The New Paradigm Test
Build only what the new technology makes possible for the first time.
- Difficulty
- Easy
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 85%
In his earliest recorded interview Bezos stated the rule plainly: if you can do something in the old paradigm, you should; when a new paradigm arrives, look for the things you could not do any other way. The mechanism is a filter applied to an idea list. For each candidate you ask what physical or economic constraint the old world imposed, and whether the new technology actually removes it. Ideas that survive are impossible-before rather than merely better. Applied to books in 1994, the constraint was shelf space: a physical superstore could stock perhaps eighty thousand titles, while three million were in print, so a true superstore with exhaustive selection simply could not exist in any other way. The failure mode is the banner ad — a magazine page pasted onto the web — rather than inventing the feed.
Origin
Stated by Jeff Bezos in his first recorded interview, given outside a Seattle conference in the mid-1990s, and highlighted as a playbook theme on the Acquired episode on Amazon.com.
Core principles
- 01If it works in the old paradigm, it should stay in the old paradigm.
- 02A new medium earns its cost of adoption only when it removes a hard constraint.
- 03Porting an old format to a new platform captures none of the new platform's power.
- 04The best ideas look impossible rather than merely improved.
- 05The test is a filter, not an idea generator.
How to run it
- 1
Name the old paradigm's binding constraint
For each candidate idea, write down the specific physical, economic or informational limit the previous technology imposed. Vague constraints produce vague ideas.
Pro tip Constraints are usually about space, distance, time, cost per unit or information asymmetry.
- 2
Test whether the new technology removes it
Ask directly whether the new platform eliminates that constraint or merely makes it slightly cheaper. Marginal improvement does not qualify.
Watch out Beware ideas where the new technology removes a constraint nobody was actually complaining about.
- 3
Discard everything that still works without the new technology
If the idea would function in the old paradigm, it belongs there. This is the cut that removes most of the list.
Pro tip Force the cut by writing the old-paradigm version of the product; if it reads fine, delete the idea.
- 4
Find the native format, not the port
Do not take an existing format and place it on the new platform. Invent the format the platform makes possible — the feed, not the banner ad.
Pro tip Ask what the platform does that has no analogue at all in the old world; build around that.
Watch out Ported formats feel safe because they are legible to customers and investors, which is exactly why they get funded and then fail.
- 5
Pressure-test the survivors on demand
An impossible-before product still needs someone who wants it. Check the removed constraint was one that genuinely frustrated customers.
In the wild
Barnes & Noble and Borders superstores stocked roughly eighty thousand titles, while about three million books were in print worldwide. The constraint was physical shelf space, and no amount of retail skill could remove it. The web removed it entirely, letting Amazon offer a store with exhaustive selection that, in Bezos's words, simply could not exist in any other way. Long-tail titles that previously meant a special order and a month's wait became a normal purchase.
→ A category-defining product whose core promise was impossible in the previous paradigm.
The hosts use the banner ad as the archetype of failing the test: a magazine advertisement pasted onto a web page. It works in the old paradigm, so by the rule it should have stayed there. The native format the web actually made possible was the feed, which took another decade to invent.
→ A decade of value captured by whoever eventually built the native format rather than the port.
Common mistakes
Confusing cheaper with impossible
A 20% cost reduction is an optimisation, not a paradigm shift, and rarely justifies the risk and adoption cost of building on brand-new technology.
Shipping the skeuomorph
Porting a familiar format onto a new platform is easy to explain and easy to fund, which is why so much of every platform shift is wasted on it.
Treating the test as an idea generator
The rule filters a list; it does not produce one. You still need domain knowledge to generate candidates worth filtering.
Is it for you?
Best for
Founders and investors filtering a long list of ideas during a platform shift.
Not ideal for
Mature markets where the winning move is execution and distribution rather than a new capability.
From the transcript
“you basically want to find things that you could not do any other way”
“don't go create the banner ad and slap it on the internet and be like see it's like a magazine but on the internet you…”
“with that huge diversity of products 3 million books in print you could build a store online that simply could not exist in any other…”
From the episode
Amazon.com