The first thing I would say, it's not the annual income of a company that would determine the database system, rather, the amount of data, the manipulation of that data and the accessability of the data.
I created an ERP for a manufacturing company that averages 4 to 7 million annually. The FMP based solution tracks the orders through the facility as well as shipping. Filemaker is not able to handle the machine processes, which run on SQL based systems... but the FMP ERP can extract, via ODBC, all SQL based collected data needed in the FMP solution.
Yes Filemaker can be 'mission critical.' This facility produces hundreds of millions of units per year that can, and are, tracked (down to the unit). If and end user has purchased a box of these items, they can be traced down to the serial number of each unit. And this has been ongoing since October 2000.
With FMP the customers can create, design and submit their orders online and track the orders through the facility, and when shipped, track the order to destination.
The most important part of this FMP solution is not Filemaker itself, but rather the database design, allowing FMP to handle those aspects at which it excels, and offloading aspects to which other systems excel, and collecting all of the data back into the FMP ERP.
Filemaker cannot be all data things to all data needs, but excels at being a great manager of data from a wide array of data sources.
I know that this reply is somewhat 'general', lacking detailed specifics, but I thought it best answers your question as to whether other companies use Filemaker in mission critical, large enterprises.
i concur.. its all about how much data, how many users, and how many transactions you need to process
What type of business is it? You mention parts, is it distribution or manufacturing?
If it is manufacturing how complex is the product? If it's a relatively simple product with a small number of parts like a pens and pencils for instance, FM could possibly handle that. If it's a product as complex as a car I'm afraid that FM wouldn't cut it. Eventually most manufacturers want to do MRP and that takes some heavy resources and the programming required to produce the MRP output is shockingly complex.
The company I work for faced a similar dilemma years ago. Management knew they eventually wanted to take the company public and the CFO felt that Wall Street investors would be very wary of a larger and growing business running on software (FM) built to serve the SOHO/workgroup market. For those reasons we went the pre-packaged ERP route and use FM to fill in where the core ERP comes-up short. It works pretty well.