Email Roundtable Discussion: Email Delivery Wins


Kath Pay: Bill asked me to lead a discussion based on my article/case study: 0% to 100% in Gmail? Yes, It Happened, and Here's How!

We tend to treat deliverability as a dirty secret and more often than not only ever hear bad stories - so let's flip it and share our successes.

So I thought this is a perfect opportunity for those who have had a nice deliverability win (i..e had a deliverability challenge in the past but are now sitting pretty) to share what they did to turn their situation around.

You can make it anonymous i.e. 'a brand I knew', 'a client I worked with' etc or shout it from the rooftops that it was your brand - I mean, most of us have been there - right?

If we can keep this discussion real then we should have an interesting discourse. By 'real' I mean not just listing the best practices required for good deliverability according to consumer expectations, legislation and ISP requirements but also involve the fundamental reason why are sending email in the first place - and that is to make money.

Successful deliverability is an art - a delicate blend of all 4 key factors (consumers expectations, legislation, ISP's requirements and business goals) - so let's discuss!

Dela Quist:

Great job and case study it’s always good to see Touchstone users achieving great things particularly around deliverability. When it comes to inbox placement all deliverability experts speak regularly about the importance of open rates, but very few speak about the importance of Subject Line testing as the best way of achieving this. This may be because they fear the tests themselves might have a negative impact.

This is why Touchstone is so valuable, it uses AI to replace humans in tests. Only winning subject lines ever go out, meaning you don’t upset folks on your list with tests that fail.

I would also like to share something important we have learnt about deliverability using Touchstone, Subject Lines do affect deliverability. While we are all pretty much agreed individual words don’t affect deliverability, we have found meaning and/or sentiment does. What we have found is a better than average open rate does not always mean better than average delivery rate. Here’s an example, say you chose to use a misleading SL that got a lot of people to open the email. The chances of failing to deliver to the openers expectations will lead to an increase in spam complaints and unsubscribes, in turn leading to poor inbox placement.

Touchstone reports on delivery as you will see in the example below, taken from a we did test for a horoscope company. The clear winners were the last 2 tested one with a predicted open rate 33% above their average and the other 12% above their average.


Kath Pay: 

Thanks Dela for those insights. and I totally agree. Achieving high open rates using a 'click bait' approach will only cause you pain in the long run...not only deliverability issues but also loss of trust by some of your subscribers/customers by being misleading.

Remember how as email marketers, we used to discuss whether short or long subject lines were best? And more often than not, short was touted as being preferable? Well that was because when they were tested, the short SL came out as the winner when using the open rate as the success metric. Why? Because short subject lines are more often than not generic. And being generic, people can read into it what they want to, hence the short SL appealed to more people. BUT opens don't necessarily mean clicks and conversions. We've found that these short, generic SL are less likely to deliver clicks and conversions - simply because the offer or content in the body of the email didn't meet the expectation of the consumers.

However, long SL are more likely to be specific. Therefore, they're less likely to appeal to a wide audience, but rather the right audience and are therefore more likely to result in higher CTR's and conversions.

So the question should not be 'should my SL be short or long?' but rather 'should my SL be generic or specific?'. And the success metric to use? The one that supports your objective.

For example. During the deliverability reconstruction, our objective was to build up our reputation with the ISP's, hence we chose the open rate as the success metric, as this is the metric the ISP's base engagement on. So we tested both generic and specific SL's in Touchstone, and found (to no one's surprise) that the generic SL's delivered higher open rates.

Now that we're in the maintenance phase of deliverability, our objective is to increase conversions, so we've changed our success metric to being CTR's (due to client's technology issues, conversion % isn't able to tested) and hence our SL's are more specific now.

Has anyone else been on a similar journey?

Jaffer Ali: 

Kath, PulseTV is on a similar journey with our political newsletters from our e-zine site. We have two political e-zines; Conservative Review and the other is Progressive Review. We will be creating an Independent Review soon. We are using an aggressive subscriber acquisition strategy that results in a low open rate of 10.74%. These are emailed 3X per week. 2 content deployments and one solo ad for a PulseTV product.

We want to get the open rate to 20%+ so we have a lot of work. We will be sending a hot button political poll that is not click bait, but burning issues of the week to both Conservative Review and Progressive review. We are doing the programming now rather than use an outside Survey Monkey or other. First poll will be on impeachment. We will monetize the email with Power Inbox and the landing page with Adsense plus PulseTV offers. But the REAL goal is to increase open rates and increase in-boxing.

I like Dela’s overall philosophy of *cumulative open rate* and that is disturbing for these political newsletters as it is much lower than our daily deal cumulative open rate. We are hoping that topical issues will increase engagement…increase our e-commerce sales plus increase ad revenues. 

Kath Pay: 


Ooohh sounds like you have a good challenge on your hands with a worthy 'email meister' at the helm (i.e. you)

I love that you've identified that your real goal is 'to increase open rates and increase in-boxing' and it sounds as though you have a solid plan in place to achieve this goal. Then once achieved, your goal will change (as our did) to simply being revenue (via high open rates, deliverability and engagement) - but one step at a time right?

And I know that you're one person who totally 'gets' that deliverability isn't for deliverability's sake but rather to serve an end purpose (revenue) - you are a pragmatist like me :-)

I'd love to hear if you have other tactics lined up to help you double your current open rate. Any Subject Lines tactics or sending /segmenting strategies in mind?  And have you performed any interesting tests?

Jaffer Ali: 

This is a lot like back to the future as we used to be heavily into the e-zine/newsletter business. As for subject lines, the great polarized polling questions will/should carry the day. But we also will *go local* and do “Iowa” polls for those in Iowa… We discovered that local polls outperformed national polls. Our experience is that people “want” to feel their opinions count (although my personal POV is that we have one party in the US…the War Party).

We also will do other polls for non-political e-zines. We have 4 different humor newsletters: Clean Laffs; Bizarre News; Laff-A-Day (R-rated) and Phunny Pictures. We are going to do a poll to identify the funniest comedian of all time and list all the way back from Jack Benny and Lucy to George Carlin and Kevin Hart. We BELIEVE what PCH discovered. Let people participate or do something (take a stamp and place it on a an image) and their engagement increases. People who say *yes* to putting a sign on their lawn will vote something like 80% more than someone w/o a sign. Of course this is bad statistics, but directionally makes sense.

Kath Pay: 

Absolutely.It sounds like you're calling upon Robert Cialdini's Commitment and Consistency principle. Makes total sense. Not only will it help you with engagement and deliverability but also should go towards increasing the CLTV as well :-)

Sound exciting Jaffer! Good luck with it!

Dela Quist: 

 *cumulative open rate* and that is disturbing for these political newsletters as it is much lower than our daily deal cumulative open rate. We are hoping that topical issues will increase engagement…increase our e-commerce sales plus increase ad revenues"

The best advice I can give you on that is to think of your "yet to openers" as your most valuable segment, they are after all the largest segment you and most people have. More importantly you need to create a segment with its own test plan for Subject Lines, time of day/day of week, Frequency, Cadence etc. and separate tracking and reporting.

This list must only contain yet to opens, so every time someone opens an email the have to be transferred to a separate "active" list. Then monitor the performance of every email these folk receive what you begin to learn over time is which SL's and/or content types create new openers. Its like a card game where the aim is to get rid of all your cards.

I probably don't need to tell you this, but from a deliverability perspective you may have to snowshoe this segment in with your active segment when you send so your open rate averages across both.

Hope that makes sense

Brian Sisolak: 

We are just wrapping up ramping up a new client, and it’s pretty textbook so a fun one to share. The client had one major challenge, as a Union the email list is what it is, there is no more list building. So there is no way to make up for a blocked domain, so vigilance was paramount. Wanted to share some hard data.

The original dataset was 161,450 records. We used KickBox for cleanup, and for $735 removed 4,139 bad emails, 19 role and 89 malformed emails. As this list had seen very little previous email, we were concerned about high hard bounce rates.

Client is using Acoustic (aka IBM/Silverpop), with a dedicated IP and custom branded domain setup.

Even with strong (37%) open rates, Microsoft was/is still giving us trouble. We immediately noticed the Microsoft issue and pulled back, while maintaining the ramp up on other domains. By slowly ramping up in terms of overall volume and send speed (using the built-in rate limiting feature in Acoustic), we took it slow and low.

After backing off Microsoft, we ramped up again (we did not apply for mitigation). Still have inconsistent open patterns there, but suspect this client will be good in the long run.

After we sent the first email to all records, we updated the subject line and re-sent to non-openers. Given the client’s sending pattern we will only have a few emails a month, and I don’t want to let the new IP rest too long. This non-opener group looks to open ~17%, so are sending over the next few days at just 1k emails per hour (this is deploying right now, after running a 1k test).

We will then start with email content #2 slow and low again…

Overall Data

Domain Provider Open Rates

Kath Pay: 

Thanks for sharing Brian - appreciate it.

I also note that you made a key point: "The client had one major challenge, as a Union the email list is what it is, there is no more list building."

Lack of new subscribers with their wonderful engagement definitely is a key challenge when it comes to deliverability and was something I realised that I had unintentionally omitted from the case study. 

When I was delving into the data I realised that the list growth was static (or rather negative due to unsubs). So we explored why and found that technology had failed them and the new signups for the past 3 months hadn't been receiving any emails. Improving deliverability is reliant upon engagement, and who are some of the most engaged? New signups!!

So, we rectified the technology issue and are now happily sending to the new signups. We also have plans to increase the number of signups by adding a signup form to the website - currently they can only signup via a popover or by purchasing - so there's a huge opportunity to be had there.