First-Principles Disruption
Rebuild a broken process from fundamentals instead of inherited convention
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 96%
First-Principles Disruption starts by defining what a process must accomplish, then questions every inherited step rather than optimizing the routine as given. Blomfield illustrates the mechanism with an early consulting assignment: instead of manually counting and grouping items on a large jewellery website, he wrote an Excel script to collect and tally the data. The technical solution produced the requested result far faster, but it conflicted with the firm's hourly-billing incentive. That tension matters: a better process can still be rejected when it threatens the surrounding business model. The framework therefore ends not with automation alone, but with a comparison of outcomes, constraints, and incentives so the operator can distinguish a genuinely necessary step from convention that protects an incumbent system.
Origin
Blomfield links this approach to his habit of seeing a process, asking why it works that way, and looking for a much faster alternative. Host Steven Bartlett identifies the pattern as first-principles thinking. Extracted from The Diary of a CEO.
Core principles
- 01A familiar process is not necessarily a sensible process
- 02Start with the required outcome rather than the inherited workflow
- 03Use technology where it removes avoidable effort
- 04Test whether an apparent constraint is real or merely commercial convention
How to run it
- 1
Define the required outcome
Describe the result the process must deliver without referring to its current steps.
Pro tip Use an observable output such as an accurate tally rather than a vague goal such as efficiency.
- 2
Expose inherited assumptions
Write down why each current step exists and whether that reason still applies.
Watch out Do not assume a familiar step is a real constraint merely because everyone follows it.
- 3
Rebuild from fundamentals
Design a direct path from the available inputs to the required result, using automation where it materially reduces effort.
Pro tip Look for order-of-magnitude improvements rather than cosmetic gains.
- 4
Run a bounded prototype
Test the alternative on a small, reversible task and compare its output with the incumbent method.
Watch out Verify accuracy as well as speed.
- 5
Check the incentive system
Identify who gains or loses if the new process replaces the old one. Decide whether adoption requires changing incentives as well as technology.
Pro tip Treat resistance as information about the business model, not automatic proof that the idea is wrong.
In the wild
Asked to count and group roughly a thousand items on a jewellery website, Blomfield wrote an Excel script that scraped and tallied them instead of doing the work manually. The output arrived in about an hour, but the consultancy preferred the manual route because it could bill more hours.
→ The prototype demonstrated a large speed gain while revealing that hourly billing, not technical difficulty, protected the old process.
Common mistakes
Automating before defining the result
A faster workflow is useless if it produces the wrong output. Define and verify the required result first.
Ignoring incumbent incentives
Technical superiority does not guarantee adoption when the existing model benefits from inefficiency.
Is it for you?
Best for
It is best for operators who can inspect a workflow closely and test a simpler alternative.
Not ideal for
It is not ideal where a step is legally mandatory or where changing it creates unmanaged safety risk.
From the transcript
“you sort of start from physics and build build up from there really”
From the episode
Monzo CEO On Death Threats, Depression & Digital Banking Wars - Tom Blomfield