Field Notes #043: Hard to Fake

David and I went to a live My First Million event the other day.

Sam Parr talked about how podcasting is a very hard format to fake.

“I can fake a blog post. I have faked a blog post, many times. But if I’m talking for an hour, twice per week, eventually I’d be found out if I were faking it.”

Here’s the deal: evolutionarily speaking, those who can successfully fake it are conferred an advantage akin to free lunch. But in most evolutionary games, it’s impossible to fake it for long; thus, costly signaling becomes the evolutionary advantage. Here’s how Rory Sutherland describes it:

“Signaling theory proposes that animals (including humans) may send honest signals about desirable personal characteristics and access to resources through costly biological displays, altruism, or other behaviors that would be hard to fake.”

So if you could write a blog post using AI in 1/10 the time it takes someone writing it manually, you cut corners but still get awarded the results from that effort.

And honestly, it does work. Try as they might, Google still tends to rank content that is faking it. That’s why HubSpot still ranks page 1 for “A/B testing” despite having factually inaccurate information. Realistically, if you are good at faking it, you actually can win. It’s just a short-term game, and we like to play the long game here 🙂

As generative AI improves, it’s going to get easier and easier to fake a surface level of knowledge. We can use Surfer or Clearscope to invoke the right keywords and blog post structure.

However, the second order effect of this is that, when everyone can fake it relatively easily, the value of doing so dilutes quickly.

Even today, faking it with regards to content puts you on thin ice.

I can tell you personally, having written a bunch of thin affiliate listicles on my personal website, that today’s page 1 ranking is tomorrow’s page 4 ranking as soon as the competition pulls the skyscraper technique on you. Eventually, you’ve got to compete on merit.

Faking it also tends to bring the wrong rewards.

You can easily fake LinkedIn and social authority. I once hired a LinkedIn agency to write my posts, and they had a whole network of people who would comment and like your content to trick the algorithm into showcasing you. They used a bunch of emojis and hyperbolic, clickbait language.

The result? I got a TON of engagement, but it was all from the wrong audience. I was attracting the wrong people. I played a stupid game and won stupid prizes.

On a company scale, there’s only so much you can fake it through hyperbolic marketing promises, but eventually, your product has to actually be good and keep those promises. Rand Fishkin talked about this in relation to the concept of the minimum viable product. It turns out, people are disappointed when you promise a 10 and get a 2.

That’s why I’m big on leaning into unique strengths and picking a few things that are hard to fake and hard to replicate. In content, that tends to be a given format (like podcasting or IRL events) or on the subject matter expertise or data that is unique to your brand.

The problem is that things that are hard to fake also tend to be hard to do, as well as hard to scale. But because of those barriers, they give you a very powerful moat if you can pull them off.

As AI lowers the barrier to entry to many forms of content creation, it’s important to find and index on a few things that you can do well, through efforts, that others will have trouble replicating.