The lights are on...

Yesterday I received a second email from a major CRM/marketing tech business*. I mentioned the first in the sequence in my piece on optimising the funnel. The second deserves an article of its own.

The email had clearly been built on the back of two pieces of data:

• A company that I did some work for about four years ago, and

• That company being interested in, or needing, lead scoring

Neither was accurate. As a result, I received an email that wasn’t just irrelevant. It made the sender look incompetent, taught me to ignore them in future, and actively damaged the brand.

How does this happen? From the tactical perspective the answer is pretty straightforward. At a guess:

• The organisation looked at the account I was last associated with, albeit at least four years ago, and

• The account has not yet enabled lead scoring

The second is more speculative. It could be that product marketing, or customer marketing, have a push on to ensure that lead scoring is adopted by as many customers as possible. It doesn’t really matter. Those are the immediate causes of a useless email actively damaging to the brand.

But what are the underlying causes? Let’s talk about that.

Drowning in a sea of automation

Marketers have more power than ever before. We can publish more, send more, create more, promote more, analyse more, optimise more, iterate more. More, more, more. How do you like it? Don’t answer that question.

We have this ability thanks to a tech stack that gives us a vast number of data points, the ability to send and publish at almost infinite scale, at a marginal cost of something close to zero, and with the rise of AI agents, we don’t even have to do that. We can just sit back and let them do the job for us. 

So why is so much email marketing, or marketing in general, of such poor quality? Don’t forget, I wasn’t receiving email from a scrappy start-up. I was receiving it from one of the world’s leaders in marketing automation. They are supposed to be really, really good at this. How is that possible?

As with almost any question worth answering, there is more than one answer. Here’s a few:

• The low marginal cost is the problem. It’s free to send email, or somewhere close to it. It’s easy to measure clicks and even conversions. But nobody measures (and nobody can accurately measure) the slow corrosion of brand equity: it takes place silently, in thousands or millions of heads, over a time scale that doesn’t show up in a quarterly report. The incentives are all one way: send more, never mind the quality.

• Garbage in, garbage out. This was an immediate cause of the specific email I received. Whatever workflow was created for this campaign, it was not able to figure out that someone who hasn’t logged in for 4 years probably isn’t actively using that account. Automated systems at scale are rarely built well enough to avoid them making errors at scale. The rise of AI makes this truth more important than ever. We hear plenty in the press about the dangers of AI agents gone rogue. But those same risks apply here: if you haven’t got a very, very clear framework, things are likely to go wrong, and potentially badly wrong.

• (Clever) humans removed from the process. It is self-evident that anyone with a brain would not have sent the email I received. But more importantly, a human with a brain could almost certainly have written a single email, that was sent to every person who downloaded that whitepaper, that performed more effectively than the customised version. For context, the whitepaper was about AI, the email didn’t mention it. This point leads me to the big one:

• We love tech so much that we’ve lost sight of WHY. We have the capability to send millions of ‘personalised’ emails, so goddam it we’re going to do it! Sometimes it feels as if nobody is asking the fundamental questions: what are we trying to achieve? Will this work? Who are we talking to and why?

There are probably more. Who can speak to the internal workings of any specific organisation? What should be most concerning is that AI is only going to make this worse. Today most workflows are, at least to some extent, watched over by a human. In a world of AI agents the ship sails autonomously and nobody stands on the bridge looking out for icebergs. 

That opens up a lot of exciting possibilities, but also some rather unpleasant ones. How should we move forward from here?

Returning marketing to fundamentals

Earlier I pointed out that the email was not from a scrappy start up. But of course I had that completely the wrong way round. A scrappy start up would never send an email like that. They would likely have a small group of customers and prospects. They would almost certainly put a premium on personal, direct communication: doing “things that don’t scale”.

That is a reminder that human judgement should always remain at the heart of marketing, and broader go-to-market. As discussed above, ‘activity’ is commoditised at this stage. Judgement most definitely is not. And the former without the latter is not just ‘noise’ (which would be bad enough). It can be actively damaging to your business.

To conclude, and addressing the human being in the equation, here’s four things to bear in mind before spinning up your next campaign:

• Know the truth, and respect the truth. Don’t build large systems on best guesses or what was true years ago. If your data is patchy and inaccurate, don’t build on it until you fix it.

• Don’t be ashamed to simplify. If the alternative to building on inaccurate data is a simple, straightforward message that makes sense, sent to a relevant audience, then fine. Do that. Don’t feel embarrassed that you aren’t up to speed with the latest possibilities. The truth is you are way ahead of them.

• Understand the why. Why are you sending a message? What do you want to happen? Will what I am doing help? It is far, far too easy to do things in 2026 without having clarity around why you are doing them. But that clarity is what distinguishes meaningful communication from junk.

• Think of the negative externalities, and exercise restraint. Just because they are not measured, doesn’t mean they are not there. Unsubscribe and engagement rates tell only part of the story: step back and ask the fundamental question “would I want to receive an email like this?”

As per the title of this piece, a huge amount of this sounds probably sounds desperately old-fashioned. But I am not opposed to AI or any other part of the tech stack that has the potential to make marketing more effective and easier to deliver. Far from it. I just believe that now more than ever, we need to direct those efforts as thoughtfully - and impactfully - as possible.

*I know the default is assume I made this email up. I didn’t. If you really want to talk it over, drop me a line…