In a way this is a restatement of the efficient-market hypothesis. If the market is efficient, then a product that is inferior in most or all dimensions will be forced out of the market by one that is superior. Therefore, if you believe in efficient markets and observe that both products are still offered for sale, it's a reasonable inference that the superior product must be inferior in some way that you haven't yet observed.

Usually where this heuristic fails is on the margins, when markets are not efficient and you're observing a change in technology or production techniques. In this case, the correct inference isn't that the product is inferior in some way you haven't observed, it's that the superior product will eventually end up winning and taking over the marketplace.

I think your argument somewhat presumes that we all agree on some platonic ideal of a product.

Let's pick restaurants as as product: I have some favourite restaurants that for me beat most other restaurants on multiple dimensions (taste, freshness, service, location, price). And there are very popular restaurants I rate poorly on multiple dimensions.

I know other products that I think are both cheaper and better than the competitors, but I can't think of any products where I have many many purchases within the sample space of different products.

A nit pick: you're describing something more like perfectly competitive markets through to monopolistic competition.

The efficient market hypothesis is that the prices offered for purchase and sale factor in all relevant information about the thing being traded. It's less about whether that good is "better" or "worse" than a substitute, more about what information is revealed by prices.

Doesn't open source software kind of throw a spanner in this idea? In quite a few fields, the best option could be released for free and made by a team of volunteers, and if so, it would logically be better than its paid alternatives in pretty much every possible way.

There are important exceptions, though. It's well demonstrated that wine price influence taste perception, for example [1], [2], [3]. And many of us have seen tech consultants charge prices all out of line with value delivered.

[1] https://www.caltech.edu/about/news/wine-study-shows-price-in... [2] https://www.sciencedaily.com/releases/2017/08/170814092949.h... [3] http://www.openculture.com/2017/11/expensive-wine-is-for-dup...

When it comes to products, marketing is also one of those dimensions, and can easily swamp the others. It's a good assumption that if you hold price steady, the more marketing, the worse the product. We don't see this because it is often canceled out by economies of scale, which of course are a result of size, and the larger the relative size, the higher amount of marketing as compared to other brands.

At the same production scale, you should expect the most marketed product to be the worst.

There's a school of thought that expensive advertising is actually "signalling" higher quality to consumers. Why spend so much on something substantially less appealing than available substitutes?

It doesn't help that size is partially a result of marketing. At the same production scale, the most marketed product may not be staying at that scale as long (assuming the marketing is being done somewhat efficiently, which perhaps isn't necessarily a good assumption).

> Assuming price is one of the dimensions, this is generally a true statement. Otherwise we'd have a bunch of best-in-class products that are cheaper than inferior competitors.

You say this like it's an absurd condition but we can observe exactly this condition all the time, with new manufacturing techniques and labor practices creating better products for less resources all the time.

One place I love to observe this phenomenon is cars, where often even new inexpensive cars are superior in many dimensions to older luxury cars (including passenger amenities). But brand attachment and carefully seeded misleading advertising allows luxury vendors to keep their prices high despite a slower product development cycle.

Would this be dissimilar to undercutting a competitor? For example Ryzen 3rd gen chips seem to out perform, and undercut the price, of Intel chips with a similar feature set. Maybe brand is a dimension?

Sometimes this happens for a period of time, but in the medium- or long-run, the manufacturer of the superior product will raise prices. Undercutting is just a way of getting attention in the short-term.

The other alternative is that the inferior and more expensive version simply stops being produced.

Wait until tomorrow until you say that, the third party benchmarks aren't out yet. ;)

Brand is certainly a dimension in some cases (see: Apple products) but it's inconsistent per consumer.