.
Strategy changes/gains:
Technical changes/gains:
The goal is to build trust with customers and provide them with useful information they couldn’t get anywhere else. The development of highly personalized nurture campaigns will help you serve up helpful, relevant content based on where the customer is in their personal journey.
Finally, develop a more comprehensive multi-channel onboarding campaign. This means that instead of relying on email to reach an audience, leverage SMS, email, and social media together.
Originally published Sep 20, 2017 10:00:00 AM, updated November 28 2017