On 15 July 2020, I bought a 12 inch inner bicycle tube for my 16 inch bandsaw, on amazon.nl. I paid 2.51 euro for it, which was quite cheap compared to other shops. But a few weeks later I noticed that the price had gone up - and not a little but to almost 9 euro - more that 3 times the price I paid!
Of course all prices on amazon (and similar websites) are "flexible", and are determined by automated (and complex) supply-and-demand algorithms, but this was a bit shocking. So I decided to write some code to monitor this trend. What else could I do? ;-) Well, obviously this has been done before by other site that track prices, like
camelcamelcamel.com and
keepa.com - but for amazon.nl this wasn't available at the time. And besides, it's fun to build.
I have included 2 other items on amazon which I looked at, and for which I noticed big prices jumps too. I fact, I didn't buy either of them because the price had gone up while I was still thinking about them. I'm not sure if amazon really intended that end result of their price changes...
I also included amazon.de, to see if the prices on there did roughly the same - which they do, more or less. Nothing shocking anyway, although the price difference between NL and DE for the 3rd item (Steel Ruler Set) is, and stays, quite large.
I'm not quite sure what to do with this information, but it sure is interesting. The price of the 12 inch tube did drop to it original level eventually, but that took quite a few weeks - and then went up again to 5 euros shortly after.
As a side note: Originally I had the database job scheduled to run every minute, instead of every hour. I had no idea yet how often the prices changed, so I thought I'd make sure to capture as much data as I could. However, it turns out Amazon does not like that kind of behavior very much (can't blame them really...), so after about a day I got 500 server errors for almost every try. Polling every hour seems to work fine, at least for a few months.
Note: May 10th 2021 I stopped the database job. The ted_amazon_logs table was completely filling up my Users tablespace (8112 MB and counting), and that was a bit much for my small VPS. I probably should disable logging for that table, but I think the experiment has gone on long enough to prove my point.