I'd think about it less as trust and more as risk. Specifically, risk arbitrage.
The originating buyer wants quality work, probably as a one-off, but doesn't know or want to know how to find someone like that. Everyone is high risk.
A matchmaker company in the middle has ongoing relationships with the end contractor, knows their work is decent, and provides a framework / legal liability / insurance on top. The contractors they know are lower risk.
So the matchmaker can charge {full cost of high risk - slight discount} while knowing they're actually only taking on {lower risk}.
Where it seems to go pear-shaped is when the matchmaker gets too large and can no longer individually vouch for their contractors (e.g. IBM Services and globally integrated service companies).
Trust is an appropriate metric to consider. Often, the matchmaker company isn't being evaluated on material criteria like performance; they're contracted because there is some preexisting relationship (not always direct) between decision makers at the two firms, or because they hold a certain reputation. You could say that risk to firms are lowered because liability is transferred, but that's sort of orthogonal to my point that the point of existing from the point of view of the middlemen is not to trade in risk but it trust. Risk is a commodity, trust is scarce.
The originating buyer wants quality work, probably as a one-off, but doesn't know or want to know how to find someone like that. Everyone is high risk.
A matchmaker company in the middle has ongoing relationships with the end contractor, knows their work is decent, and provides a framework / legal liability / insurance on top. The contractors they know are lower risk.
So the matchmaker can charge {full cost of high risk - slight discount} while knowing they're actually only taking on {lower risk}.
Where it seems to go pear-shaped is when the matchmaker gets too large and can no longer individually vouch for their contractors (e.g. IBM Services and globally integrated service companies).