POINT OF VIEW

AI isn't here to take my job. It's here to do the boring stuff.

Why the best AI in the world still needs an email marketer sitting next to it, and why that's good news, not a pink slip.

Let's clear something up: AI is not here to take every email marketer's job. I've spent more hours working alongside it than I'd like to admit, and I can tell you exactly what it is: the most talented, hardworking, fast-learning assistant I've ever had. It's also forgetful, occasionally overconfident, and completely convinced that coding for email is the same as coding for the web. It is not. It's harder.

Point AI at a website and it'll write clean, modern HTML that would make any front-end dev proud. But email lives in a parallel universe where it's still 1999, tables are load-bearing, and Outlook renders with Word's engine because of course it does. So I don't rely on AI to code my emails. I work with it. It's a phenomenal second set of hands for markup validation, spotting the typo I glossed over, and catching the unclosed tag before it catches me. It saves me real time. It just needs a grown-up in the room who's shipped enough email to see the disasters coming.

The future of email marketing isn't human or machine. It's the marketer who knows how to drive the machine, and knows when to take the wheel back before it merges us all into Outlook's rendering engine.

Open rate needs an asterisk now.

Open rate isn't gone. It's no longer a number I'd build a KPI around.

Apple Mail Privacy Protection pre-fetches images for a huge share of inboxes now, so "opened" often means Apple's servers checked in, not that a person read anything. A report that leads with open rate is mostly describing Apple's infrastructure.

Open rate still has some use. It can flag a broken send or a subject line that's badly underperforming. I don't want it anchoring a KPI target, and that goes for click-to-open rate too, since it's calculated from the same inflated open count. Delivery rate, click-through rate, and opt-out rate hold up better. I build reporting frameworks around those, and treat open rate as a helpful sanity check rather than the headline.

A campaign ID only earns its keep if every system agrees on it.

The value isn't the number. It's that email, social, web, and the CRM all point to the same thing.

A campaign rarely lives in one channel. The same push shows up as an email send, a paid social flight, a landing page tagged in GA4, and an opportunity logged in the CRM, and if the ID doesn't carry through cleanly across all of it, nobody can actually see the campaign, just fragments of it in four different tools. That's the real cost of a loose ID convention: not a wrong guess on one send, but four channels reporting on what looks like four different campaigns.

I treat the ID as connective tissue across channels, not a label that only means something inside one platform. The same ID, or a cleanly mapped equivalent, threads through the email platform, the social ad manager, GA4 UTM parameters, and the CRM, so a campaign can be traced end to end instead of stitched together after the fact. It's a small discipline, but it's what lets marketing, sales, and ops look at one campaign, across every channel it touched, and agree on what happened.

An MQL means more when both teams built it.

A lead score works better once sales helped write it.

I've watched marketing defend a scoring model and sales quietly ignore it, and both sides usually have a point. Most scoring models get built in a room without sales in it, then handed over as settled fact instead of a starting draft.

"Marketing Qualified Lead" only means something if both teams agree on what qualified means, and that agreement needs revisiting as the business changes. I treat the handoff as an ongoing relationship between two teams, not a one-way pipe. Teams that keep re-agreeing on what "ready" looks like tend to trust each other's numbers more.