In most countries there are plenty of examples of the State (and it’s associated public organisations such as Municipalities, Regions, etc.) buying IT solutions that ends in a disaster. In Denmark we have a pretty long list of such failed, publicly owned IT projects.
Everybody that has been part of IT development knows, that innovation doesn’t always work, and sometimes it just goes wrong. That has to be accepted as a starting point. However, there is something else that our politicians and public servants needs to be aware of before they launch tenders for new IT systems: the purchasing and development methodology is often wrong. It is not only bad project management that causes the issues, but also the way the projects are laid out from the beginning. This causes not only bad use of our public money as well as bad solutions for all of us, but it also undermines the objective of building tech clusters, for example in relation to welfare technology. In Denmark, where I reside, our last left wing government wanted welfare technology to be a Danish export success, and so does the current right wing government, so the local industry receives government grants. But grants is not the key, customers is key.
(NB: I have invested in a few companies that sell to the Public sector – e.g. Sekoia, Famly, Peergrade as well as my own construction related company, GenieBelt – but I hope my arguments clearly shows, that my agenda is not entirely selfish).
To understand my point about how the public purchasing processes are hurting us all, then we need to look at how the projects that the State et al are buying are normally organised. They will almost always use the “waterfall development process“, i.e. first a very loooooong specification is made by analysts, then this specification is shipped over to the developers who then start coding for a few years. The actual development process will be structured in some phases, but fundamentally progress is seen as flowing steadily downwards. It is sequential and not iterative. Several years after the first notes on the specification it is finally delivered to the users.
More and more often this is not the right approach for developing systems. There are multiple problems, the first of which is that often the detailed specifications are totally out of tune when the project is finished many years later. Both user needs and technology has moved on.
In the tech startup world this is pretty obvious. After the good old .com days we have increasingly used more agile development methodologies with emphasis on iterative processes. We will try to find the MVP as quickly as possible. We look for the UX and feature set that as a minimum gives a path ahead in terms of how to achieve the core benefit of the product. Often the hypothesis is changed multiple times in the process, my own journey with GenieBelt is a good example of this, where we on a monthly basis would ask our selves, our data and our beta customers if we were on track. After the MVP we will move on, and iterate towards a commercially viable MVP, etc. etc. Always iterating, driven by monitoring actual engagement metrics, always prioritizing the important stuff and cutting out all the nice-to-have clutter. It is impossible to wait 5 years and many Millions of EUR/DKK before we know if we have something that works.
There is a lot of advantages of this model, e.g. we will quicker get to the core of the problem that needs to be solved, and we will quicker start to challenge various hypothesis on what the solution should be. Less risk of spending years of coding on something that makes no sense. Furthermore, the ultimate solution will in general be better, because the specifications are constantly being updated based on real user engagement, and not what an analyst believed some years ago. And that is also the way we should build welfare tech solutions to the public sector.
Why are not more public tenders based on this development model? I don’t know for sure, but there is some rigidity in the system. Also, the big and well-established players in the market (which in Denmark would be KMD, IBM, CSC, etc.) has an interest in keeping the public sector in this development cycle, since their entire business system is tuned to these kind of processes. But if it doesn’t change we will stay where we are: huge specifications that are partly irrelevance, year long projects that are delayed and too costly. The big guys should learn from the small innovative startups to earn their right to keep on serving our public sector.
But the problem is actually even bigger. In our part of the world (Denmark/Scandinavia/Nordics) we have a unique expertise in welfare technology. This can be used to create internationally leading tech companies, that based on the (small) home market export our state of the art solutions to the rest of the world. But one of the pre-requisites of this is, that we actually have a home market. A market where we fine tune the commercial PMF, where we start to build scale and recognition. If we can’t offer our startups a home market, then it is very-very difficult in the welfare tech space to build an internationally leading company. Since the public sector has monopoly power on big parts of the welfare system, then we need politicians and public servants in charge to recognize this issue, because otherwise we are missing a significant opportunity – on top of making it worse for our own public sector.
A concrete example will make it obvious what we are dealing with. 3 years ago, the newly founded company Famly launched a platform for communication and administration in the daycare/kindergarten/nursery sector. PMF was found, the team started selling after a while, and now they have sold to more than 200 institutions in Denmark. This summer Famly received a couple of Mio. EUR for their growth in UK, DE etc. There are other players in the market, but Famly is the Scandinavian company moving the fastest due to a super slick product with some of the highest engagement metrics I have ever seen.
All looks pretty good, because there are another couple of thousand institutions in Denmark alone, which will be a solid home market to support a growth journey where tens of thousands of these customers could benefit from the platform in the coming 5-10 years. But then something happened last year: a public IT company (Kombit) decided that it was about time that the Danish municipalities should have a solution for their nurseries. Instead of asking the existing providers on the market, they instead made a looong specification, and have recently put the project up for tender. In total the project is planned to last 5 years, and costs several hundred millions of kroner (7,4 DKK = 1 EUR). Almost all Danish municipalities have signed a contract supporting this project, even before the specification was finished. And I am sure, the municipalities did it without really checking the market.
So now this market will freeze, because the municipalities (which are 90%+ of the customers in Denmark) will wait until 2020 (!) before they can get a solution they need TODAY. And then they will pay a lot of money and get a product I bet is worse than if they had asked existing providers to start delivering TODAY. Plus of course a monopoly has been created for this kind of solution since all the municipalities decided they would be part of the tender.
On top of this, the existing startups are excluded from the tender, since it is stated that in order to be considered as part of the tender, then a provider needs to have DKK 250M in revenue, equity of 200M and be profitable over the last three years. They might as well have written: “all startups, fuck off!”. Sorry my bad language, but I am a tax payer and voter in this country, and it is demotivating to see this unfolding.
I dare to argue, that if the municipalities today choose to buy from the existing providers, then they would not only get a solution immediately that would provide clear value, but they would also ultimately end up with a better solution by 2020 at a lower cost.
The public sector in the Nordics is about half of our economy. If the public sector does not design purchasing processes that are supporting the development of tech clusters, then we are shooting our selves in the foot. In some sectors this is more important than others, in welfare tech it is totally essential.
UPDATE: just to be clear, there are plenty of good examples, where the public sector in Denmark/Scandinavia works well with small, progressive companies. In GenieBelt we have several forward looking customers among the municipalities as well as government agencies, and Sekoia could not become a success unless many municipalities/kommuner took a bet on the solution a couple of years ago. The point is, we need much more of this!