Some years ago the city of Stockholm initiated a digitalization project. Named "Skolplattformen", its purpose was to provide a digital platform within education. The project is unfortunately a catastrophic failure and with over a billion SEK later the city has decided to replace the whole thing. With yet another platform. As a tax payer this is not great to say the least. Such complete failures are still far too common both in the public and private sector. Let us explore some of the core problems and what successful digital transformation might require.
To keep it short, here are some of the recurring issues often reported about the platform project:
- Complete failure to achieve the intended results and actually making things worse
- Sky-high lifetime costs
- Hiding issues such as serious security vulnerabilities
This is "digital transformation" at its worst and a schoolbook example of what not to do. Pun intended.
The goal was to provide operational support for teachers, administrators, pupils and parents. Reduce the workload, make things simpler and save time. The usual nice sounding words. Somehow the project managed to achieve the exact opposite results. How is that possible?
Procurement and contracting is the enabler of successful digital transformation
Large public sector projects are often delivered by the same suppliers with questionable track records. When the buyers don't have the right competence, they are at the mercy of the suppliers. Unfit procurement dooms these projects to mediocrity or worse. The current contracting is broken and must be changed. Lowest bid can't be the dominant criteria.
Procurement's primary task is to weed out the uninteresting suppliers. How? A good supplier will show a history of successful results. Likewise they are happy to provide convincing evidence to back up their claims about their suggested solutions.
"Customer collaboration over contract negotiation" is no excuse for lousy contracting. The most effective way to identify strong suppliers is to have them sign a "no cure, no pay" type of contract. You need to incentivize the suppliers to do the right thing, not just take your money. Design the relationship into a partnership. This is nothing new and you can do it even in the public sector.
In a nutshell: The people with the money must start asking the right questions to set the bar at the right level. That is how we can start seeing industry-wide improvements and higher success rates. If they don't, there's only one thing to do and that includes a mirror.
Control over the wrong things does not contribute to successful digitalization
All projects exist only for one reason. We invest in different initiatives only for one reason. They must improve how well a set of needs of a set of stakeholders are being met in a given context. As an example of one such need - Spotify makes it a breeze to find the right kind of music for any situation.
What is it that we do to help the projects succeed? We deploy controls. What is it we try to control? Tangible things such as delivered features and the amount of work done. Often this is command-and-control in disguise. The problem is:
They tell us nothing about the results achieved. At best they are a proxy, at worst a distraction and a hinder. For "Skolplattformen", they tell nothing about how much administrative workload we have reduced.
In any non-trivial projects the results are what we need to have control of. That is the only way to avoid failures. The solutions must be able to evolve in small steps whilst taking you closer to your goals.Your processes and the culture need to help you all the time focus on results. Everything else is noise. Even the proxies.
Unclear goals - The enemy #1 of successful digital transformation
To be able have control over the desired results, they need to be as clear as possible. How would you go about ensuring satisfactory results if they were unclear to you? How do you know when you have simplified the teachers' work enough? What kind of simple is that? How do you know when your system is secure enough?
To design optimal solutions we need a clear understanding of the solution space. For that we need clear objectives as they fence in the available space for solutions. If the desired results are unclear this becomes impossible. You need to get rid of unnecessary ambiguity. Don't believe me? Let me paraphrase one of the many frustrated parents who must use "Skolplattformen":
"Using the app makes me regret having children in the first place"
There's no better way of getting rid of ambiguity than quantification. It is an extremely effective way of communicating difficult subjects clearly. Putting numbers on intangible things cuts through the nonsense like nothing else.
Getting control over your critical objectives starts with quantifying them. If they are critical, you must get rid of the ambiguity. If you only take one thing from this article, it's this. Study people like Tom Gilb and learn how to truly get these goals clarified and under real control. Anyone can learn how to quantify objectives, even non-economical ones, in a short time.
Confusing Means with Ends
"We need to go all-in on digital!" echoes the command from the boss's room. While this might be a great idea, it is a strategy to achieve something. What that something is you must clarify as already stated. The more interesting question is: What if there was an even better strategy to achieve the same thing? Would you still cry "Digitalization!!"? If so, why?
The point is nailing down the strategies too early without considering their efficacy is dangerous. It's excellent if you have convincing rationale and evidence why a strategy is a good one. If you don't and still want to constrain the design process by requiring it, we are on thin ice.
Have you considered your current services might be generating tons of failure demand? What if you could get rid of that without no digitalization? Software is always a liability so make sure your default stance isn't more software.
Besides, who keeps selling the need for more digitalization? Exactly.
Closing thoughts
Successful digital transformation relies on successful software-intensive product development. It can produce wildly different results. For success the basic competencies all revolve around success itself. They have nothing to do with workshop facilitation, 1-on-1s, user stories or certifications. The key is understanding the nature of the work and the requirements it puts on how the work itself should work.
We all can learn much from history. There are plenty of people who've already learned the hard lessons, we don't have to do the same thing. A quote that I keep coming back to almost weekly is from none other than Dr. Deming:
The knowledge how to avoid these catastrophic failures isn't protected by secret societies. It is out there. It is completely unacceptable in 2023 that we still encounter such complete collapses. We need to set much higher expectations on ourselves and our peers. That is the only way we get out of this ridiculous situation.
Especially when the taxpayers' money is at stake it should be a crime to repeat the same mistakes. The large consultancies and their shareholders might object to that of course. It might be an uncomfortable way forward but in the end it's a win-win proposition for everyone.