This is a summary of Bob Dorf‘s Business of Software 2012 presentation.
Why most startups fail
Most startups fail to scale. More startups fail from a lack of passionate customers than anything else.
Code is only half the job. You can pretty much build anything these days. Technology isn’t the problem, figuring out who your customers will be and finding passionate customers is the hard part.
If you are a passionate founder, putting in 20,000 hours gives you about a 1 in 8 chance of success.
Every founding team needs three members: a hacker, a hustler, and an artist. Every morning the hacker and the hustler should get together. Then the hacker should spend all day on product development and the hustler should spend all day on customer development.
Way back in the day, building a successful company was all about rigor and process, but things have changed.
Most startups fail because they
- Assume customer problem is known.
- Assume product features are known.
They assume it’s a simple linear process of concept/seed round -> product dev -> alpha/beta test -> launch/first ship.
We assume the customer’s problem is known, and how to solve it. Yet as a startup, we have no customers!
Startup Enemy #1: Business Plans
Business plans belong in the creative writing department, not the business development department of a university.
You should always be innovating. Always ask what can I change to make it better. Constantly seek customer feedback.
Always challenge your business model. No business plan survives first contact with customers. Ask Webvan.
What is a startup? A startup is a band of pirates who get together temporarily to see if they can make all the parts of a business plan come together. They are in search mode. Once they’ve done enough searching, they can put together a business plan based on facts. By definition there is no such thing as an eight year startup. It is a four year failure.
Startups should have more of an operating plan than a business plan. Alex Osterwalder’s business model canvas is a great start. The canvas has nine components, but the most important ones are:
- Value proposition – What problem do we solve?
- Customer segments – For whom do we solve it?
- Customer relationships – How do we get, keep, and grow them?
- Revenue stream – How do we make money?
The customer segments should be defined as tightly and narrowly as you can. Customer relationships is about the three most critical activities of any business. It is about your promise/pledge to the customer.
Take a couple of people to discuss and create new business models. When you’re done completing the business model canvas, you’ve got product-market fit. However, all you have are nine best guesses. How do we turn the guesses into facts? Go out and talk to your customers to find out if they agree with your guesses. This is where customer development takes over.
The customer development process is customer discovery -> customer validation -> customer creation -> company building. Customer discovery and customer validation form the search phase of the customer development process. Pivoting occurs as one cycles between customer discovery and customer validation. Customer creation and company building form the execution phase of the process.
The search phase is key. You can learn the execution phase in business school. The search phase takes you through the process of validating all the guesses you’ve made in your business model.
The Minimum Viable Product
Discovery begins with the minimum viable product. Develop the fewest possible features to make the point. You can use the MVP to test new ideas.
You want to have something customers can interact with, so create something as soon as possible that customers can play with – even if it doesn’t “work.” The reactions you get from customers that try your MVP are better than the reactions you get from simply telling them about it. Reactions are incredibly helpful in making design decisions.
An example of creating an MVP to test your value proposition is the diapers.com story. Diapers.com put up a website and started taking orders before they had any diapers in stock. They just wanted to know if their idea was worth pursuing further. They ended up driving all around town to different stores, buying diapers and shipping them themselves. When the orders grew, they rented vans to transport the diapers. They lost money in the process, but they weren’t testing profitability. They were testing the business model to see if they really had a business. They were testing their value proposition – their consumer promise.
An incidental loss per sale is worth the information gained when testing your value proposition.
An MVP is all about talking to customers. You want to get to your MVP as fast as you can so you can get honest feedback quickly. Ask your customers:
- If it didn’t suck so bad, how would you find it?
- What do other products do to solve the problem?
- What could ours do better?
The pivot is the heart of customer development. Pivoting is the iteration between customer validation and customer discovery. The pivot is fast, agile, and opportunistic.
Pivot if you start to hear consistently from 20 – 40 customers that something needs to change. Not when only 2 – 3 customers complain.
When pivoting, go back to the business model, make the change, go back to the customers, and ask if the better idea solves their problem. You cannot delegate customer validation. As the CEO, you MUST be involved.
Pivots in large companies happen all the time, but it usually comes with a firing. In a startup, a pivot is almost a celebration – you found a better way to reach customers, you found 20 more customers, etc.
One problem with pivots: premature evacuation (i.e. not talking to enough customers). Are you collecting enough data from your core customers to make a smart decision on a pivot? Three people told you it was bad, and now you are giving up? No. Go talk to 20 more customers before making a decision.
You want to pivot as quickly as you can, as long as you are careful about premature evacuation. The pivot cycle time matters. Do it as quickly as you can, because you’ve got the ticking time bomb of a declining bank balance. Speed of cycle minimizes cash needs.
How do you know when you are done?
You are never really done, but you can scale back when you know who the customers are, how you will find them, how the product fits them, etc.
Use the business model canvas as your scorecard to know when you’re done.
Customer validation is like putting gas in a race car with a turkey baster (i.e. tweak the variables to see if the same customer quality/stick rate/cash conservation continue to work).
Make sure that all your guesses are validated by customers. Customer validation is a great activity to perform no matter how happy you are with your product.
Always look for passionate users (i.e. those who want the product to be better just as badly as you and your investors).
Join the Business of Software Community.
Get the latest updates from Business of Software free to you. You will receive conference news, blogs from some of the world's best leaders, and access BoS video content as soon as it goes live by signing up here.