The Startup Owner's Manual

date Aug 28, 2016
authors Steve Blank and Bob Dorf
reading time 26 mins

Reading period: 3 August 2016 - 28 August 2016

MBA !== Running a startup

Today, after half a century of practice, we know unequivocally that the traditional MBA curriculum for running large companies like IBM, GM and Boeing does not work in startups. In fact, it’s toxic.

Companies = knowns, Startups = search

Companies execute business models where customers, their problems, and necessary product features are all “knowns.” In sharp contrast, startups operate in “search” mode, seeking a repeatable and profitable business model.

Startups == search mode

startup? A startup is not a smaller version of a large company. A startup is a temporary organization in search of a scalable, repeatable, profitable business model. At the outset, the startup business model is a canvas covered with ideas and guesses, but it has no customers and minimal customer knowledge.

Hero with a 1000 faces

At the outset of the voyage, the path is unclear and no end is in sight. Each hero meets a unique set of obstacles, but Campbell’s keen insight was that the outline of these stories is always the same. There are not a thousand different heroes but one hero with a thousand faces.

Untested hypothesis

Instead, they combine agile engineering and Customer Development to iteratively build, test and search for a business model, turning unknowns into knowns. Winners also recognize their startup “vision” as a series of untested hypotheses in need of “customer proof.”

Founders’ primary role

Founders, not employees, must search for a business model. The best way to search is for the founders themselves to get out of the building to gain a deep, personal, firsthand understanding of their potential customers’ needs before locking into a specific path and precise product specs.

Products to customers vs products to sales / marketing

The core of Customer Development is blissfully simple: Products developed by founders who get out in front of customers early and often, win. Products handed off to sales and marketing organizations that are only tangentially involved in the new-product development process will lose.

Traditional irreversibility of the waterfall method

This process starts with the founder’s vision, which may be expanded into an MRD (and a product requirements document), and expands further into detailed engineering specifications. With those in hand, Engineering begins implementation fueled by cold pizza and long nights and weekends. Once a waterfall process starts, the proverbial train has left the station and the product is nearly impossible to revise.

The unknowns in a startup

For new products like Webvan, the business plan fails as a roadmap because both the product and the customer are unknown.

For most startups, these nine flawed assumptions are the most toxic of all:

  1. Assuming “I Know What the Customer Wants”
  2. The “I Know What Features to Build” Flaw
  3. Focus on Launch Date
  4. Emphasis on execution instead of hypotheses, testing, learning and iteration
  5. Traditional Business Plans Presume No Trial and No Errors
  6. Confusing Traditional Job Titles with What a Startup Needs to Accomplish
  7. Sales and Marketing Execute to a Plan
  8. Presumption of Success Leads to Premature Scaling
  9. Management by Crisis Leads to a Death Spiral

Product feedback from customers

Yet without direct and continuous customer contact, it’s unknown whether the features appeal to customers. Fixing the inevitable product mistakes after building and shipping the entire product is costly and time-consuming, if not deadly. It can render the product obsolete by launch.

Traditional product development

The traditional product introduction model focuses engineering, sales and marketing on the all-important, immovable launch date.

Tolerance of wrong turns

Neither management nor investors tolerate “wrong turns” that result in delays. In fact, traditional engineering schedules have test cycles with the names alpha, beta, and release but rarely allow time to improve the product. They’re still geared to putting out the original product with minimal bugs, though.

Not enough demand…

Time after time, only after launch does a startup discover that not enough customers visit its website, play the game, bring their friends, or convert to orders. Or it discovers that early customers don’t scale into a mainstream market, or the product doesn’t solve a high-value problem, or the cost of distribution is too high.

Premature scaling

Like all startups focused on executing to a sequential product introduction plan, Webvan hired vice presidents of merchandising, marketing and product management—all oriented around executing a given sales and marketing strategy instead of listening to customers and discovering customer needs.


In a startup, these metrics don’t track progress against the startup’s only goal: to find a repeatable and scalable business model. Instead, traditional metrics get in the way. Instead of asking, “How many days to the beta test?” or, “What’s in our sales pipeline?” a startup’s board and management team need to ask specific questions about results of its long list of tests and experiments to validate all components of its business model.

Financial metrics that matter!

No matter what, directors and founders must stay focused on one financial metric that always matters: cash burn rate and number of months’ worth of cash left in the bank.


They must be eager to search for a repeatable and scalable business model. Agile enough to deal with daily change and operating “without a map.” Readily able to wear multiple hats, often on the same day, and comfortable celebrating failure when it leads to learning and iteration.

Startup progress !== revenue plan or product launch

But in a majority of startups, measuring progress against a product launch or revenue plan is simply false progress, since it transpires in a vacuum absent real customer feedback, instead of searching for an understanding of customers and their problems and replacing assumptions with facts.

Used to searching vs used to plan…

Searching for a business model requires a different organization than the one used to execute a plan. Searching requires the company to be organized around a customer development team led by the founders.

Missing departments when you are searching

There are no sales, marketing or business development departments when you are searching for a business model. If you’ve organized your startup with those departments, you are not really doing customer development.

Provision for failing quickly, cheaply

We will have failures and we will screw it up several times before we get it right.” In contrast, a traditional product introduction plan makes no provision for moving backward. To do so would be considered a failure. No wonder most startup founders are embarrassed when they’re out in the field learning, failing, and learning some more.

Paying customers == product validation

In a single-sided market (one where the user is the payer), a steady stream of customer purchases validates the concept far more solidly than lots of polite words. There’s no surrogate for people paying for a product.

Release early

By contrast, IMVU launched a buggy early product roughly 120 days after it was founded. Amazingly, some customers loved the buggy product enough not only to pay for it, but also to give the founders what they wanted: feedback (and money).

Company vs startup

“Graduation day” arrives when the startup finds a scalable, repeatable business model. At this point it’s fundamentally no longer the temporary search-oriented organization known as a startup — it’s a company! In a sometimes-bittersweet transition out of startup mode, company-building refocuses the team’s energy away from “search” mode and to a focus on execution, swapping its informal learning- and discovery-oriented Customer Development team for formal, structured departments such as Sales, Marketing and Business Development, among others, complete with VPs. These executives now focus on building their departments to scale the company.

Customer development manifesto:

  • Rule No. 1: There Are No Facts Inside Your Building, So Get Outside.
  • Rule No. 2: Pair Customer Development with Agile Development
  • Rule No. 3: Failure is an Integral Part of the Search
  • Rule No. 3: Failure is an Integral Part of the Search
  • Rule No. 4: Make Continuous Iterations and Pivots
  • Rule No. 5: No Business Plan Survives First Contact with Customers So Use a Business Model Canvas
  • Rule No. 6: Design Experiments and Test to Validate Your Hypotheses
  • Rule No. 7: Agree on Market Type. It Changes Everything
  • Rule No. 8: Startup metrics differ from those in existing companies
  • Rule No. 9: Fast decision-making, cycle time, speed and tempo
  • Rule No. 10: It’s all about passion
  • Rule No. 12: Preserve All Cash Until Needed
  • Rule No. 11: Startup job titles are very different from a large company’s
  • Rule No. 13: Communicate and Share Learning
  • Rule No. 14: Customer development success begins with buy-in

Product development

Customer Development is useless unless the product development organization can iterate the product with speed and agility. If Engineering builds the product using waterfall development, it will be deaf, dumb and blind to customer input except during a short period when it’s specifying the product.


The best startup founders don’t hesitate to make the change. They admit when hypotheses are wrong and adapt.

Simplest test + designing the test

Start by asking yourself, “What insight do I need to move forward?” Then ask, “What’s the simplest test I can run to get it?” Finally, think about, “How do I design an experiment to run this simple test?”

Making reversible decisions

Startups should as policy, make reversible decisions before anyone leaves the CEO’s office or before a meeting ends. Perfect decision-making is both unimportant and impossible, and what matters more is forward momentum and a tight, fact-based feedback loop to quickly recognize and reverse bad decisions.

Quick decisions

Learning to make decisions quickly is just part of the equation. Agile startups have mastered another trick: tempo — the ability to make quick decisions consistently and at all levels in the company.

Cash spending

The goal of Customer Development is not to avoid spending money but to preserve cash while searching for the repeatable and scalable business model. … The Customer Development process preserves cash by not hiring any sales and marketing staff until the founders turn hypotheses into facts and discover a viable product/market fit.

Search for pattern and not one-off

Search not for the one-off revenue hits but rather for a pattern… Business model: A business model answers the basic questions of how the company makes money. Is this a revenue play, or is it a freemium model seeking users? Something else? Who’s the customer?

The company spent $5 billion building a business over eight years without ever focusing on four key questions:

  • Have we identified a problem a customer wants to see solved?
  • Does our product solve this customer problem or need?
  • If so, do we have a viable and profitable business model?
  • Have we learned enough to go out and sell?

From hypotheses to facts

So the No. 1 goal of customer discovery amounts to this: turning the founders’ initial hypotheses about their market and customers into facts.

Founder role

Get Out of the Building Facts exist only outside the building , where customers live, so the most important aspect of customer discovery is getting out of the building, in front of customers. And not for a few days or a week, but repeatedly, over weeks if not months. This critical task can’t be assigned to junior staffers and must be driven by founders. Only after the founders have performed this step will they know whether they have a valid vision or just a hallucination.

Building mainstream product is perilirious

In a startup, the first product is not designed to satisfy a mainstream customer. No startup can afford to build a product with every feature a mainstream customer needs all at once. The product would take years to get to market and be obsolete by the time it arrived. Instead, successful startups solve this conundrum by focusing development and early selling efforts on a very small group of early customers who have bought into the startup’s vision.

Build for small group…

The idea that a startup builds its product for a small group of initial customers rather than devising a generic mainstream spec is radical.

Build an MVP…

A third, more productive approach is to develop the core features of the product (incrementally and iteratively with agile engineering methods), with the feature list driven by the vision and experience of the company’s founders. This is a minimum viable product.

indispensable features

Engineers tend to make a product bigger and more perfect. The MVP helps them focus the most important and indispensable features. Your goal in having an MVP is not to gather feature requests to change the product or to make the feature set larger.

What is the problem you are solgin?

Study the problem. Is your product solving a mission-critical company problem or satisfying a must-have consumer need? Is your product a must-have or a nice-to-have? When your product solves a problem that costs customers sleep, revenue, or profits, things are definitely looking up.

Single sales channel initially

Don’t try to launch a product via direct sales, chain stores and direct mail all at once—it’s almost impossible to succeed at all three. The big exception, of course: physical-channel launches with simultaneous web-marketing (and sometimes sales) support.

What is the preferred channel os sales?

It’s almost always smartest—and safest—to first observe existing buying patterns and habits for similar products and product categories, since customers are demonstrating their preferred channel by spending their money there.


seek a provisional answer to just one question: is the company entering an existing market, re-segmenting an existing market, creating a new market or cloning

In an existing market

In an existing market, your startup is the weakest player with the least resources. Therefore, attacking the strongest players head-on is foolish. You want to choose strategies that acknowledge your weaknesses and play to your agility. If there’s a dominant player with more than 74 percent market share, don’t attack that market head-on.

Demand creation

Getting customers, or demand creation, has four distinct stages in the physical channel: awareness, interest, consideration, and purchase.

Pull in the customers

Unlike the door-to-door salesmen of yesteryear, your job on the web is to “pull” customers to you rather than to push your product at them. The web offers a near-limitless tool set to help you pull customers in.

Some metrics to measure:

  • Track start dates and sources of each customer (referred by a blogger, another site, etc.)
  • Track customers’ activity level individually. How often do they come? How long is each visit? What’s the time span between visits?
  • When do customers abandon, and what were they doing that caused them to do so?

Different customer types in early stage

Organize the metrics around “cohorts,” or common groups of customers (like “those who joined in January”), since, for example, three-month customers may behave one way while nine-month customers may be much more or less active than their newer brethren.

Design experiments with speed, tempo and fast cycle time

But how do you test? Rule number 6 of the Customer Development Manifesto says to design experiments and Rule 9 says to do it with speed, tempo and fast cycle time. So the first step in running a test of your business model hypotheses is to design simple pass/fail experiments for each test.

Designing tests

Start by asking yourself, “What do I want to learn?” Then ask, “What’s the simplest pass/fail test I can run to learn?” Finally, think about, “How do I design a pass/fail experiment to run this simple test?”

What is a successful test?

When you get a strong “grab it out of your hands” on, say, from four of your first ten customers, it’s OK to stop the test and declare it a success. The goal is speed, learning and looking for a global maximum (not a local maximum).

Reaching global maximum

Only experience and good guesswork can tell you how long to run a test, and while shorter is always better, be sure you’ve given yourself the opportunity to reach your global maximum.

Starting customers

A solid discovery effort will usually involve 10 to 15 customer visits a week, and getting face-to-face with 50 people will probably require contacting 200 customers or more. Where do the names come from? Start with the people you know directly.

Motion is not action

One of the mistakes entrepreneurs make in this step is confusing motion with action. Motion is a sent e-mail, a left voice-mail message or a note on LinkedIn. Action is a two-way conversation. So 10 conversations may require 25 e-mails, voice mails, Tweets, etc. Keep calling until the schedule’s booked with three customer visits a day.

Competitor companies

When you start a company, you should have some vague notion of what companies are in adjacent markets or are part of the infrastructure or ecosystem of your business.

Conference and traaade shows

Not only do they provide great trinkets, but conferences and trade shows are the prime areas for both talent-and trend-spotting. Ask usual questions about trends and players, but this time accomplish a few things that can’t happen anywhere else. Get demos of competitive and adjacent products.

Pre-mature media exposure?

As a reminder, now is not the time for press releases, interviews, blogs or public demos (other than to potential investors). You simply do not know enough yet to say what business you are in. If you do get noticed by the press, simply don’t return their e-mails or calls.


Explore the pricing boundaries. Ask, “Would you pay $1 million for our software?” The answer is usually instructive. Suppose customers said, “we couldn’t see paying more than $250,000 for the first set of applications.” In their minds, they had already bought the product and the bill just came due.

Good enough…

The natural instinct of Engineering is to keep adding features. But customer discovery is a race to get the MVP into paying customers’ hands as quickly as possible, so fewer features or an MVP that’s just “good enough” are far better than losing a month or even a week’s worth of customer feedback.

There are three critical questions to answer:

  1. Have we found a product/market fit?
  2. Who are our customers and how do we reach them?
  3. Can we make money and grow the company?

Hardest question

The hardest question is simple and needs an honest answer: do the customer discovery findings point to a big enough market that’s hungry for the product? This is often a painful question and, sadly, more often than not leads back almost to the beginning of customer discovery.


It’s wonderful to have lots of soft, fuzzy metrics of success, but they seldom deliver success to startups. Be sure that every hypothesis has a clear, measurable “validation checkpoint” and that those checkpoints tie into the business model.

Web/mobile checkpoint examples:

  • Every new customer invites 10 friends, half of whom sign up
  • A third of our visitors will return to the site within a week
  • A quarter of our visitors will refer an average of 1.5 friends within a week
  • Average session duration will be 10 pages or minutes per visit
  • Average order size will be $50 in the customer’s first month

Customer validation turns hypotheses into facts about basic questions like:

  • Do we understand the sales/user acquisition process?
  • Is it repeatable?
  • Can we prove it’s repeatable? (If our business model is single-sided commerce, the only acceptable proof is enough full-price orders.)
  • Can we get these orders/users with the current product?

Founder delegation

Founders who complete customer discovery often mistakenly ease up and delegate customer validation activities to Sales, Business Development, Marketing or Product Management. This is a bad idea. Middle and junior managers aren’t likely to be good at customer validation, which requires creative searching, probing, and turning on a dime—not execution of a repeatable process.

Dilemma of decision

Anyone other than a founder who learns of a serious product or business-plan flaw faces two challenges: he or she doesn’t have the authority to pivot, and he or she seldom has the courage to report bad customer feedback to the founder.

Business model canvas

The business model canvas is an excellent guide here. Most startups will focus on the four core elements:

  • value proposition
  • customer relationships
  • channel
  • revenue model

End of customer validation - epiphany moment

The moment customer validation is over is when it’s clear that there are real orders, users, or clicks—not surveys or chats. Customer validation confirms that customers will accept the minimum viable product, proves that the customers exist, figures out how to reach them predictably, and crafts a scalable plan to engage and sell many more. It’s often called the “epiphany moment.”

Promoting through customers

Encourage early customers or visitors to promote the product and your company to friends and business colleagues at every opportunity. Give them materials (e-mails, mailable links or demos), and consider rewarding them for doing so, as many online marketers do.

Word of mouth

Create YouTube videos and other shareable content that encourages exploration of the product. People respond to their friends’ likes and dislikes and are more apt to explore something recommended by a friend than something they see in an e-mail or an ad. Create robust profiles of company, product and users wherever possible on Facebook and other relevant social-networking sites.


The products you build should be instrumented to measure every click on the website or app, its source, and the action it causes or doesn’t. As a result, the management team should have at its fingertips a dashboard that summarizes every key, actionable user behavior, and delivers insights and trends that drive continual business improvement.

Number of metrics

Using your customer relationships hypotheses, identify the metrics for success in the business model. Prioritize and limit the number of metrics to fewer than a dozen, and the only metrics to be measured are those that can be acted on or improved.

Which numbers to worry about?

The metrics you use to measure and monitor your business ought to be the same ones you and your board are looking at during board meetings. If they’re only asking you for income statement, balance sheet and cash flow and not looking at these numbers, you’ve failed as a CEO. It’s your job to get your board to agree that the numbers you’re worrying about are the numbers they’re worrying about.

When to build a sales team (or not!)

Your goal is to determine who your true customers are and how they’ll buy your product. You build a sales team only when you completely understand the process that transforms a prospect into a purchaser and know you can sell the product at a price that supports your business model.

Thumbs up from early-evangelist is not a celebration time!

A common first time founder mistake is getting a “thumbs up” from an early-evangelist and opening the Champagne to toast the first sale. As any of you with experience in sales will know, don’t do it yet. Unfortunately, much can happen between the time the decision-maker agrees to make a purchase and you actually receive the check.

Value of advisors

Basically, you want two things from advisors: great introductions you can’t get any other way to key customers, talent or investors; and bold, out-of-the-box business model design thinking that will have dramatic impact on your strategy.

From requests to specs to features…

The art of being a founder lies in knowing when to incorporate those requests into the spec and declare them features. The danger and challenge is to avoid an inadvertent migration to a custom products business where all the economics change at a disadvantage to the company.

!! Never agree to exclusivity

From the customer point of view, this is a reasonable and rational request. “I’m going to take a risk on you, so I want a competitive advantage by being the only one with your product”—an exclusive. Or, “I want the best pricing forever” for being first. Tread very carefully here. Often these are requests from big-company execs with little real-world understanding of how a startup works. Don’t give up your rights to sell to other customers or you’ll end as a development arm of the first company that asks you for an exclusive.

Practice with lower-level employees

While entrepreneurs’ instinct is to get to the most important and highest-level executives they can find, keep in mind that you usually don’t get multiple meetings with “C-level” executives. Practice by calling on lower-level employees until you’re certain you understand you have a product/ market fit.

Don’t count your chickens until they are hatched

Particularly in business-to-business sales (as well as sales where husband and wife must agree), much can happen between the agreement to buy and the arrival of the check. Recognize and identify any steps required before the sale is completed and the product is delivered and paid for.


Before the company spends time on positioning, it’s a good idea to air out the conference room and get some facts. The best way to do this is with a positioning audit. An audit is an unbiased way to learn how others perceive your company and products.

Differentiation in an existing market can take one of three forms:

  • product attributes (faster, cheaper, less filling, 30 percent more)
  • distribution channel (pizza in 30 minutes, home delivery, see your nearest dealer, build it yourself on the web)
  • service (five-year, 50,000-mile automobile warranty, 90-day money-back guarantee, lifetime warranty)


Equally important, Walmart adoption of cutting-edge technology to track how people shop and buy, and its ability to deliver goods more efficiently and cheaply reduced its cost of sales to a small fraction of competitors’. By 2002, Kmart was bankrupt and Walmart was the largest company in the world.

There are three steps to answering the crucial “pivot or proceed” question

  • Assemble and review all key discovery and validation findings
  • Review the business model hypotheses and their interactions with one another
  • Focus on the “metrics that matter” in the financial model

Reports to graphs

The best way to “translate” piles of data, reports and questionnaires is to make it as visual as possible. For best results, lock the founding team in a room for a day or two to walk step by step through each hypothesis.

While not every diagram is appropriate for your startup, review

  • a work-flow map of the prototypical customer that diagrams how customers do their jobs or live their lives both with and without the new product
  • an organizational/influence map showing whom consumers or businesspeople interact with, how often, and how those people influence buying decisions
  • customer archetypes
    • how they earn and spend their money and use their time
    • a market map showing where your customers will come from
    • a channel or sales roadmap diagramming how sales will happen
    • a fully updated business model canvas


  • how many months’ worth of cash remain in the bank?
  • how fast are we burning through what’s left?

We guarantee that this different approach to finance will get you thrown out of just about every business school in the country, but it’s all that’s needed at this stage to determine whether your business model passes the validation test and is ready to move on from validation to scale in customer creation.

Five metrics that matter in this spreadsheet

  • Revenue comes from three channels Direct revenue from web sales, net of acquisition cost Direct mobile sales revenues, after subtracting referral incentives Revenue from the appstore channel, after subtracting channel fees and marketing costs
  • Cash burned or made during the period
  • Cash remaining at the end of the quarter

16 metrics that matter:

  1. New activated users
  2. New user acquisition cost
  3. Total active users


It’s time to reflect, thinking seriously about the company’s genuine chances for scalable, profitable success. It’s a hard decision, particularly for self-confident entrepreneurs who pride themselves on their tenacity, and problem-solving abilities. “I’m an entrepreneur. I make things happen against all odds” just doesn’t cut it when staring at the hard facts and statistics.