Email marketing automation: what is actually worth automating

What is genuinely worth automating

Automation earns its keep where the same message needs to reach the same kind of person at the same moment, over and over. That is a narrower list than most agencies will sell you.

• The follow-up nobody does. An enquiry comes in, someone means to follow up, the week gets busy. This is the most valuable automation most businesses will ever build, and it is not complicated.
• The welcome sequence. New subscribers are never more interested than in the first week. Most businesses waste it.
• Abandoned cart or abandoned enquiry. They were ready and something stopped them. Ask why.
• Re-engagement. People who went quiet. Cheaper to recover than to replace.
• Post-sale onboarding. Reduces refunds, raises repeat rates, and almost nobody bothers.

The warning that should come with every automation pitch

Automate a broken process and you just get to be wrong faster.

If your follow-up emails are unhelpful, automating them means sending unhelpful emails to more people, more reliably, forever. If your data is a mess, automation will personalize incorrectly at scale, nothing says "we value you" quite like an email addressed to "Hi FIRSTNAME".

So the first job of an email marketing automation agency is not building sequences. It is looking at the process you already have, fixing what is broken, and only then making it run without you.

The manual task that should have been automated

The clearest case we see: a team copy-pasting leads between a web form, a spreadsheet and the CRM by hand. It works, until it does not. Records get missed, follow-up is slow, and eventually nobody trusts the numbers in any of the three systems.

Automating that hand-off end to end does two things at once. The follow-up finally happens within minutes instead of days, and the reporting lines up, because there is only one path the data can take. The team gets back hours of clerical work they never should have been doing.

That is usually worth more than any clever nurture sequence, and it is the first thing we look for. It is also less about email and more about workflow automation generally, the email is just where you notice the problem.

When not to buy automation

• Your list is small and your process is fine. Automation multiplies. There is nothing to multiply yet.
• Your data is a mess. Fix tracking and data first, or you will automate the mess.
• You have never sent a single manual follow-up. Do it by hand ten times first. You will learn what to say, and only then is it worth encoding.
• The offer is not working. Sending it automatically will not make people want it.

The rule we apply: do it manually until it works, then automate it so it always happens. Doing that in the other order is how businesses end up with an elaborate sequence nobody has read, running perfectly, forever. Our email marketing work starts with the sequence that already earns its keep.

Automation is only as good as the data underneath it

Every automation makes decisions on your behalf. Send this, wait three days, send that, skip anyone who purchased. Those decisions are only as sound as the data feeding them, and most businesses have never audited that data.

The failure mode is not dramatic. It is a customer who bought yesterday receiving an email begging them to buy. It is a lapsed enquiry getting a welcome sequence eight months late. It is "Hi FIRSTNAME" going out to two thousand people. Each one is small and each one costs you a little credibility.

• Check the integration properly. If the events flowing into your email platform are incomplete, your segments are fiction dressed as precision, Klaviyo's documentation covers what a correct setup looks like.
• Check the exclusions. Every flow needs to know who should not receive it. This is the step people skip and then apologize for.
• Check the timing against reality. A three-day cart reminder for a product people research for six weeks is badly calibrated, however well written.
• Check what happens on failure. When the data is missing, does the email send anyway with a blank space in it? Usually, yes.

This is why we treat data and tracking as part of the automation job rather than a separate project. An automation built on unreliable data does not fail loudly. It works perfectly, forever, on the wrong information, which is a far more expensive way to be wrong.

Fix the data. Then automate. In that order, every time.

What to actually measure

Email reporting is full of metrics that feel meaningful and are not. Two in particular waste a lot of everyone's time.

Open rates have been unreliable for years. Privacy features and image proxying mean a meaningful share of reported opens never happened, and a share of real opens are never recorded. Judging an automation on opens is judging it on noise.

Click rate is better, but still a proxy. People clicking is not people buying, and it is entirely possible to optimize your way to a highly engaged list that generates no revenue.

• Revenue per recipient. The number that matters. It survives list-size games and engagement theatre.
• Revenue per flow. Which automation is actually earning? Usually a small number of them carry everything.
• Time-to-first-response for enquiries. For a follow-up automation, this is the whole point.
• Unsubscribe and spam rates. The early warning system for a program that is sending too much to people who never asked.

To get any of that, the events have to flow correctly between your site, your email platform and your CRM, which is a tracking and data problem before it is an email problem. Google's recommended event structure is a sensible baseline for the website half of it.

Frequently asked questions