When you’re optimizing email automations, it can feel easiest to simply toss out the existing emails and start from scratch. After all, the existing emails:

  • Were likely written based on guesses and hunches, not data;
  • Typically underperform, which is why you’re looking at optimizing them; and
  • May be entirely outdated.

And they may not even be worthy of the title “control.” That’s because a proper control should have proven itself; just because a thing exists does not make it a control, does it?

No, it doesn’t. However, your clients or your team are going to refer to it as the control… so roll with it. 😇

First, where not to start.

We’ve established that you should not start by throwing out the emails and starting from scratch. But what else should you avoid when optimizing an email automation?

Avoid rewriting any part of any email yet. You don’t know what’s working and what’s not. That subject line you hate? How’s it performing? We need to understand performance before we start developing hypotheses for how to improve performance.

Avoid commenting on writing style. Does an email feel particularly long to you? Or far too short? Are the paragraphs massive, or are there so many bullets you feel shot through? Does the voice feel dull to you, or is the tone way off, in your opinion? Great. You’re not the prospect, though, so keep it in your optimization notes – don’t start making assumptions or recommending changes yet.

Avoid solving guessed-at problems. Have the discipline to first fully identify the problem. And have the deeper discipline to do the work that comes before identifying problems. That is, have the discipline for what follows here…

Where to start? Start with an audit of these two parts of the email automation.

First, audit the cadence itself. Get out your Miro board or Whimsical map – this is when you sketch some boxes and lines! Keep in mind this is largely a thinking-free exercise; all you’re doing is documenting what’s there, without any commentary.

We begin with a map, like the one shown above. The map shows if-then scenarios, triggers and boxcars, as well as core flows and segmentation. This map is likely to be more detailed than the automation map that’s in the CRM itself, so you want to create it outside the CRM. I use Whimsical.

Once the flow is mapped, we can audit it.

In this audit, you are looking for:

  • Cadence. How much time passes between the trigger action for the automation (e.g., new lead successful form submission) and the first email appearing in the lead’s inbox. How much time passes between the third email and the fourth?
  • Necessary practices. If the first email is a double opt-in email or not. If it is, do further emails hinge on the confirmed opt in, as they should?
  • Triggers based on action. If a lead clicks to register for a webinar… and two hours passes… and they haven’t registered for the webinar, what happens in the flow?
  • Triggers based on inaction. If a lead stops opening on Day 5, what measures are in place to bring that lead back into the flow?
  • Tagging and segmentation. Every open, click and delay tells us something about our lead. But AI hasn’t yet figured out how to tell us what those actions and delays mean about each prospect. So we rely on tagging leads with what their interactions tell us. These tags need to be documented.
  • Lead scoring. Align your Sales team’s lead scoring sheet with actions across the flow.
  • Black holes. We’ve worked with too many businesses that have 100,000+ leads on their list, but only 3,000 have converted to paid and only 18,000 open emails with any regularity. What are the other 79,000 addresses up to? The challenge typically is that when a lead exited a flow or graduated to a new status – and then nothing came of that – they got dropped into nothingness. IS there always an appropriate next action or step for a lead? What’s the next automation they go into?

Second, map performance for each email in the flow. Write the following next to each email:

  • True open rate (avoid letting Apple mess you up!), average
  • Click thru rate, average
  • Conversion rate, average
  • Unsubscribe rate, average
  • Revenue attributed over time
  • Revenue attributed in the last 30 days

When it comes to revenue, I prefer revenue directly attributed to the email. This is a “last” attribution rather than an “each” attribution, for the ad folks in the room. So while all 8 emails in a flow may “each” contribute $100,00 to $800,000 in total revenue over the lifetime of the flow, Email 4 is the “last” email touched before $600,000 in revenue was generated; I would say Email 4 is freaking killer and everyone needs to be careful not to mess up anything to do with Email 4 — and while we’re on the subject, in our next phase, let’s really analyze what’s so freaking great about Email 4 so we can do more of that!

What happens after you’ve created the map?

Next, you get to start using your brain. I look at the map as a whole. And I make big green circles around the parts that look so good we should basically hold them sacred and big red circles around the problems we’ve gotta get after immediately.

What gets a green circle?

🟢 Emails with an unbalanced [high] amount of revenue attributed to them.

🟢 Emails with very high click-thru rates.

🟢 Emails with very high conversion rates.

🟢 Emails with ABNORMALLY high open rates.

What gets a red circle?

🔴 No double opt-in email. (Exception: automation is not for new leads.)

🔴 If the “day zero” email(s) has an open rate under 40%.

🔴 Emails with ABNORMALLY high unsubscribe rates.

🔴 Emails with ABNORMALLY low revenue.

Then there are yellow circles for…

🟡 Cadence issues, such as what could be too much time between emails or not enough time between emails.

🟡 Segmentation issues, such as too many segments (where perhaps more conditional messages would simplify sends) or not enough segments (where diverse groups are lumped into one for the convenience, often, of the sender).

🟡 Black holes, and dead ends for live leads.

What happens once you’ve got a bunch of circles everywhere?

We also need to zoom out from the flow itself to audit sending practices. Compliance affects deliverability.

  • Is DMARC set up?
  • Is opt-out easy (i.e., one click)?
  • Are we sending from a real address or a no-reply?
  • Do we violate CAN-SPAM, GDPR or CASL?
  • When was the list last scrubbed to remove disengaged subscribers and bots?
  • Is the from name and/or sender (and reply) email trustworthy?
  • Is your sending domain trustworthy?

If an email sender has a poor reputation, their emails won’t land in an inbox. Identify your sender score with a tool like this, which also helps you optimize your sends overall.

Finish what you started with a Report on Email Automation for your brand.

Document what you’ve found, and present those findings to your team or client.

You’ll want to continue with a deeper heuristic analysis of the messages in the emails, including not only what you’re sending and how it’s written but also where the offer appears, if it’s optimized and much more. But first you need your team aligned on what’s currently happening with the automation. So present the above before moving on.

Next, advance your email copywriting skills: