Bridge-Slicing
Hiring an FDE from a deployco can cost roughly a million dollars a year.
Crudely, these engineers need to generate four thousand dollars a day of customer value to break even. As such, these engineers need to be highly targeted, highly focused, and operating with an Auftrag (or mandate) from the top.
The key to this is problem selection. FDEs are incredibly effective at solving a very limited set of problems; these problems must be found, and the FDEs manoeuvred into position to tackle them.
This leads to a jagged combination of humility and hubris: humility about the problems you can solve, and hubris about the ones you can’t. I once spent some time with a FTSE 100 company which was experiencing flat revenue growth while managing a complex supply chain. In the course of trying to find the problem that would warrant buying FDEs, we realised that while we had solid plans for their supply chain, the executives were instead focused on bringing down the average age of their customer base. FDEs can’t solve that!
Similarly, there’s a great story about Palantir going to Tyson Foods, having saved the company over $100m in supply chain costs. Mr Tyson says, thank you so much Mr Palantir, that’s great work, but it’s chickenfeed compared to my real problem - which is that half our chicks can’t lay any eggs. Can you fix that?
Expensive engineering must only be used to solve important problems; but some of the most important problems aren’t amenable to engineering.
I was reminded of this while reading about HS2 and the Turweston Green Bridge. TGB is 99 metres wide, an enormous concrete edifice that bridges the cap between some fields and some other fields. Why was it built? Michael Dnes argues that it’s an engineering solution to a political problem.
Nobody was brave enough to argue that it is OK to cut a farm in half to build critical national infrastructure: so the cowardly yet expedient answer was to build a bridge, diffusing immense cost across every British taxpayer while avoiding some more targeted costs on costs on local people, and the planners, consultants, and politicians tasked with building HS2.
To quote Dnes:
“What if, instead, this repeated tendency to over-engineer is actually the solution to failings elsewhere in the system?
…
You can handle these choices in a lot of different ways, using different approaches. In the case of the Turweston Green Bridge, you could take the risk that your senior colleague will flounce out of the cabinet and declare their opposition to HS2. But the option they went with instead was to build a great big bridge to reduce the local environmental impacts and mitigate the impact on the local community.
If you look at the development of HS2, you see that happening a lot. Infamously, the process of choosing a route through the Chilterns turned into a process of choosing a route underneath them, as no amount of digging was more offputting than the idea of facing down angry locals. It was easier to build a £120m bat tunnel than to have a fight with Natural England about how much bat protection is too much.
…
If we see this as a matter of choices between different pathways, the preference for heavy engineering and big structures isn’t just about building things – it’s about building things instead of doing something harder, and following the pathway of least resistance. It is not just that those conversations are hard; it is that the engineering is easy.
…
In the case of HS2, all that extra engineering to solve issues that eluded the policymakers (local buy in; design requirements; cabinet politics; regulatory escalation) ended up as a pile of expensive new bits of construction. And, when the bill was totted up, the calculator caught fire.
….
The engineers are saving the policymakers, the strategists, the regulators, the stakeholder managers and the politicians from problems that they can’t solve for themselves.
Often I think that it would be helpful for policymakers to think about the negative externalities their policies will create. Who is going to flounce out of Cabinet? If nobody is flouncing, you are doing something wrong. All good policies create negative externalities and tradeoffs; want to free the slaves? Someone’s going to be out-of-pocket. Want an industrial revolution? We’re going to have to cut down some trees. We want negative externalities! Give us bullets to bite!
The TGB was an engineering solution to political dysfunction; if you’re deploying FDEs, you should seek to exploit political dysfunction to make your engineering successful.
Your job is to create value; but it’s easier to do that by moving value around than by creating net new value. In order to maximise your chance of success, you should - must! -do both.
A personal example: at one of my clients, we built software that touched both IT and HR; power, responsibility, and pain was split across these functions. As part of that build, we of course wished to create value by increasing power while reducing pain; but we could also distribute power and pain between those functions. The IT team lived in the basement and ate pizza; the head of HR was close with the CEO. As such, we designed our project to take power from IT and gave it to HR: in swapping a small pie for a bigger one, we can slice it differently.
The problem with the TGB was that in redistributing value, the bridge destroyed it in the aggregate.