Accumulated Accidents Test
Find legacy practices that a disruption is likely to permanently displace.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 98%
The Accumulated Accidents Test starts by refusing to treat the pre-disruption world as an optimized equilibrium. Analysts reconstruct how an established practice emerged and ask whether it reflects genuine effectiveness or merely a sequence of historical constraints, conventions, and compromises. They then compare that practice with alternatives adopted during a shock. If the new method is cheaper, easier, more measurable, or otherwise superior, the old structure may not fully return even after the immediate disruption ends. The framework therefore separates temporary volume changes from persistent behavioral and institutional change. It is especially useful for identifying legacy processes vulnerable to digital replacement and for testing whether a crisis beneficiary has a durable runway rather than a one-time demand surge.
Origin
Tom Slater credited colleague Dave Putschnowski with applying the accumulated-accidents idea to the post-COVID investment landscape on Masters in Business.
Core principles
- 01Existing practices are not necessarily optimal merely because they are established.
- 02Many industry conventions result from layered historical accidents.
- 03A disruption reveals which practices persist by inertia rather than merit.
- 04Superior alternatives adopted during a shock may prevent a return to the old norm.
- 05Structural change must be separated from temporary demand spikes.
How to run it
- 1
Describe the old norm
Record how customers, suppliers, and intermediaries behaved before the disruption, including the institutions and rituals supporting that behavior.
Pro tip Describe the workflow precisely rather than labeling it simply traditional or offline.
Watch out Do not assume longevity proves efficiency.
- 2
Trace its historical causes
Identify the regulations, technologies, distribution constraints, habits, and compromises that produced the old arrangement.
Pro tip Ask what would be designed differently if the system were created today.
Watch out Avoid calling every disliked feature an accident; some constraints remain real.
- 3
Study forced alternatives
Observe which replacement behaviors gained adoption during the disruption and what users learned from them.
Pro tip Look for adoption beyond specialist or early-adopter groups.
Watch out Emergency adoption alone does not prove long-term preference.
- 4
Compare effectiveness
Evaluate the old and new approaches on cost, convenience, reach, data quality, and outcomes.
Pro tip Prioritize measurable improvements over novelty.
Watch out Account for hidden costs that were temporarily subsidized during the disruption.
- 5
Test persistence
Estimate which users have both the incentive and ability to retain the new behavior after conditions normalize.
Pro tip Separate likely demand normalization from permanent expansion of awareness and access.
Watch out Do not equate a post-shock decline with complete reversion.
- 6
Map investment implications
Identify businesses enabling persistent practices and incumbents dependent on accidental legacy structures.
Pro tip Favor businesses whose advantage increases as the new behavior spreads.
Watch out A correct structural thesis can still be a poor investment at an excessive price.
In the wild
Traditional television advertising budgets were heavily committed during annual upfront events. Connected television introduced audience data and more measurable digital purchasing, raising the question of whether the ceremony remained useful or was simply inherited from an earlier media system.
→ The test highlights data-driven advertising platforms as potential beneficiaries of a legacy practice losing relevance.
Lockdowns forced knowledge workers to learn video communication at massive scale. Demand would naturally fall as in-person meetings returned, but widespread product familiarity and an expanded user base could persist because the previous specialist, room-based setup was unnecessarily cumbersome.
→ The analysis separates temporary peak usage from a lasting increase in the addressable market.
Common mistakes
Assuming full reversion
Treating every crisis-era behavior as temporary ignores knowledge, infrastructure, and preferences acquired during the disruption.
Assuming every change persists
Forced adoption can disappear when restrictions end, especially if the alternative is inferior without emergency conditions.
Ignoring why the old system existed
Some legacy practices solve durable legal, economic, or coordination problems and cannot be dismissed as historical accidents.
Is it for you?
Best for
It is best for investors and strategists analyzing how a major shock may permanently reshape established markets.
Not ideal for
It is not ideal when the pre-disruption system was already demonstrably efficient or when replacement behavior lacks adoption evidence.
From the transcript
“One of the frameworks that I've I've found really helpful and it's worked on with by by one of my colleagues, Dave Putschnowski is a…”
“So this idea that what were the structures UM that were the norm before COVID hit. There weren't the sort of local maximum or the…”
“Because I think those are the things that are UM were unlikely to go back to UM as as COVID starts to unwind UM.”
From the episode
Tom Slater on Growth Investing (Podcast)
Tom Slater