Why can no one seem to get it right?

A generation ago, the IRS spent $4 billion developing “modern computer systems” an official later conceded “do not work in the real world.”

Flash forward to 2016 – skip over healthcare.gov and a host of others – to find the California State Bar is in the midst of rebidding a contract for computer services after the original contractor “floundered for years” and produced nothing but “seriously flawed” work product. And while it is fortunate, in this case, the Bar did not pay anything to the contractor, the backlog the systems were intended to address has only gotten worse.

To be fair, the imperatives of building a computer system for governments are far different than they are for the private sector. As we point out in our story on Page 1 and the accompanying sidebar, the demands of government customers are nearly impossible to meet in the tech world. The IRS can’t launch in beta during tax season, and as we learned with the rollout of the portal for the Affordable Care Act, when you go public with a health insurance enrollment platform you’d damn well better have all the bugs worked out.

That’s a problem no unicorn ever had to deal with. If there are bugs, fix them on the fly. If the model doesn’t work, pivot. If no one wants a platform, make a server box. As long as the investors have an exit, it really doesn’t matter.

With the cultures so different, it may not be realistic to expect private contractors to be the ones to build these taxpayer-funded systems. And yet, it’s also unreasonable to expect to have that expertise on the government payroll, especially when there are unicorns to be chased.

Perhaps the government should seed its own VC firm, one focused on developing systems for public entities. Fund companies that develop government solutions, and when the bugs are worked out sign a long-term maintenance and support contract and exit based on the future revenue stream.