To take your question seriously – one of the primary surface areas for fraud is exploiting the customer service function.
Just to work one (real) example:
1. Buy expensive face cream from Amazon
2. Take out the expensive stuff and replace it with junk
3. Complain to Amazon that you got junk and need a refund.
The challenge here becomes obvious. Even in claims where the customers was allegedly a victim, some portion of those customers are fraudsters. So any policy that results in a large, profitable company will likely result in something less than 100% of legitimate fraud cases from being satisfactorily resolved.
In the case of the small business, they are very unlikely to be targeted by industrial scale fraud and will probably just be able to solve this with personal relationships and human judgment.
That said, the price you pay the small operator is very likely to be meaningfully higher than the price of the scaled operator.
If you care more about having near 100% probability of having fraud/abuse cases addressed than you care about benefiting from the economies of scale, then yeah, go with the small operator.
In things like communities, my general sense is that there is almost no beneficial economies of scale. So I like to hang out in smaller places. But in things like information retrieval (ie. Google), the benefits are usually worth it for me to use that.
The kind of experiences that the post and I am talking about are where the platform user or the customer end are dissatisfied by the platform's conduct.
The question of the company itself being defrauded by some number of customers and having to defend against that as it becomes increasing larger is definitely valid and difficult to tackle at scale. However, I think it is a totally separate one. I do not think it is a necessity for the company to compromise the ability to make any one dissatisfied customer whole just because they are being targeted by a whole lot of fraudsters all at once. The company might have to increased their costs on their products to pay for things like additional training, tamper-evident packaging, improving supply chain management, and so on.
They aren't separate questions, and there is a trade-off between protecting themselves and attempting to do right by good-faith customers.
The correct amount of fraud the platforms tolerate against themselves is definitely >0%, and it's for the reason you're identifying.
It seems that you're assuming it's obvious to identify good-faith consumers and do right by them. In some cases that's true. If you're a loyal, repeat customer of Amazon's they will basically just do whatever you ask them (within reason). But there are many cases where the act of discriminating between fraudster and Good Samaritan is in fact the whole of the challenge.
Indeed, one of the most common exploits is to hack the account of good customers and abuse merchants that way.
There's a dark, dark underbelly of crime out there. And the cost of that is borne by all of us in occasionally-less-than-good customer experiences.
> can they make their customer whole 100% of the time?
They do make their customers (i.e., the companies who pay them for access to their users) whole. But their users (at least the vast majority of them) are not their customers; they don't pay for the service. (It's true, as the article points out, that even paid support at the big platforms is very poor, but that's because the money involved is still not even rounding error compared to the astronomical profits being made from selling access to their users.)
They can't make their users whole, or even a significant fraction of them, without breaking their business model. That's why they don't do it.
Those who have run meaningful amounts of ads in The Platforms will know quite well that they are very frequently defrauded and almost never made whole.
There is a fair (but tolerable) amount of click-fraud, etc that you basically factor as a cost of doing busines.
I'm quite skeptical of the ad model myself, but I'm relatively sure that "you are not the customer" is not altogether the most salient argument to make in this thread.
> Those who have run meaningful amounts of ads in The Platforms will know quite well that they are very frequently defrauded and almost never made whole.
"Meaningful amounts" to most ad customers is not the same as "meaningful amounts" to the platforms.
Irrespective of whether or not a platform is small or large, can they make their customer whole 100% of the time? Why or why not?