Robotized, Not Personalized

Ah... personalization, a topic almost as old as the interwebs, and here we are, figuring out if the robots can save us.

The other day, in a post on LinkedIn, I joked that:

If I ever take a knock to the head and forget what I do or what my company does, I have comfort in the stream of “personalized” emails that will tell me - "Hi Ian, I see Velocity B does <a thing>...." - honestly it's good to know these people are here, it's a service, really, maybe even a daily reminder of the mission... thank you, AI SDRs, keep doing the good work.

And of course we all have the examples where “the thing” is way off.

Just now, yes, this very moment, an email just popped into my inbox from someone who’d watched and enjoyed my podcast (which is resolutely audio only), wondering how I’d generated the clips (which don’t exist), in a pitch for video editing services

I’m excited by the potential that AI has in scaling sales and marketing outreach, and I’ve invested a fair amount of time digging into how to get that balance of the promise of scale versus what needs manual intervention.

Get that balance wrong by trusting the machine, and we are pitching video editing services to someone with an audio-only podcast.

Get that balance wrong by not trusting the machine, and we spend 30 minutes on every outreach, and we don’t get scale.

The reason why we end up with “personalized” icebreakers that are as banal as saying “you are a fellow human that breathes air” or “I see Velocity B does the thing it says on the thing that I can easily scrape”, is that personalization in cold outreach is hard.

And of course it’s hard; this is a cold email, we reliably know fuck all about them.

When we try and get fancy and ask our robot to serve something “personalized” with no previous data to go on, apart from what the public web tells them, they make assumptions.

Like the assumption that all podcasts are video, or in my experiments, an LLM drafted a wonderful email for me to send that talked about someone speaking at an event THREE YEARS AGO, as the prospect had barely posted on LinkedIn since, and it was all the context it had.

So we default to what’s less risky; we know who they work for and vaguely what that company does.

Of course, there could be something wrong with my robots or how I am training them, but if we are clutching at synthetically generated straws about our audience, and we have to default to what’s safe, why go down this personalization rabbit hole at all?

Don’t get me wrong, I understand the value of finding an icebreaker and the psychology.

Using someone’s name builds trust, as does being relatable and relevant - or, as the science says, “brings cognitive ease through relevance bias,” as the reader doesn’t need to figure out if this is for them. Plus, done well, it shows we are paying attention to them, and we all love to be paid attention to.

Personalization is acting on behavior and high-quality signals that we can act on. Like, a website visit, a genuine intent signal, buying our products, or even publishing a snarky post about personalization on LinkedIn

But, if we don’t have good data to create a genuine hook - once we get past “Hello {firstname}”, telling them what they do and what their company does not build trust, relevance, or show that we paid any attention; it's just noise.

Surely, they are going to be more interested in what they don’t know, which is likely to be something about us, our service or product.

Or better yet, the problem we solve, or a nugget of industry insight that they may not have come across.

Or better yet, how other companies like theirs have benefitted from what we do.

In my example, telling me “other podcasts have benefited from video editing” with a beautiful example might actually drive me to publish my podcast on video and give you a call.

Telling me you watched my audio show encourages me to write a snarky Tuesday 2¢.

Who needed that?

True personalization makes the email more useful to me.

If not, let's cut the crap; it’s a cold email - we should tell them how we can help.

There is a chance that's way more interesting.