Focus on What Makes Your Beer Taste Better
Own only what customers taste; rent the rest, or become the utility others rent
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 5
- Confidence
- 89%
Bezos used the analogy at the 2008 Y Combinator Startup School to sell AWS. Around the turn of the twentieth century, electricity was transformative for brewing, and the first breweries to adopt it built their own generators. That worked for a few years but was capital intensive and operationally heavy. The next generation simply rented power from utility companies and ran roughshod over the first, because whoever makes your electricity has no impact on how your beer tastes, while making it yourself does raise your cost structure. The framework has two halves. As an operator, focus solely on the attributes of your product customers care about and outsource everything else. As a strategist, notice what Bezos did rather than what he said: being a utility company, particularly an unregulated one, is an exceedingly great business. Square, Shopify, Vanta and Modern Treasury all occupy that position, supplying mission-critical infrastructure that does not make anyone's beer taste better.
Origin
From Jeff Bezos's talk at the 2008 Y Combinator Startup School, where he used European breweries and their own power generators to argue startups should build on AWS rather than roll their own infrastructure.
Core principles
- 01Whoever makes your electricity has no impact on how your beer tastes, but making it yourself raises your cost structure.
- 02Customers only reward attributes they can perceive in the product.
- 03Vertical integration into invisible inputs is a tax you pay for the illusion of control.
- 04The corollary is the bigger prize: supplying the invisible input to everyone else is an outstanding business.
- 05An unregulated utility in technology is one of the most defensible positions available.
- 06This is specialisation of labour, well understood for people, applied at the level of firms.
How to run it
- 1
Inventory everything you build yourself
List the capabilities you currently own end to end: infrastructure, tooling, internal platforms, back-office systems, physical operations.
Pro tip Include the things nobody questions because they have always been in-house.
- 2
Apply the taste test to each one
For each capability, ask whether a customer could ever detect the difference between your version and a bought-in one. If not, it does not make your beer taste better.
Pro tip Ask a real customer rather than guessing; teams systematically overestimate what is perceivable.
Watch out Cost, latency and reliability can be perceivable. The test is customer perception, not physical proximity to the product.
- 3
Rent everything that fails the test
Move the invisible layers to specialists. The first-generation breweries were not wrong to want power; they were wrong to generate it, because it raised their cost structure with no product benefit.
Watch out Outsourcing something customers genuinely can taste is the mirror-image error and is far more damaging.
- 4
Redeploy the freed capital and attention
The gain is not just cash saved. It is operational labour and founder attention returned to the attributes that determine whether you win.
Pro tip Track where the freed engineering time actually went; if it dissipated, you captured only half the benefit.
- 5
Look for the utility position on the other side
Ask whether the infrastructure you understand could be supplied to everyone else. The prize is being the mission-critical layer other companies need but which does not differentiate their product.
Pro tip Bezos evangelised AWS directly to tiny startups and future startups, which bought Amazon roughly a five-year lead in cloud.
Watch out Utility economics only work at scale; a sub-scale utility carries the cost structure without the position.
In the wild
When electricity arrived it let breweries produce vastly more beer than before, and the first adopters built their own generators. That was fine for a few years, but it was capital intensive and demanded operational labour to run. The next generation of breweries rented power from the utility companies instead and outcompeted the first generation, because the source of the electricity had no effect on the taste of the beer while the generators had a real effect on cost.
→ The renters beat the builders on cost with an identical product.
Bezos used the analogy as a marketing pitch to persuade startups to build on AWS rather than roll their own infrastructure. It worked twice over. Startups did focus on their own products, and Amazon captured roughly a five-year lead in cloud by flying people from Seattle to the Bay Area to evangelise relentlessly. The hosts note that the more important lesson is what Bezos did rather than what he said: he was building an unregulated utility.
→ A marketing analogy that both freed customers and built one of the most defensible businesses in technology.
Common mistakes
Confusing control with differentiation
Owning an input feels safer but adds cost without adding anything the customer can perceive. The generator did not improve the beer.
Outsourcing the thing customers actually taste
The rule cuts both ways. Handing off the attributes that define your product is how companies commoditise themselves.
Missing the utility opportunity in your own stack
Teams apply only the outsourcing half and never ask whether the infrastructure they built could be sold to everyone else as a mission-critical layer.
Is it for you?
Best for
Startups and operators deciding what to build versus buy, and platform businesses hunting a defensible layer.
Not ideal for
Businesses whose differentiation genuinely is the infrastructure, where the input is exactly what customers taste.
From the transcript
“you should focus solely on the attributes of your product that your customers are going to care about everything else your infrastructure doesn't matter outsource”
“being a utility company is an exceedingly exceedingly great business and particularly being an unregulated utility company”
“if you can provide a critical mission critical piece of infrastructure that other companies can use that they need but doesn't make their actual beer…”
From the episode
The Playbook: Lessons from 200+ Company Stories