Le manuel du propriétaire de démarrage : résumé et examen
barre de progression
The Startup Owner’s Manual: Summary and Review
Critique de livre
The concept of The Startup Owner’s Manual is that it’s like a car repair manual but for startup owners. In this book, Steve Blank and Bob Dorf detail all the work (and fun) that needs to happen before launching a product. With so much information, it’s a big book. In fact, the size is probably how it most resembles an auto manual. “Don’t read too much at a time.” This caveat appears before the book even starts. Take it seriously.
The organization can be confusing. While this manual takes the reader sequentially through the necessary tasks of starting up the startup, it follows two different tracks at the same time. There are the Web/Mobile Channels — your IT, new-economy type channels — and the Physical Channels — brick and mortar. The authors say it’s helpful to read both sections, regardless of what channel your product happens to be. After some initial confusion, this reviewer personally found it thought-provoking to contemplate both channels side by side. And the differentiated font is useful.
With all the information presented here, it’s still easy to get lost. To help with that, there are gobs of checklists. So if checklists are what you need, by God, you will find them here (the appendix contains 60 pages of checklists).
Getting Started (Chapters 1 & 2)
The authors remind us that the old product roll-out process is totally wrong for startups. That process is appropriate when customers are known and the market is well-defined. This is often not the case for startups.
Customer development, then, is very important for startups, and there are a number of phases in Blank and Dorf’s framework:
- Customer Discovery: develop hypotheses and test them with customers.
- Customer Validation: test sales, see if people will buy, and see if you can scale up.
- Customer Creation: marketing.
- Company Building: transition to a sustainable enterprise.
Each step is iterative and can be repeated, as needed. Furthermore, at any point, a pivot or change of strategy may be necessary. And that’s OK — it’s not a failure.
Step One: Customer Discovery (Chapters 3–7)
You have to understand your customers; don’t make assumptions about them. One of the seminal points that Blank and Dorf make (as do many others) is that customers do not live where you work. In order to meet them at their level, you have to “[g]et out of the building.”
Startups should aim to develop the first product for a small target market via a Minimum Viable Product (MVP). The object of the MVP is to get a product out there for the early adapters (called earlyvangelists here) to play with. Putting out a MVP forces developers to focus on the important features, not the bells and whistles. The product can be refined once there’s customer feedback.
Business Model Planning: The authors recommend the Canevas du modèle commercial (laid out in Alexander Osterwalder’s Génération de modèles commerciaux) as a good tool to help explain how the company plans to make money. There are a number of components to the canvas:
- Value Proposition Hypothesis: Includes product vision, features and benefits, and a description of the smallest number of features it would take to make a stand-alone product (i.e., the MVP). Along the web/mobile track, the low fidelity MVP is used to test if you’ve identified a problem people care about and it employs user stories instead of feature lists.
- Customer Segments: Includes customer problems, needs or passions; customer types (for example, end users, influencers, recommenders, buyers, etc.); customer archetypes; a day in the life of a customer; and a customer organization and influence map.
- Channels Hypothesis: For physical products, this includes a description of how the product gets to the customer. Think about the suitability of channels for your product, and the authors recommend that startups pick just one channel that has the most potential for your product. For web/mobile products, consider the advantages and disadvantages of different channels.
- Market-Type and Competitive Hypothesis: Describes which market a product fits into: established market, new market, extension of an existing market, re-segmentation, etc. Once you know the market, you know something about the competition, and you can come up with a preliminary plan for that.
- Customer Relationships Hypothesis: Describes how you will get, keep, and grow customers. For physical products, there are four “get” stages — Awareness, Interest, Consideration, and Purchase. Devise several different “get” strategies and test them. Customer retention programs come into play during the “keep” phase, and the “grow” phase focuses on selling more stuff to existing customers. For web/mobile products, the “get” phase has two stages, Acquire and Activate, which loop back on themselves. The “keep” phase might involve some of the same retention programs as for physical products, but might also include components like emails, digital support, etc. The “grow” phase involves getting current customers to spend more or bring other customers to the product.
- Key Resources Hypothesis: Describes physical, financial, human, and intellectual property resources. This component also includes a dependency analysis: what are all of the things that have to happen that the company can’t control, and what are the risks if the things you’re dependent on end up failing? Contingency plans are important, here.
- Partners Hypothesis: Involves listing all of the partners you’ll need and identifying what you need from them and what they’ll get in exchange.
- Revenue and Pricing Hypothesis: Includes the following questions: 1) How many items will we sell? 2) What’s the revenue model? 3) How much will we charge? and 4) Is there evidence that this is worth pursuing?
Use this tool beyond a one-time plan and update it once a week. Over time, you’ll have a record of the company’s evolution.
Size of the Market: When it comes to estimating the size of your market, one way to think about it is as follows:
- Total Addressable Market (TAM): the universe where your product lives
- Served Available Market (SAM): the people that can be reached via your sales channel
- Target Market: those who are your most likely customers
You’ve got to do research here. Look at industry reports, press releases, libraries — anywhere you can get metrics that help you gauge the size of the market. It’s going to be tricky to estimate brand new markets, so consider looking at adjacent markets and see if you can find any comparable companies.
Testing Customer’s Problems: You have to get out of the building to discover:
- How well you understand the customers’ problem
- How important the problem is to the customers, and exactly how many customers are talking about it
- If the customers care enough about it to tell their friends
Broadly, the problem discovery process includes the following five steps (and note: this process goes faster for web/mobile products than for physical products):
- Design experiments for customer tests: Short, simple, objective pass/fail tests that aim not just to collect data, but also to gain insight.
- Get ready for customer contact: For physical products, contact potential customers that you can test your ideas on. For web/mobile products, this involves developing a low fidelity MVP, which could be as simple as a brochure or web page that states the problem, shows people your solution, and solicits customer feedback.
- Test how customers perceive the problem: A problem presentation (not a product demonstration) for physical products that illustrates your understanding of the problem and your proposed solution. For web/mobile products, this is where you test your low fidelity MVP, validating your hypothesis and all your assumptions about the customer.
- Understand customers: This goes beyond simply validating assumptions about customer problems; it entails understanding how customers go about their day/lives and how they are (or are not, as the case may be) currently solving the problem.
- Get information on the market and competition: Understanding the environment in which a physical product is operating may involve going to trade shows, conferences, competitor lunches, etc. These tools also work for web/mobile products, as do traffic-measurement tools like Alexa.
Testing Your Solution: Following customer problem testing, this phase tests the solution, and the authors detail five steps here:
- Update the business model and team: See how the new information fits with the old hypothesis, and decide if a change of strategy (pivot) is needed or whether proceeding makes sense. This is also the time to share what you’ve learned so far about the customer with top management and investors.
- Create a product solution presentation (physical) or high fidelity MVP test (web/mobile)
- Test the product solution: See what customers think by asking them about everything, including features, channels, etc. And for web/mobile products, measure customer behavior.
- Update the business model (yes, again): Using what you’ve learned from the solution tests, update the business model and decide (yes, again) whether to pivot or proceed. Low customer enthusiasm at this stage is a huge red flag!
- Find board members: It’s good to get some friendly outside help.
Verifying the Business Model: Still not at launch yet, this phase of the customer discovery process involves answering three critical questions:
- Have you found product-market fit? You have to be sure your product is a good fit with the market. Is this something a lot of people need? How well does it solve the problem it’s made to solve? How many people would buy this thing?
- Who are your customers and how do you reach them?
- Can you make money and grow the company? Figure out if your solution is a winning proposition. Crunch the numbers and do a rough estimate to see if there’s any chance that you can make money with your new product.
And, you guessed it, at this point you should once again pause and decide whether to pivot or proceed.
Step Two: Customer Validation (Chapters 8–12)
Customer validation means test selling at every point in the process (all the while keeping costs low — it’s not the time to scale up yet). Blank and Dorf detail four phases in the customer validation process:
- Phase I. Get ready to sell.
- Phase II. Go live. Try to sell.
- Phase III. Refine your product. Position your company.
- Phase IV. Analyze. Pivot or proceed.
Founders cannot lead from a distance here; they have to be directly involved in process.
Phase I — Get ready to sell
Steps for this phase of customer validation really are different for each channel, so it’s helpful to consider each separately.
For physical products, this phase begins with making a positioning statement: a message (it should be pithy but compelling) that explains why people should buy your product. Then, customer-focused sales and marketing materials should be developed largely from the information that you generated in the hypothesis you created for the customer discovery process. (Remember, acquiring customers in this channel is a four-stage process of awareness, consideration, interest, and purchase.) The authors recommend that this is an opportune time to bring sales professionals on board.
You and your team will also need to develop a sales channel roadmap which covers the organizations in the organizational food chain, the relationships in the distribution channel, and how money moves in the channels. When it comes to managing channels, work on the relationships with your channel partners. Think about what you need from them and what they need from you. And don’t expect your channel partners to help generate demand.
Then you will develop the sales roadmap, which accounts for every step from the first sales call to the contract signing. Keep in mind what you’ve learned about the customer and revisit some of the materials (especially the organization and influence maps) you developed back in the customer discovery stage. Make a model of the purchase process, and identify key influencers.
A Customer Access Map can help you identify organizations that customers belong to (if you’re selling to the public) or key deciders in the organization (if you’re selling to a business). Put all the maps next to each other, also considering the sales strategy, and write an ImplementationPlan that shows everything that has to happen before selling your product.
For web/mobile products, the Get Ready to Sell phase involves first refining plans and developing tools to acquire and activate customers. Remember: acquisition is when the customer first hears about the product and activation is when customers participate, enroll, or do something. Don’t launch the Acquire effort unless the Activation program is ready for customers, and the key to the latter is engagement. Also, measure everything!
Then, using the information you’ve gathered to build previous versions, you’re going to build a high fidelity MVP, which will have more features than the last one. It’s definitely more polished, but it’s still not the complete, finished product. Be sure to capture as much data as possible on customer reaction to the new MVP. From there, collect data, analyze and optimize — so it goes from the moment a web/mobile business opens its virtual doors until the day it dies.
Use the customer relationships hypothesis to guide your selection of metrics to track and prioritize the metrics. Only track things you can measure and improve: acquisition (How many visitors to your site? Were they referred from another site?); activation (How many activations? What pages did they look at on the site?); and referrals (How many people were referred from existing users? What’s the acceptance rate of these referrals?) Use a dashboard or something similar to keep an eye on the data. Speaking of data, you need someone to crunch the numbers, and Blank and Dorf recommend hiring a data analytics chief. This person should be in senior management and they should have clout. It can be someone already on your team or someone new.
Finally, in an author recommendation that is applicable to both physical and web/mobile products, recruit and keep only board members who can help you. Do they have important connections? Do they have really good ideas that could help the company? Lots of times, it’s helpful to have customers on the board. They can bring their perspective.
Phase II — Get out of the building and sell
In this phase, you are validating your business model hypothesis, but you are still testing (by making real sales). Don’t try to scale up yet — the object of your sales is to test your business model.
For physical products, you don’t need a lot of customers, just a few. Remember those early adopters/earlyvangelists? These people are your natural market. But don’t forget that what motivates them probably isn’t what motivates your average Joe. There are different kinds of customers (for example, early evaluators, earlyvangelists, scalable customers, and mainstream customers), and you should have a separate strategy for each of them.
Customer interest for physical products is tested by seeing how much people will pay for the shiny new thing. Maybe give people a discount from (what you say is) the full price, but not too much. (Predictably Irrational par Philip E. Tetlock et Dan Gardner has some great insight on pricing schemes.) In any event, get out there and sell. Once you have sales, you will have data. Collect it all. But pay attention: Did you get enthusiastic customer response with your Shiny New Thing™? Not so much? Is it because of something that needs to be fixed with the test sell, or is it something with the product? Watch for your pivot or proceed moments here.
Your approach to sales will depend on whether you’re selling to consumers or companies. With companies, you’re trying to reach executives with decision-making power. With the public, you’re segmenting the market. Pull out your old Sales Roadmap; that’s an important way to communicate with your sales VP. Write the roadmap as a flow chart, and include each step of the process. Don’t forget: after a sale is made, someone has to follow up with the actual transaction and see that the customer gets their thing. Make sure this is part of your roadmap.
Finally, once you have customers, you can test the sales channel. Approach a potential channel partner and share some of your metrics and projections; tell them your idea. You should have questions for them about their channel — for example, what kind of percentage do they want? They might go for your offer right away, or they might want to test your idea first by maybe selling your product in their stores first to see if they do well. Check out a number of different potential channel partners.
Shifting gears, for web/mobile products, you need to prepare optimization plans and tools. Optimization is all about getting more out of everything — for example, if you have a 6% activation rate, try to push it up to 10%. Try for measurable improvement in each thing you measure. This is a continuous process that won’t end until the business either folds or goes public. Optimization should be focused on increasing volume, reducing cost per activated customer, and increasing the conversion of visitors to users. The authors offer some advice on optimization testing:
- know what you’re testing and why
- don’t over-test
- tests should be controlled and follow accepted protocols
- remember the value of the customer over their lifetime, not just for this one transaction
For web/mobile products, the strategy for getting more customers depends on your business model. For example, subscription services might offer a free trial to new users. Be sure to attend to issues: what’s important, what’s disappointing, what results in the best customers, etc. Test a wide range of incentives and discounts. And don’t optimize too many things at once. You need to be certain about causes and effects.
But do optimize the “keep” and “grow” components. It’s cheaper to keep old customers than to get new ones. When focusing on “keep,” go back to your customer relationship hypothesis, where you developed some ideas for customer retention (loyalty programs and so forth). Launch a few of these. Be sure to track everything. Some programs are cheaper than others, so track your costs. In the “grow” segment, get your customers to buy more (perhaps give discounts for large orders and similar tactics), and have your existing customers draw in new customers.
Finally, you need partners to drive traffic to your new website, and there are strategies for using partner sites to promote yours. But don’t be surprised if potential partners aren’t excited by your cold call. Try to emphasize the benefits of partnering with you.
Phase III — Positioning phase
Your positioning should match your market. For existing markets, you want to be different, yet credible. You want to solve a problem that’s important to people.
For a new market, you want to explain your vision. What is your company trying to change?
At this point, go back and look at the Positioning Statement you wrote during the customer validation phase. How did it survive the validation process? If the customers didn’t love it, do you understand the reasons why? Also, meet with the key influencers that you identified back in the customer discovery phase. Now is the time to meet with them. Test to see if these influencers and other industry analysts will say good things about you.
Finally, don’t let anyone tell you that you need a PR agency — your customer development people, with input from the product development people, can do the job just fine at this point. But do conduct a position audit to see how others see your company and your product. This will give you a baseline to compare against.
Phase IV — Pivot and/or proceed
This is probably the most critical pivot and proceed you will have. Review and analyze all the material you’ve generated; make sure that there are solid answers to the hypotheses. Look at your business model and how it’s changed through this process. Do your metrics provide evidence that this is a scalable endeavor that could turn a profit? This is important, and the authors provide insight into “Metrics that Matter” (for example, total number of units sold, selling price, etc.).
Review your financial models. Revenue growth curves are different depending on the market type: growth in a new market company will look like a hockey stick on a graph, slow at first and then taking off; an existing market shows a steady rate of growth. The numbers that you have are your best guess, but remember that nothing is certain, even this late in the game. Also reevaluate the business model, value proposition, and product delivery to be sure they are the right fit. The revenue should be high and the costs low.
In short, look at everything again. It can be hard after all this work to go back to the drawing board, but it’s OK if that’s what you must do. Otherwise, if everything checks out, take a night off. No, make that a week. You’ll need the rest. There is still plenty more work ahead.