Economics International

Consulting and expert testimony in economics, finance, and statistics

For research and analysis
+1-503-928-6635

Lessons learned from Oregon’s state and local technology failures

Cover Oregon’s fizzled launch has been a high profile disaster for the state. After spending $160 million, Oregon’s health insurance exchange had exactly zero people sign up for private insurance in the first two months.

Now Cover Oregon claims to have enrolled 7,300 people. However, the fine print from the Cover Oregon press release reveals that the suspiciously round number is actually an estimate. In fact, Cover Oregon has not indicated that a single person has paid their first bill, which is the industry’s measure of enrollment.

As with any bureaucratic failure, fingerpointing is an unavoidable stage. Rather than trying to figure out who messed up and when they messed up, we can use Oregon’s history of multi-million dollar software disasters to illustrate some valuable lessons.

  1. Start small. Do one thing, do it well, then add bells and whistles. Google began with a search box. That’s it. It took two years before the company ran ads. Oregon’s software projects try to do too much, too soon, resulting in busted budgets and missed deadlines.
  2. Off-the-shelf is another way of saying, “It works!” Use the resources that have already been developed and tested. Learn the lessons from other states and municipalities. We Oregonians like to think we’re special, embodied in the state’s old marketing slogan “Things look different here.” But, fact is, we’re not that different from any other state in the U.S. and in many cases we don’t need one-of-a-kind software solutions.
  3. Know your limits. Software firms specialize in software, so use them. Don’t try to do it yourself. If you don’t know what you’re doing, then you shouldn’t be overseeing the project. Cover Oregon was urged to hire a system integrator to oversee implementation of the Cover Oregon online exchange. Instead, the state decided to be its own general contractor and found that its staff had no idea what was going on.

A history of budget busting software projects

Oregon has a 20-year history of budget busting software failures. Cover Oregon is just the latest and the first to catch national attention.

  • This year, it was revealed that the Oregon Employment Department burned up $6.9 million on a software project that never worked and expects to spend another $1.2 million to go back to the old system. Even worse, the Oregonian reports that the state may be forced to repay $1 million or more in federal funding because the project is being canceled.
  • The Oregon Wireless Interoperability Network blew through $600 million before it was deemed a failure and laid to rest in 2011. The project was intended to migrate all of the state’s radio systems to one system such that each of the systems could talk to each other. A scaled back State Radio Project hopes to salvage some of OWIN’s earlier work.
  • Many Portlanders still recall the city’s $35 million water billing software debacle. This was followed a decade later by a payroll and financial services software disaster costing $47 million, or more than triple the original estimate. The project was completed more than a year late, and did not include expected functions.
  • In 1993, the Oregon Department of Motor Vehicles began what was projected to be a five year, $48 million project to automate its operations. By 1996, the costs had soared to $123 million. That year, the state pulled the plug on the project, leaving the agency “saddled with useless hardware and programs” according to the Oregonian.

Start small

The Affordable Care Acts requirements for an exchange are straightforward. With the delay in the employer mandate, only individual plans have to be offered on the exchange. The exchange is supposed to use your age and tobacco use status to determine which plans are available to you and whether you qualify for Medicaid. Then it collects you and your household financial information to determine whether you qualify for a tax credit/subsidy.

Instead, Cover Oregon opted for a cutting-edge exchange site that did more than was demanded by the Affordable Care Act. In addition to allowing consumers to shop for commercial health insurance, the Oregon site was supposed to be a one-stop-shop that allowed small-businesses to shop for insurance and to enroll eligible participants directly into the Oregon Health Plan, the state’s version of Medicaid.

Years earlier, Oregon had a similar experience when OWIN busted its budget in part because the state decided to adopt a “highly sophisticated network design.” Portland’s water billing and payroll fiascos were guided by a vision to develop “sophisticated” systems to allow for virtually endless variation to satisfy an expansive range of users and situations.

In contrast to the Cover Oregon experience, Kentucky is known for having one of the most trouble free exchange sites. As reported in USA Today, Carrie Banahan, executive director of the Kentucky exchange relates, “Our system doesn’t have a lot of bells and whistles.”

Google is a $356 billion company that began with a simple ugly search box. It was simple and it was ugly, but it worked. And it worked better than anything else. It took Google more than 15 years to add all the services it offers today. And the search box looks almost the same as it did on Day 1.

Even the moon landing was a series of small steps. Watch The Right Stuff. The moon landing was just one giant leap in a series of steps that began with test pilots, space monkeys, Earth orbits, and so on.

It’s OK to use off-the-shelf solutions

Politicians seem to have a deep disdain for off-the-shelf software: “We’re different. We can’t use commoners’ software.” But, let’s face it. We’re not that special and as a state we’re too small to support spending on a one-of-a-kind software solution. As a state, we’re not that much different from the other 49 states and our local governments are not that much different from the thousands of other local governments throughout the country.

Because Cover Oregon opted for a cutting-edge exchange site that did more than required, it was more difficult for the state to use off-the-shelf solutions. The Oregonian reports that Oracle’s implementation of the Cover Oregon exchange had thousands of lines of custom software code.

The OWIN experience is instructive here. The state opted for a costly custom implementation when off-the-shelf solutions were already out there. A post mortem of the project concluded management “failed to fully utilize the available wireless infrastructure assets of both the public and private sector. Lower cost alternatives appeared to be readily available.”

Portland’s water billing fiasco was driven, in part, by a demand by the commissioner in charge to develop custom software to handle his vision of future water rate reform. An audit of the city’s payroll and financial services software mess found that the original scope included non-essential items including the ability to generate custom reports.

Oregon was also given nearly $100 million by the federal government to be one of the first states to set up an ACA exchange. While there is some pride in being a pioneer, it comes at cost of having no one to learn from. As the old saying goes, “The early bird gets the worm, but the second mouse gets the cheese.”

Again, Cover Oregon could have looked to the disastrous OWIN project for some lessons. A review of the project found the project management failed “to fully utilize cost savings lessons learned from other states.”

Know your limits

Ben Folds sings, “So why you gotta act like you know when you don’t know? / It’s OK if you don’t know everything.” It’s a motto that public sector project managers would be wise to adopt.

USA Today notes that states successful in implementing the ACA exchanges had employed “systems integrators” to coordinate communications between the states’ websites, their Medicaid enrollment systems and with other state and federal systems. These states recognized that they don’t know everything and hired someone who had the necessary knowledge.

The Oregonian reports that Oregon was urged to hire a system integrator experienced with Oracle’s architecture to provide expertise putting together the Cover Oregon exchange. Instead, the state insisted on using its own staff, who seemed overwhelmed with the project and the challenges presented.

Summary

The Cover Oregon disaster is just the latest in a long line of public sector software project failures. These failures provide a useful set of lessons for future projects: start small, look at existing solutions, and hire expertise when you don’t have it yourself.

While the lessons are expensive, one can only hope that future projects can avoid repeating the same mistakes that have been made over the past 20 years.

Originally published at Oregon Business.