Incumbent Disruption Response
Ship the disruptive product while deliberately managing cannibalization
- Difficulty
- Expert
- Time to result
- ~ongoing to results
- Steps
- 7
- Confidence
- 95%
The framework begins when a technology stops merely improving an incumbent product and enables a different product that can replace it. Leadership must acknowledge the shift, organize around a native response, and ship quickly enough to learn from real users. The team then maps individual use cases: some directly cannibalize the legacy franchise, while others complement it or create new categories. Distribution is staged accordingly, with experiments measuring product quality, adoption, revenue, and damage to the core. Fragmented research and product teams should consolidate around a shared platform when scale economics demand it. The goal is not to prevent all cannibalization; it is to control the transition while ensuring a competitor does not own the new platform.
Origin
The hosts derive the framework from Google's response after ChatGPT transformed AI from a sustaining improvement to an existentially disruptive product. Sundar Pichai issued a code red, merged Brain with DeepMind, standardized Google on Gemini, and staged AI distribution across products.
Core principles
- 01A sustaining technology can become disruptive when its product form changes.
- 02Incumbents must build the new experience instead of defending only the old one.
- 03Cannibalization should be measured by use case rather than discussed abstractly.
- 04Complementary applications offer safer places to build early momentum.
- 05Rapid action and rash action are different.
- 06Organizational structure must match the strategic emergency.
How to run it
- 1
Reclassify the technology
Determine whether the technology still strengthens the incumbent product or now enables a substitute that performs the customer's job differently.
Pro tip Evaluate the end-user experience, not only the underlying technical capability.
Watch out A technology can shift from sustaining to disruptive suddenly when a new interface unlocks demand.
- 2
Declare an operational response
Set a clear mandate to build and ship native products on the new platform. Give the effort executive urgency and explicit authority.
Pro tip Translate strategic language into concrete product and delivery expectations.
Watch out A declaration without changed priorities, staffing, or incentives will not alter execution.
- 3
Consolidate critical capabilities
Unify overlapping teams and standardize on a shared platform where fragmented efforts dilute talent, data, and compute.
Pro tip Name one accountable leader for the integrated platform.
Watch out Forced consolidation can destroy productive autonomy unless the strategic rationale is explicit.
- 4
Map cannibalization by use case
Identify which customer interactions the new product replaces, which it enhances, and which create entirely new demand.
Pro tip Use query, workflow, or customer-level data instead of a single company-wide cannibalization estimate.
Watch out Treating every new use case as either harmless or fatal produces poor allocation decisions.
- 5
Stage distribution
Introduce the new experience through selected surfaces, users, and scenarios while preserving the legacy product where it remains stronger or more profitable.
Pro tip Use opt-in modes and targeted defaults to gather evidence before wholesale migration.
Watch out Overly cautious staging can give a disruptive competitor time to become the default.
- 6
Expand complementary categories
Invest aggressively in applications where the new technology creates value without directly eroding the legacy franchise.
Pro tip Use incumbent assets such as data, infrastructure, and distribution to establish early leadership.
Watch out Safe adjacencies cannot substitute indefinitely for confronting the core disruption.
- 7
Tune the transition continuously
Monitor adoption, quality, unit economics, legacy growth, and competitive movement, then adjust exposure and investment repeatedly.
Pro tip Maintain separate dashboards for value creation and value capture.
Watch out Short-term financial stability can conceal long-term erosion of the underlying franchise.
In the wild
After ChatGPT's launch, Google merged Brain and DeepMind, established Gemini as the common model, launched a separate Gemini application, introduced AI Overviews for selected searches, and offered AI Mode without redirecting all Google traffic into a chatbot. It also accelerated non-search products such as video generation.
→ Google became a credible frontier-AI competitor while search revenue and company earnings continued growing during the early transition.
A mature software vendor sees an AI-native competitor automate its core workflow. It launches an AI workspace for selected customers, integrates assistance into low-risk tasks, measures seat and usage cannibalization, and expands the new interface as retention and unit economics improve.
→ The vendor learns the new model before competitors capture its customers while avoiding an uncontrolled migration.
Common mistakes
Protecting the legacy product at all costs
Refusing to ship a substitute does not preserve demand if competitors can deliver it. It merely transfers control of the transition to them.
Replacing everything immediately
A disruptive product may still have quality, safety, or economic weaknesses. Wholesale migration can destroy trust and profits before the replacement is ready.
Confusing speed with recklessness
Rapid execution requires focused decisions and staged learning. It does not require launching unverified products into every customer interaction.
Is it for you?
Best for
Large incumbents facing a new product architecture that performs their core customer job differently.
Not ideal for
Startups without a legacy franchise, distribution base, or existing profits to protect.
From the transcript
“we need to build and ship real native AI products ASAP.”
“you need to be laser operationally in all the details to try and figure out where is it that the new product is actually cannibalizing…”
“Rapid but not rash, you know.”
From the episode
Google Part III: The AI Company