Austan Goolsbee and Peter Klenow compared overall inflation (red line) with inflation for goods that are sold online. They controlled for the fact that many goods like housing and fresh produce cannot be easily delivered by online vendors by matching the kinds of products that are sold online with similar products that are sold in stores and found that prices have tended to fall online more than in stores. The black DPI line below is their digital price index which they compared with similar goods in their offline CPI in red.
More useful for the average consumer is their finding that the best month to buy stuff online is July because prices drop the most in July with a smaller price drop in November. One possible reason online prices are lower in July than other months is that Prime Day is in July. That is Amazon’s big sale day which rivals anything in November.
However, the authors didn’t look at every single item that is sold online, so perhaps the apparent reductions in prices in July is mostly due to the large drop in prices for apparel. Their work makes it absolutely clear that the best month to buy apparel online is July and the worst months are September and October.
Bricks and mortar apparel stores have much less price volatility and have two peaks and troughs each year rather than the one large price drop seen for online apparel. They also have data on recreational goods which go on sale in November online:
You can also check the price cycles for each product that Amazon sells by checking out the data on Camelcamelcamel.com. Amazon changes prices on their website 2.5 million times per day and sometimes you can see patterns in the data such as the following graph showing the price of A Charlie Brown Christmas which fluctuates annually between about $16 and $2.
The best time to buy it is generally in the summer. That is some serious price discrimination. Keepa offers a similar service, but I haven’t bothered because they want users to register to fully use the service.
More importantly, if you pay any attention to online reviews, you should always check fakespot.com first before believing any ratings on Amazon.com. ReviewMeta has a similar service. I’ve become a lot more cynical about Amazon reviews after seeing more data about them.
[…] spotting algorithm to automatically assess the chances that each Trump tweet is fake or not like FakeSpot.com does to weed out fake Amazon […]