Well, let's get specific. I used to specify something like 3 million dollars worth of rubber/steel suspension bushes a year.
We had two vendors, both of which were (laughing) full service suppliers.
One of these had access to a good non-linear FEA consultant , who would model each bush as it was designed and flag up the likely stiffnesses, and any durbaility issues, and could also take a stab at any injection moulding problems with the part.
The other did not. he relied on copying similar parts, modified to get the required characteristics, and then progressive mods to his proto tool to get to the required spec.
I would expect a development program with the first vendor to be cheaper in the long run, even if we had to pay directly for the consultant, since the chances are that the first-off parts would be close enough for useful work.
Therefore, if vendor A came in with a cheaper quote, I could trust it, as it was based on better technology. The risk was less as well (proto tools have a 6 to 13 week lead time, and time is our major risk, although cost is our main driver).
Cheers
Greg Locock