Fast Tech, Slow Change
It may seem like technology will speed processes and improve outcomes, but that isn’t always the case. Here’s why.
In January 2009, we assumed we struck gold. After three years of experimentation we finally reaped real results in our efforts to spread public service innovations across cities. Billions of people around the world got subpar public services, stifling their ability to thrive and, many times, furthering inequities in communities, and we found a way to spread what worked, faster. Our fix: using public procurement, the function governments use to buy goods and services, as a strategic problem-solving tool. Over the next decade we worked to refine this methodology and demonstrate its impact, and eventually tried to productize and scale it with the help of technology. However, we soon learned that what looks good in theory may not stand up in action.
A History of Inequity and Slow Public Change
Public innovations typically take an average of forty years to gain wide adoption in society, making social change a slow process. For example, shared public bicycles — first invented by citizens of Amsterdam 1965 — took more than 30 years to catch on. It was only in 1998 that what today has evolved into CitiBike-like programs started showing up in a recognizably modern and technology enabled form, and the path forward was very slow. By 2004, a mere 13 cities had a bike share program. By 2010, that number swelled to 400. Today, bike share is a highly visible innovation that mayors like to replicate, but it still took what many would say is far too long to propagate. Unfortunately, when you look at social innovations in areas like early childhood development, nutrition, education or policing the adoption rate of new innovations is even slower. We set out to change this.
Public procurement came up on our radar because one of its hallmarks is that it is designed to source the best solutions in the market. We found that by teaching cities to open up their procurement programs to other workflows, they could more reliably discover novel ideas such as bike sharing across all public services. With time, we were able to demonstrate some compelling numbers about the opportunity for improvement. For example, we found that the average municipal procurement team had knowledge of less than three percent of the solutions on the market. That was one of the reasons why public bike systems were slow to scale. Buyers simply didn’t know the full scale of what was out there. In addition, even when they asked for new ideas, government procurement teams received an average of just two bids per contract opportunity.
We discovered that the best way to fill this knowledge gap was to get more participation from the market. Why settle for two bids when our research showed that hundreds of vendors could solve the problem in more creative ways? If we could get more ideas and more vendors into the procurement process, we could put trillions of dollars of wasted municipal public procurement funds to better use.
Teaching and Showing — the Secret to Success
Our approach to this opportunity became widely known as problem-based procurement — procurement challenges or calls for innovations. We helped cities replace detailed specifications with problem statements. In the case of Philadelphia, for example, the process of problem framing helped the city realize that — in order to reduce its high traffic fatality rate — they needed to forgo buying smart cities technologies. They could make the streets safer by accessing data and doing analysis to understand how crashes happen. Once they did that they could make more nuanced interventions. The end result: Instead of spending millions on technology they hoped would fix their issue, they spent $30,000 to get the detailed insights. We helped them come to this solution via consulting services with a strong focus on capacity building. We guided city teams through the various steps, augmenting their work with sophisticated market research and vendor engagement. And we’ve done this many, many times for a wide variety of public organizations.
Our results speak for themselves: 135 governments in 35 countries including cities such as New York, San Francisco, Paris, London, Norfolk, Va., Long Beach, Calif., Mexico City, Rio de Janeiro, Fukuoka, Japan, Da Nang, Vietnam, and Melbourne adopted our approach. Our procurements received an average of 24 proposals (instead of just two), bringing many new ideas into government. Vendors changed, too. Cities got bids from a variety of service providers. In fact, 98 percent of all bids coming in using this process were from small, minority, and women-owned businesses. Previously, it was the same large corporations that had a stranglehold on the bidding process. After working with us, governments reported routinely finding better answers to their needs and that they saved an estimated 30 percent per transaction. And, much like the adage of teaching someone to fish rather than just catching it for them, we trained more than 2,000 public servants to carry out this work on their own. And, always a good sign, we had an increasing number of copycats around the world, replicating our approach.
But despite all these successes we were concerned. How would we ever get this lesson out to a market of 551,000 municipal governments, which collectively conduct tens of millions of procurements every year? We all — our team, clients and funders — believed that we should build technology that would make our approach more accessible and lower cost. It seemed that the only way forward would be to productize our approach, something that filled us with both excitement and anxiety. We had another goal, too. In addition to multiplying our impact, we also hoped we could eliminate some of the pain of serving governments as consultants. It would be wonderful to banish the constant delays in projects that strained our nerves and challenged our financial sustainability.
Given that we had revenue from governments, we saw it as a logical step to build a Software as a Service (SaaS) offering that would help city halls across the world lower costs and effort by more than 90 percent compared to consulting. Our investors and funders were concerned about a hybrid business model — offering both consulting and a SaaS offering. They demanded that we choose one or the other, so we became a tech startup and pulled out of all our consulting arrangements, going all in on productizing. We decided we would measure our success by growth in revenue and transactions.
How Far We’ve Come
Fixing public procurement is a distinctly Slow Lane problem. There is massive public interest but it is highly complex, involving bureaucracy, culture, mindsets and human behaviors. Our clients seemed eager for us to productize, but over time we began to realize that many of our fundamental assumptions were flawed. For one, we thought that if we had worked offline with 135 cities and 2,000 public servants, we’d be able to offer it to even more governments at a fraction of the cost and effort, exponentially opening up the market. That, it turned out, wasn’t true. The reason: Many Slow Lane problems don’t suffer from logistical or financial bottlenecks but from a lack of readiness. Looking at the problem under this different light made us realize that, by definition, we had probably captured the what was for us the whole market. As a result, our growth was much slower than expected, and we were making a fraction of the revenue we anticipated. We had also underestimated the benefit of human contact in the process. The richness of our relationships with city employees and the leadership function we held during our consulting work could not be duplicated in a SaaS offering. It’s not surprising that we missed this. It was easy to overlook since we used a highly standardized playbook to work with cities. It was automatic for us. We didn’t realize that our process also featured empathy, encouragement, and human support that helped our clients complete the process. Yes, the tech product we were offering had all the same functions and workflows, but lacked the important intangibles of human contact and relationships.
Another lesson we learned was that we were serving a need with technology that, for our customers, only came up two to three times a year. This meant the tool sat dormant most of the time and that, when it was time to do procurement, users either simply forgot they had it or forgot how it worked. This meant we found ourselves in the impossible position of having to try to figure out when someone might need our tool and remind them to use it. Technology, for those users, was simply not the right answer. All three of these lessons led to new product ideas, but the rate of adoption remained very slow.
It is hard to get a product right in the best of circumstances. In the case of a public interest opportunity like putting public procurement to better use, we had to recognize that it’s a problem that cannot be solved in the lifecycle of a startup. We simply cannot force traction on a complex issue that will always move at a slow pace, especially if we are pursuing change rather than just servicing the status quo — something that the govtech sector generally does. As I come out of this decade of learning, a few new principles look promising going forward.
The first is that we need to find ways to prepare ourselves to be resilient for a journey of change that may take decades. This includes having funders who don’t squeal at building open, societal platforms that may take years to gain the kind of traction that most public innovations only see after 30 years or more. Secondly, to do this, we need to embrace collaboration in entirely new ways. Over the years technologists’ competitive mindsets have rendered us ineffective by carving out unique selling points and differences over our common mission for change. The journey to success in a Slow Lane mission will not be a winner-takes-all scenario, but one in which we spread the burden as widely as possible to stay in the game. Third, we need to find ways of placing technology into a context that also requires human nurturing and engagement. For this we need a unifying vision and mission, but also an appreciation by all stakeholders that every contribution is equal. And finally, we need to align around models of governance and openness of data and information that should be the standard of public interest technology missions.
In the current funding and collaborative environment it is too hard to pursue a path of true openness and shared production and leadership. But it has become evident to me that, as a design principle, we need to find ways of enabling the openness of critical public infrastructures and movements for change.