It's hard to reason from his post, but it seems like he's spending money on branded search terms. E.g. queries that include his website in its name. The effectiveness of branded search terms is absolutely worth discussing, and it's pretty easy to run a small scale test on if it works by comparing organic and paid traffic after the fact.

But the issue is that this article doesn't differentiate between branded and non-branded search terms. Did he run campaigns on non-branded searches like "Registration checker UK" to get build awareness of his service for new customers doing broad searches? Did he compare his organic non-branded search engine presence with his results when running paid ads?

Also, his service doesn't seem like a great match for the advertising channel either. It's hard to turn a profit when your service costs £8. The product - channel fit [0] doesn't seem right. So all in all, not such a great experiment.

https://brianbalfour.com/essays/product-channel-fit-for-grow...