I know a platform where they actually check the deliverable, before the delivery. Of course, that requires staff, and it’s a more specialized platform, so they don’t need specialists from a-z to check.
However, that’s the only thing that would actually be useful for customers. With those standard tests, there’s no guarantee that the seller didn’t cheat, and no guarantee that the seller himself or herself did take the test. Could have been their dog, this is the internet, after all. (which would be okay if then they outsourced all their gigs to their dog, of course, but typically, they’ll outsource to cheaper workers than their dog)
Too expensive, takes too long for a platform as big as Fiverr, sure, unless the fees would cover it. 20% from sellers + 5% from buyers is a pretty sizable chunk, after all. Still, that won’t happen, of course, but even checking 5 (of course) randomly chosen deliverables for correctness/against the order specifications/whatever applicable in the category/for that gig/order would make much more sense for determining whether a seller is able to deliver what their gig promises, which is what counts, after all, for a customer.
If I knew that Fiverr did random quality checks on actual deliveries of sellers more or less regularly, just to pick a number, 5 per year per seller, or something? Yeah, that would inspire some trust in me.
Knowing a seller who sells a gig as an expert in branding was able to tick the right field “Brand management is the analysis and planning on how that brand is perceived in the market.” as the obvious choice among others that don’t make sense at all for almost anyone able to read (oh wait… but anyway, let’s pretend those tests weren’t broken and there was a correct answer for a moment) in a multiple choice test? Not so much.
But yeah, commonly, people don’t know and don’t invest time to know, so you can will them to trust a label they didn’t know how it happened quite easily, I suppose.