Thread
Having Co-founded a billion dollar company (@Udemy) and raised US$20m from @a16z for @MavenHQ:
@Gaganbiyani knows a thing or two about finding great startup ideas.
Here's the framework he uses to evaluate startup ideas:
@Gaganbiyani knows a thing or two about finding great startup ideas.
Here's the framework he uses to evaluate startup ideas:
The traditional approach in startups is that you can't predict whether people will want your product.
So, you build a 'Minimum Viable Product' as fast as possible to see if users find value in the idea.
If they don't, you iterate until they do.
This isn't Gagan's approach.
So, you build a 'Minimum Viable Product' as fast as possible to see if users find value in the idea.
If they don't, you iterate until they do.
This isn't Gagan's approach.
Instead, Gagan uses 'Minimum Viable Testing.'
An MVT is a test of a critical hypothesis—something that must be right for the company to stand a chance.
For example, AirBnB had to be right that strangers would pay to stay in others' homes, or else the business fails.
An MVT is a test of a critical hypothesis—something that must be right for the company to stand a chance.
For example, AirBnB had to be right that strangers would pay to stay in others' homes, or else the business fails.
So, rather than building any software, they tested this assumption in the simplest manner possible.
They provided air mattresses in their living room, free wifi, and breakfast to attendees of an upcoming conference (where all the hotels were booked out).
They provided air mattresses in their living room, free wifi, and breakfast to attendees of an upcoming conference (where all the hotels were booked out).
The team were able to validate—with real-world data, minimal risk and cost—that people were willing to stay with strangers.
In fact, this model allowed travellers to have a unique, personable experience with the city and other like-minded people.
The rest is history.
In fact, this model allowed travellers to have a unique, personable experience with the city and other like-minded people.
The rest is history.
So, how does an MVT compare to an MVP?
MVP = a basic early version of a product that looks and feels like a simplified version of the eventual vision.
MVT = a specific test of an assumption that must be true for the business to succeed.
MVP = a basic early version of a product that looks and feels like a simplified version of the eventual vision.
MVT = a specific test of an assumption that must be true for the business to succeed.
How do the strategies differ?
MVP = there is no strategy: you throw things at the wall until it sticks. You iterate from an initial product.
MVT = take your time to discover a strategy but once you have one, move forward with conviction. You build with confidence.
MVP = there is no strategy: you throw things at the wall until it sticks. You iterate from an initial product.
MVT = take your time to discover a strategy but once you have one, move forward with conviction. You build with confidence.
What's wrong with MVP's?
(1) MVP’s overemphasize building
New products don’t succeed because of the number of features they provide.
They succeed because of one core insight that customers actually care about.
MVT's focus the attention on this core insight and not features.
(1) MVP’s overemphasize building
New products don’t succeed because of the number of features they provide.
They succeed because of one core insight that customers actually care about.
MVT's focus the attention on this core insight and not features.
(2) MVP’s cause founders to obsess about customer opinion
Customers rarely know what they really want, and even if they do, they’ll want different things.
If you rely on customers, you’ll build incremental improvements instead of a novel breakthrough.
Customers rarely know what they really want, and even if they do, they’ll want different things.
If you rely on customers, you’ll build incremental improvements instead of a novel breakthrough.
(3) MVPs are often overbuilt and make horrible core products
Once you start writing code with an MVP, you start to add technical and product debt.
If instead, you run an MVT, you start from a fresh slate and are actually building the longer-term vision.
Once you start writing code with an MVP, you start to add technical and product debt.
If instead, you run an MVT, you start from a fresh slate and are actually building the longer-term vision.
So, what does starting a company look like with a MVT?
1: Immerse yourself in a new industry
2: Determine your user's jobs-to-be-done
3: Identify your value proposition: what promise can you make to help a user with their jobs-to-be done?
Now, it's time for the MVT—not MVP.
1: Immerse yourself in a new industry
2: Determine your user's jobs-to-be-done
3: Identify your value proposition: what promise can you make to help a user with their jobs-to-be done?
Now, it's time for the MVT—not MVP.
The Minimum Viable Testing Framework
(1) List the riskiest assumptions that might lead your business to succeed or fail:
• Why might this not work?
(2) Test your assumptions through Minimum Viable Tests:
• Determine if your idea actually works in reality
(1) List the riskiest assumptions that might lead your business to succeed or fail:
• Why might this not work?
(2) Test your assumptions through Minimum Viable Tests:
• Determine if your idea actually works in reality
(1) Riskiest assumptions commonly include:
• Building something people don’t want (test: force people to pay)
• Execution risk (harder in reality than theory)
• Marketing/Distribution (can you reach users)
• Market size
• Profitability
• Building something people don’t want (test: force people to pay)
• Execution risk (harder in reality than theory)
• Marketing/Distribution (can you reach users)
• Market size
• Profitability
(2) Minimum Viable Tests
Pick your riskiest assumption and devise a specific test for that assumption.
There is no specific 'prescription' for how to do this because it depends on the nature of the company, industry, and specific risks.
This is where creativity is required.
Pick your riskiest assumption and devise a specific test for that assumption.
There is no specific 'prescription' for how to do this because it depends on the nature of the company, industry, and specific risks.
This is where creativity is required.
Here, it's critical to disregard the future state of what the solution might look like.
Instead, focus on the lowest risk and cost solution you can deliver in the shortest possible timeframe.
This will often involve leveraging assets—just like AirBnB did—you already have.
Instead, focus on the lowest risk and cost solution you can deliver in the shortest possible timeframe.
This will often involve leveraging assets—just like AirBnB did—you already have.
Once you've tested enough risky assumptions to have more confidence about your product viability, it's time to build an initial product.
With the help of your target customers, iterate on the product until you have nailed product/market fit.
Then, it's time to scale!
With the help of your target customers, iterate on the product until you have nailed product/market fit.
Then, it's time to scale!
If you'd like a deeper dive on the MVT process, I highly recommend reading Gagan's initial article:
review.firstround.com/the-minimum-viable-testing-process-for-evaluating-startup-ideas
review.firstround.com/the-minimum-viable-testing-process-for-evaluating-startup-ideas
That's an effective framework for evaluating startup ideas—courtesy of
@gaganbiyani.
Follow me @_alexbrogan for more mental models, concepts, and frameworks
@gaganbiyani.
Follow me @_alexbrogan for more mental models, concepts, and frameworks
