Is the Lean Startup approach applicable to smart city projects?

Aug 4 2015

The lean startup
Last week I read Eric Ries business bestseller “the lean startup: how constant innovation creates radically successful businesses”. Despite this somewhat grotesque title, I found it a great and revealing book on how startups can be more successful. His main argument is that startups –but established firms as well- can achieve faster and better results if they adopt a more scientific approach in their innovation efforts. Too may startups and innovation projects fail because they build things that people don’t want.
Rather than developing a full-fledged product or service and then try to sell it, Eric Ries -a serial entrepreneur- advises startups to make a “minimum viable product” (some sort of beta version). After that, they should systematically find out how to improve it, by involving clients –preferably early adopters. This prevents you from building a great product that nobody wants. The term “validated learning” is central in his book. Companies/entrepreneurs must learn not just to act on hunches or great visions (Ries rightly calls them assumptions) on what customers might want. Rather than shooting in the dark, they must carefully measure how customers react to (small) changes in the product or service, and learn from that. Early customer feedback is central. To learn what works, startups must run series of experiments, and draw the right conclusions from that. Why? Because the market in which startups operate is very uncertain and unpredictable. You cannot know beforehand what customers want (and typically customers don’t know what they want themselves either). Traditional market research is not helpful in this context. You rather learn by doing, experimenting systematically. Action research! Ries claims that in this way, innovation becomes manageable, and lots of waste is avoided: if you work like this, you don’t spend months on building something that you might think is perfect, but that in the end nobody wants.

Smart City projects often fail
I have seen many smart city pilot projects fail, too, or lingering in the valley of the living dead. Cities worldwide do many experiments with new technologies, especially in renewable energy and information technology (smarter traffic information, new types of energy provision, smart grids, electric car infrastructure, sharing economy etc). There is a lot to gain here: cities could become more sustainable, liveable and efficient with the help of these new technologies. And for companies, this is a huge new market (worth $1.5 trillion according to Forbes). But too often, in spite of initial enthusiasm, smart city pilots don’t scale up, and slowly fade out. They don’t evolve into a sustainable business that can continue to make a difference without government subsidy. Why? I think because what lacks is a scientific and systematic approach. Most smart city projects start from a hunch, based on some assumptions about why the project might work and why it might scale up after the pilot stage. But are these assumptions correct? What lacks is rigorous validated learning; projects typically fail to collect relevant metrics on which decisions can be based and assumptions can be tested; customers are insufficiently involved during the process. As a result, it often takes too long before it is understood if a project will take off or not, and whether a pivot or turnaround might be necessary.

Like the tech startups in Ries’ book, smart city pilot projects operate in an unpredictable and unstable environment. Beforehand, it is hard to predict if pilot projects in the city will be successful, because the technology is relatively new, there is no established market, and the customers often don’t know what they want. So, experimenting and learning are important to find out, and intense customer involvement is essential. It’s a pilot, after all. A traditional linear innovation approach (with engineers developing a smart solution, and then see if it work or not) is unlikely to work. But this is exactly what happens too often.

In a way, smart city pilots are more complex than the startups that Eric Ries discusses in his book. Here it is not a single company selling something new to customers. Typically, there are more partners involved in smart city projects, public and private (tech providers, utilities, housing corporations and municipal departments are the typical partners), with different organisational goals and incentives, and smart city projects often have a mix of business and societal aims. This makes it harder to act fast, and to admit failure (nobody likes to take the blame) and to change strategy if needed; also, different partners might perceive the progress of the project differently. In many smart city projects, the team is already happy when they have the budget and the “go” from the parent organisations.

A more systematic approach is fruitful, and the university can help
Most smart city projects show a striking lack of systematic working. They have implicit assumptions that need to be tested but are taken for granted; they lack decent metrics to track whether the project is heading in the right direction; there is often too little early end user involvement, and it is often unclear under what conditions the pilot can be scaled up. There must be more experimentation, and validated learning. Much is to be gained here. I see great scope for collaboration between smart city organisations (many cities have platforms where smart city projects are coordinated) and universities in these fields; Research teams and students could be involved early on in smart city projects. They might help to make implicit assumptions explicit, to test them, and to set up experiments in order to build something that end users really want, to define metrics and collect/interpret the data with the team. Eric Ries’s lean startup method is a good starting point to start thinking about how to increase the success chances of smart city pilots, how to avoid waste. Urban innovation must be managed better!