What if the Project Is a Tattoo?
A Blog Post about Business Cases for Major Projects
A few weeks ago, I published a piece on making business cases better. It included turtles. All the best posts should contain turtles, but - sadly - most don’t.
Anyway, after I published, several of you got in touch with similar thoughts. I got this great piece of feedback:
Your writing style is very engaging.
Thank you!
I’ve made a note to use the ‘haircut vs tattoo’ analogy from now on! Not sure how/if the ‘just do it’ model applies to big projects though, they are more ‘all-over-body tattoo’ decisions!
It’s a fair challenge.
You can’t “test and learn” your way into building a new rail line. You can’t try a tram system, see how it goes and undo it if the reaction’s poor. And no one’s going to let you iterate your way into tunnelling under London.
So how do we make good decisions when the stakes are high, the investment irreversible and the consequences play out over decades?
Let’s take that challenge seriously - because it deserves to be taken seriously. And let’s break it into two parts: costs and benefits.
Costs:
When it comes to major project estimates, we can take the Bent Flyvbjerg approach described in the previous post. This works just as well for massive projects as tiny ones. Indeed, it’s where he started.
He discovered that 90% of large projects go over budget. Rail projects typically overshoot by 45%, bridges and tunnels by 34%, roads by 20%.
But the solution he proposes is simple: don’t estimate your project from scratch. Take the outside view. Look at a reference class - similar projects delivered elsewhere - and start from there.
This is exactly what we explored in the last blog post when we talked about “superforecasting.” Great forecasters start with the average of similar things, then make small, reasoned adjustments. They don’t build a complex cost model from the bottom up.
So when it comes to forecasting costs on major projects, the tools already exist. The problem is that we tend to start from scratch each time.
Benefits
If the cost side can be handled with data and discipline, the benefit side is much harder.
The truth is, we can’t know what the benefits of major infrastructure will be - not over decades. And when we pretend we can, we often end up misleading ourselves and others.
Take my two local tube lines. To my east is the Central line. In the 1930s, when it was extended into what was then Essex, no-one could have begun to forecast that its primary purpose would be to take financial services workers to Stratford to change onto the DLR (running over a repurposed freight line) into the London docks to work in banks and investment firms.
There’s no model in the world that could have predicted that. So had BCRs existed then, they’d have been a nonsense.
To my west is the Victoria line. It’s the second newest tube line at just over 50 years old. When it was planned, it was forecast to carry 50 million passengers per year. It has - so far - peaked at 302 million in 2019. Hmmm - a tad out.
It will eventually get rather crowded in here, Ma’am
And yet, in the modern business case world, we demand that someone, somewhere, assigns a Benefit-Cost Ratio to major projects. We ask for a single number to summarise fifty years of unknown, unpredictable social, economic and spatial change.
That’s not evidence-based planning. That’s darts.
If we want to make good decisions, we need to return to something that is strangely unfashionable in the modern public sector: well-argued, well-written words.
Describe the problem. Describe what success looks like. Explain the rationale. Make the case.
Because a nonsense number is less convincing than a paragraph of well-chosen prose. (You could argue I would say that on this, ahem, blog).
You don’t have to take my word for it.
When I interviewed Jonny Mood, Director at the National Audit Office, for my podcast, I asked him how we should think about value for money in major investments.
His answer was crystal clear:
“If you're entirely making your decision on a BCR, you're missing the bigger part of the picture.”
This is the UK’s public spending watchdog saying: don’t let the number do all the talking.
He’s right. Some things just can’t be boiled down into a spreadsheet. That doesn’t mean we abandon rigour - it means we apply it differently.
Use data to understand costs. Use words to articulate value. And stop pretending that false certainty is more credible than honest uncertainty.
If you still don’t believe me, then let me present to you the most data-driven company on earth: Amazon. This is a firm so successful in its investments than it has gone from not existing when I was a child to on-and-off the most valuable company on the planet.
They are rigorous that proposals in Amazon are made long-form on six page written documents, which are then read by everyone in the room in person, in silence. I have described this approach as part of my “top 20 ways organisations can move faster”. Investments in Amazon are made by argument. It works for them - it can work for us.
The Real Problem: Our System Doesn’t Want That
Here’s the deeper issue: even when a local authority or transport body does articulate a clear strategic rationale, it often doesn’t matter. Because by the time it gets to Treasury, it’s already been reduced to a number.
And the Treasury aren’t even looking at the number itself - they’re trying to guess what number the Office of Budget Responsibility will apply to the number.
I mean, seriously, this is mad.
But we don’t have to do it this way.
We treat this process as if it were some natural force of British life, like the weather or England batting collapses. But it isn’t. The Treasury is a human institution. It was created by people and it can be changed by people.
Of course, that’s easier said than done. (If you want a great conversation about how hard it is, have a listen to my podcast episode with my utterly brilliant Freewheeling Associate and former senior manager at the Treasury Katie-Lee English.)
But here’s why I’m slightly more optimistic than usual: the current Transport Secretary actually sounds serious about devolving power. More than any of her predecessors, she seems to understand what it means. She worked at TfL. She knows what local empowerment can unlock.
Now she needs to do it - and be backed politically to do it.
Why It Matters: Britain Is the Outlier
The UK is the most centralised large country in the OECD. 96% of taxes are raised by central government. Even where taxes are theoretically local, like council tax, caps and rules are still set in Whitehall.
Sam Freedman, in his excellent book Failed State, puts it perfectly:
“Concentrating power at the centre of government, and destroying state capacity outside of it, while at the same time massively increasing the scope of what government covers, is a core reason for our policy paralysis.”
He’s right.
We don’t just need to fix how we write business cases. We need to fix who gets to decide what’s in them.
Local communities should be empowered to define what is valuable to them.
How Do We Solve the All-over-Body Tattoo
Big projects are hard. The stakes are high. The tattoo is permanent.
But we can still make good decisions - if we start with realism, speak in plain English and trust the people closest to the impact.
Because the future’s still uncertain and will remain so. Accepting that is the most important point in any piece of planning.
But if you’re a developer of a major project:
Use reference class forecasting to estimate costs.
Use clear, strategic narrative to describe benefits.
Stop demanding numerical precision where it’s impossible.
And if you’re the Government:
Give local and regional governments the power - and funding - to plan their own futures.
Trust them to define the outcomes that matter.
Reform the system so that not everything has to pass through a single model in SW1.
If you’re the Treasury:
Get out of the way
I’ve previously described my experience of the new line from Oxford to London that I was closely involved with at Chiltern Railways. We delivered this on-time, on-budget, and it’s a great example of doing things right. Have a read.
