Every federal government contract that has even a hint of favoritism is litigated to all hell, where lawyers comb through everyone's emails going back years to dig up any dirt they can find. If it was a big problem, we'd know about it.
Nepotism is alive and well at the state and local level. This is partly because state and local governments are far less sophisticated than the federal government in terms of process and people, and also because there is not a strong culture of litigating contracts.
Big organizations have transaction costs associated with them. That's kind of inevitable. But the question is whether eliminating a few billion here and there in transaction costs is worth forgoing potentially huge benefits from public/private partnerships.
I'd say overall it is beyond what an early stage startup can do (non-dod-focused), without a great partner. Selling products is a lot easier than selling ongoing services, but the profit is all in services. A lot of the difficulty was due to it being a classified contract (the work itself wasn't, but the work locations and interoperability were). It's probably feasible for a successful small business with a full time person, particularly if the business is set up to go after government from the beginning.
Big companies aren't cost-free as clients, but there's less "if you do X wrong, you could go to jail" (which essentially never happens absent willful fraud, but still). The citizenship requirements also make it really hard for tech companies.
Remember the fun with Countrywide and the sweet heart loans? How some politicians or their friends do so well in the stock market or get into IPOs at a level not afforded to the public. If not family is it not worse when those who support your campaigns are the ones more likely rewarded with the contracts?
So yes, while direct attempts may be protected against nothing stops the train of indirect benefits.