Skip to main content

The Mystery of the $190

Once upon a time, when Amazon was more new and less evil, I was diligent about linking mentions of books to Amazon along with referral codes. I never made a lot of money by doing this but occasionally I had a pleasant surprise when I had to work out how to pay a US dollars cheque into my bank. The rewards from Amazon.co.uk were even more occasional but still pleasant.

These days I’m less diligent, partly because of Amazon’s increased level of evil, but partly due to apathy. The only place I still regularly link to Amazon like this is when adding a new book to my reading list, and that’s more as a way of saying “find out more about this book here” rather than “Buy this on Amazon!” (As I mentioned recently I’ve been wondering where else I can link to instead.)

Given my dwindling usage of affiliate links I was surprised when I received a letter from my bank recently saying they’d received a payment of US$190 from Amazon.com (I have figured out how to avoid receiving cheques). Given I usually only get payments of US$10-20 every few months this was a surprise.

So I went over to the Amazon Associates site, and found the graphs, and filtered things down, and there was February 2020:

A chart showing earnings of $52 on 10th February, $43 on 11th, and $27 on 12th
My Amazon Associates Reports chart for February 2020

Strange!

Looking at the table of items sold that month I wasn’t much wiser:

A table of 20 items, mostly books, with the top three being 'Economics in One Lesson', 'The Theory of Poker' and 'How to Lie with Statistics'
My Amazon Associates ordered items table for February 2020

A lot of books, one or two of which are linked to from my site. But others seemed too… similar… and purchased too often, to be random things people had happened to chuck in their basket at the same time as following a link from here. It seemed more likely these were visitors from Gyford.com rather than, say, The Diary of Samuel Pepys but it still didn’t make much sense.

Then I remembered that while I usually use the same Associates Tracking IDs everywhere I haven’t always been lazy about this – I used to set up different IDs for different websites. So I filtered by those and – ta da! – all those purchases had come from Septivium, a site I keep forgetting I started (and stopped), a domain I always have to think hard about paying to renew.

But, still, why had the long-dead Septivium suddenly arisen and been the source of all these purchases? I googled for mentions of the site’s uncommon name over recent months, and nothing stood out.

Then I remembered that, while I removed the also-slightly-evil Google Analytics from my personal site last year, I hadn’t got round to doing it on other sites and, yes, I still had it set up on Septivium. So, after a poke around in there I discovered this for the site’s traffic in February 2020:

A chart of very low traffic except for a spike on 10th February of about 10,000 pageviews
My Google Analytics chart for Septivium.com in February 2020

Surprising, for a site that’s not been updated in seven years. Looking at the Landing Pages, almost all those users had arrived on Ask MetaFilter’s best introductory books, a post from 2009. All GA could tell me about the source was that these were visits from “(direct) / (none)”, which wasn’t very helpful.

Where had these thousands of people – there were as many users as pageviews – come from?

Google no longer supports searching for link:http://... so I tried a different “backlink checker” website for the page’s URL, and looked through all the decade-old links trying to find recent ones and, at last, I found a couple of pages by David Perell linking to that Septivium page.

First, this long essay, News in the Age of Abundance which, near the end, includes a link to the Septivium post (“the best introductory books on basically every topic”). I had to look at the essay’s source to discover it was published on 3rd February 2020. (Put dates on your posts and articles, people! It’s the basics!)

Second, this page on threader.app, which collated some of Perell’s tweets, including this one from 6th February 2020:

David Perell @david_perell

Hidden gem alert!

Here’s a list of the best introductory books on basically every topic.

Topics include aeronautics, hunting, lexicography, ancient Egypt, animation, architecture, biology, linguistics, oceanography, microfabrication, and typography.

septivium.com/b/2009/05/07/mefi/

Given Perell is apparently very popular, and given the dates, he seemed like a likely source, although still a few days early for that spike on 10th February…

But, aha! Perell has a weekly newsletter that goes out to 20,000 people on Mondays, and 10th February was a Monday! Presumably he included a link in the email (the only archive ends in July 2019).

I guess that’s it. I think we can now class The Mystery Of The $190 as solved.

Going back to that post on Septivium, and also reading the contemporaneous MetaTalk thread about it, there’s some discussion about my use of my own Amazon affiliate links, given I’d compiled the list from an excellent Ask MetaFilter question. Overall, in the end, it all seemed fine for me to have done this but I said that, anyway, “I don’t think I’ll be retiring on the proceeds 🙂”.

Clearly, despite today’s excitement, I’m still not going to retire on $190, although it’s a big enough chunk to be worth donating to MetaFilter, which I’ve done.