Author: Eric Rees
The Lean Startup: How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses Eric Ries 2011
Definition of a startup: total uncertainty (The Lean Startup)
The Lean Startup: A startup is an enterprise that offers a fundamentally new product or service. The key to defining a startup is a novelty, a journey into the unknown, and the risks associated with it. In this sense, a startup can be either a small, newly created company, or a branch of a huge, long-existing enterprise.
There is a stereotype of a “startup in a garage”. However, a company in a garage is a startup if it creates an unprecedented gadget, and not a startup if it offers car repairs.
When a division of an existing company acts as a start-up, its vision and strategy change radically. A new or old company that replicates an existing business is not a startup.
The need for a startup and the conditions for its existence are a changing reality and a high level of uncertainty. If market conditions do not change and there is no need to correct something in the system, startups do not arise.
The main features of a startup are the novelty of not only the product being produced but also the methods of production, the principles of business organization, the choice of the target audience, the dissemination of information, etc.
The product of a startup is innovation. Innovation can range from scientific discoveries and advances in technology to new business models, markets, and target audiences. The product of a startup is so new that many factors cannot be calculated in advance: what problems the company will face in production, what flaws in the product will have to be eliminated, whether customers will like it, whether competitors will overshadow the startup.
Clayton Christensen, in The Innovator’s Dilemma, described the underlying problem: often innovation comes down to improving existing products or services, ie “enhancing innovation.”The Lean Startup
It is especially difficult to create something truly revolutionary in an already existing company since innovations are revolutionary in nature, appear spontaneously and non-centrally, “from the bottom up”: they cannot be introduced by order of the management.
You should not go to the opposite extreme and think that the startup is unmanageable. The path of a startup is a path of trial and error, but these are conscious experiments and mistakes that you learn from, and not a mixture of brilliant insights and random steps taken.
The false romance of a startup (“the main thing is to get involved in a fight, and there you go or disappear”) is misleading. This extremely uneconomical way of experimenting most often leads to failure and ruins the reputation of a startup as a path of innovation.The Lean Startup
The idea of a lean startup is to produce a huge amount of experiments without spending extra resources – money and time.
What should be experiments in a lean startup:
- directed (what exactly we want to know);
- differentiated (identify problems);
- have precise parameters for assessing success;
- be accompanied by a willingness to learn;
- have constant feedback in order to correct the strategy at any time.
Minimum Working Product
Since a startup offers something unseen in an environment of high uncertainty, its leaders obviously do not know very much about their product and customers. They will have to learn as they go, figuring out which elements of the process and strategy and which product features will work and which will not.
A lean startup is a rigorous method that allows you to learn by trial and error not randomly, but by evaluating success according to certain parameters.
The main goal of lean manufacturing is to create value. Value is defined, as measurable benefits to the customer. Everything else is a cost.
The traditional tasks of management are mainly aimed at stability – meeting deadlines, fulfilling the budget. But what’s the point in a product, even if it was created on time and in accordance with the budget, if the client doesn’t need it?The Lean Startup
The scientific method of the Lean Startup: systematically breaks down a business plan into elements and tests each element. Every product, every option, and every marketing move is an experiment aimed at testing a hypothesis and obtaining sound knowledge.
An experiment is not “just to see what happens”. It begins with the development of a hypothesis, and then the hypothesis is tested empirically.The Lean Startup
The two main hypotheses are the value hypothesis and the growth hypothesis. The value hypothesis assumes that a product or service will attract customers. The Growth Hypothesis assumes exactly how and at what speed information about a new product will spread.
Hypotheses are often tested using a survey, but these data are not entirely accurate: it is difficult for people to objectively assess their intentions. Hypotheses are tested by facts: the behavior of the client.The Lean Startup
The experiment uses a minimum work product (MVP). The minimum work product is a simplified version that allows you to start the “create-evaluate-learn” cycle as quickly as possible and at a minimal cost.
This version does not yet have many options that may be needed in the future and does not answer all questions related to product characteristics. The main thing is to test the hypotheses of the founders of the startup. Make sure that this product is of interest to the client.
The main mistake of an entrepreneur: is to create a product and only then check the reaction of customers. But if the product turns out to be unnecessary, all expenses will go to waste.
Even if the product is successful, all the options that you can do without when testing is a waste of not only money but also precious time. The training cycle should be made as short as possible.
The basic principle of the minimum working product is to discard any ideas that delay the start of the learning cycle. Simplify everything you can.
Experimenting with MVP takes courage. Many companies are afraid to release a product that has not been perfected because they are afraid of:
- legal problems if customers make claims;
Of course, you should consult with lawyers, warn customers that they are “running in” the product, take care of its safety.The Lean Startup
- competitors who will steal the idea and get ahead with its implementation;
It is actually very difficult to steal an idea, especially at the stage when you have just started experimenting with it. If all your advantage lies in secrecy, it is better not to take the case. True competitive advantage is not even about being the first to market (although this is essential), but about going through feedback cycles faster than others, improving the product.The Lean Startup
- damage the reputation of the brand;
If you warn customers about possible problems, you will not ruin the reputation, but, on the contrary, will gain the reputation of a courageous and disruptive startup. Large companies with a well-known name sometimes create a separate branch for such purposes.The Lean Startup
- undermine employee morale by failure.
The concept of a lean startup is hundreds and thousands of experiments. If all hypotheses were correct and a significant proportion of experiments did not lead to failures, why do experiments? Then, indeed, the traditional method of perfecting the product and launching advertising would be preferable.The Lean Startup
An experiment should be treated as an experiment, and failure should be valued as an opportunity to learn. It is important to decide in advance that failure will not force you to abandon the project, but will encourage you to reconsider those elements of the project that do not justify themselves. The Lean Startup is the perfect combination of persistence and agility.
The quality of a product is only as important as it matters to the customer. You can’t neglect discipline and hack, especially wishful thinking (this will really cause irreparable damage to the brand). Defects that hinder product development or disrupt the feedback loop should also be addressed. But you do not need to strive for perfection at the first stage.
Many early adopters, especially in high-tech areas, even prefer to deal with an imperfect product: this allows them to take part in the development and customize the product to suit their needs.The Lean Startup
The minimum work product test is conducted primarily to clarify the customer archetype.
Hypotheses about value and growth are based on a fundamental assumption: the founder of a startup assumes that his product will be in demand. Such a “leap of faith” requires great courage.
The main uncertainty factor for a startup: we do not yet know who our client is and what is of value to him. The first thing to do is find out exactly that. Any assumptions about who will become a customer and why a product should be liked can be wrong. These assumptions should be considered hypotheses and empirical data from real users should be collected to confirm them.
The only proof: the version brought to life, bought by the client.The Lean Startup
Many companies order a “customer archetype” – a portrait of the target customer. This portrait uses many of the techniques adopted by the lean startup—minimum work product development, individual observations, and quick feedback. But, unlike the concept of a lean start-up, such developments are usually entrusted to an external agency or a special department of the company, on their basis a single and final proposal is created, and this is where the feedback-learning cycle usually ends.
The customer archetype is a hypothesis. It needs to be refined while improving product design: good design is one that influences customer behavior.
The experiment with the minimum work product is aimed at clarifying the customer archetype and the factors that influence the behavior of the customer, so the minimum work product can be not only a finished item or service but also a demo version or individual service.
To test whether the Dropbox interface is convenient, the creators showed the public a video demonstrating how this program works. The developers of Aardvark, before launching the program, served the customers themselves (simulating the operation of the program).The Lean Startup
It is very important to choose the right first users on whom the product is tested and even more importantly, to understand what information you want to get in the experiment. The first user can be one person on whom the customized form of the product is tested, a focus group, and a wide range of Internet users, depending on the features of the product. The key is genuine feedback.
You can not replace the experiment with a survey. First, people can’t accurately judge their intentions, and just because someone “generally likes” a product doesn’t mean they’ll buy, use, or recommend it. And secondly, an enthusiastic startupper unwittingly manipulates the survey data and clarifies the customer archetype “in his favor.”
When Eric Ries created his first product, he invited clients into the office and asked them to test the product. A positive response inspired hope for success, and a negative feedback was ignored by the startup: if a person does not like the product, it means that he is not included in the target audience.The Lean Startup
The problem with any company is focusing on the wrong indicators. Most often, success is measured by the overall performance achieved at the end of the year or marketing campaign – this encourages the use of the usual arsenal of tools: discounts, advertising, etc.
Such “vanity measures” do not answer the main question: why has customer behavior changed? what exactly makes them buy? what is the growth forecast?The Lean Startup
The experiment is aimed at understanding the behavior of customers. Instead of collecting general indicators, a cohort analysis should be carried out, taking into account data for each consumer group. To identify specific factors influencing customer behavior, split testing is used: product versions are simultaneously offered to several groups of customers, and differences in the behavior of control groups are revealed. The main thing is to ask a question before the experiment (what exactly we are testing), and not to build causal relationships in hindsight.
Innovation Accounting Method: Can a Startup Grow?
Experimentation will be a waste of time if the results do not induce the company to change (sometimes dramatically) its strategy.
The experiment is the basis of the quantitative method of accounting for innovations. It triggers the create-evaluate-learn feedback loop.
Experiment planning works in reverse order: first, the startup decides what needs to be learned, then what needs to be assessed by accounting for innovations in order to obtain the necessary knowledge, and finally, what version to create for the experiment.The Lean Startup
The innovation accounting method is an alternative to the traditional system, a planned method that allows you to evaluate the company’s success and confirm it with facts.
At Toyota, which is the foundation of the lean start-up, the main slogan is “genchi genbutsu”: “go and see for yourself”, get first-hand knowledge.The Lean Startup
Innovation accounting training consists of three stages:
- Create a minimum work product for a basic experiment.
- Product optimization – micro-changes, after each of which new experiments, split testing, and cohort analysis are carried out.
- Decision: whether a turn is needed.
The minimum working product allows the first stage to determine the basic parameters of the growth model: customer acquisition rate, product life cycle, etc.
First of all, it is necessary to check those initial hypotheses embedded in the business plan that are associated with the greatest risk. If the barriers to sustainable growth cannot be removed, the entire startup concept should be abandoned in time.The Lean Startup
At the second stage – product optimization – the growth mechanism is tuned. Any new option – whether in product development, marketing, pricing, etc. – aims to strengthen growth drivers.
The viable growth of the company is provided by customers attracting new customers.
Main sources of growth:
- word of mouth;
- the appeal of fashion products seen by early adopters;
It is necessary to distinguish between the objectives of advertising: whether it is aimed at attracting new or retaining old customers.The Lean Startup
- repeat customer referrals.
Companies try to incentivize repeat referrals through subscriptions, discounts, etc. But in some cases, repeat referrals are undesirable (for example, warranty repairs) and the focus should be on new customer acquisition rates.
You need to determine your growth mechanism and the indicators that are important to it.
- Sticky growth is customer retention. If the rate of new customer acquisition exceeds the loss rate, then “accumulation of customers” occurs.
For sticky growth, advertising and the speed of dissemination of information are not very important. The main task is to make customers reapply more often: offer discounts for loyalty, send out catalogs, add options.The Lean Startup
- Viral growth is the rapid, “epidemic” spread of information from the client to client. The viral growth mechanism can also be tested and quantified. The main parameter is how many people each client can bring along on average. With a viral growth mechanism, attracting new customers is in itself a value for the company, so companies often choose to distribute their product for free, deriving indirect income (most often from third-party advertising).
An example of viral growth is social networks. Each user by the very fact of using it involves more and more new customers. The forerunner of social networks is network marketing.The Lean Startup
Again, social networks, email servers, etc. are an example.The Lean Startup
- paid growth. The cost of attracting new customers is inevitable (not only advertising but all the experiments with new options and quality optimization). The only question is the ratio of income from the appearance of new customers and the cost of attracting them.
For paid growth, first of all, marketing is important, which can be ignored with viral growth.The Lean Startup
Growth is not due to product improvement, but due to an efficient and well-chosen growth mechanism. The mechanism of growth at each stage of a company’s life can slow down, it needs to be changed, and this again requires a constant feedback loop.
The value of the scientific method of accounting for innovations is that it saves resources as much as possible, allows you to monitor the viability of the company, and correct the course in time. At the first stage in the feedback cycle (“create”), the experiment determines the attractiveness of the product to customers, at the second (“evaluate”) growth mechanisms are optimized, and at the final stage (“learn”), conclusions are drawn – and turns.
A pivot is a maneuver in which a startup builds on what it has already learned and makes a U-turn. Types of turns:
- The pivot of the consumer segment.
From B2C to B2B.The Lean Startup
- Zoom – what has been considered an option becomes a staple.
Instead of a social network, Votizen has created a service for communicating with voters.The Lean Startup
- Reducing – a separate product is included as an option in the overall functionality.
- Displacement – a minor product displaces the main one.
Such a turn was made by the Potbelly Sandwich Shop chain, which originally sold sandwiches to attract customers to an antique store.The Lean Startup
- Platform Pivot – transition from platform to application and vice versa.
- The pivot of customer needs is one of the most important. A need that is more important than the one addressed by the original product may come to light.
Pivot is not only applied to the product – it is possible to pivot the business architecture, growth mechanism, monetization method, distribution channel, etc. Pivot – structured changes aimed at testing another hypothesis about the company’s product, its business model, and growth mechanism.
The viability of a startup is determined by how many turns it has in stock , and how many times it has time to make fundamental changes to its strategy. The faster the turns are made, the longer the “runway”. All Lean Startup methods aim to lengthen the runway.
Time to make a turn prevent:
- “vanity indicators” (the company collects general growth data and does not see the need to urgently change something);
- fuzzy strategy, unwritten indicators (again, failure goes unnoticed);
- fear (pivot is a rejection of the original strategy).
The main way to lengthen the “runway” without attracting additional financial resources and spending time is to speed up the feedback, and for this:
- speed up the “create – evaluate – learn” cycle as much as possible;
- simplify the basic product as much as possible;
- test one element of the product or business model in each experiment;
- constantly analyze the collected data, make reporting available to all employees;
- turn-on time.
“Just in Time”: Japanese Lean Manufacturing Method
The principles of lean manufacturing were developed by the Japanese company Toyota.
- Instead of expensive specialized equipment – simple machines.
- Small batches and small stocks in a large number of warehouses.
- Quickly reconfigure machines and the entire production.
Production and supply in small batches seemed counterintuitive. However, it is this approach that provides benefits for both the company and the client:
- Small batches are delivered more frequently and do not have to wait.
You need a part of a certain color for a car. It is promptly delivered from the nearest warehouse, the central warehouse receives an order for replenishment of stocks, the plant – for the shipment of this part. If feedback works regularly and orders are fulfilled daily, the client does not have to wait for planned deliveries in a month.The Lean Startup
- There are no expired and unclaimed goods.
If the goods are delivered to the grocery store once a day, a lot is taken with a margin – and becomes stale the next day.The Lean Startup
- For the company, the advantage of small batches is faster feedback: shortcomings will be noticed faster.
The andon principle at Toyota plants: any worker can stop production if they notice a problem.The Lean Startup
Offering one option after another in split testing is the same small-batch method. The “immune system” works:
- Unsuccessful changes are removed.
- All employees involved in the problem receive a notification.
- Before proceeding further, an analysis of the error and its causes is carried out.
Startups are gradually adopting this method of continuous deployment.
Five why method
This method was developed by Toyota and adapted for a startup – when parsing an error, you should look for the cause, not the culprit.
- Why did the crash happen?
An overload has occurred. There was a problem with the server.The Lean Startup
- Why did this situation arise (technical reason)?
Lubricant was missing. The subsystem was not working properly.The Lean Startup
- Why was the correct operation not ensured (system error)?
The pump is not working well. The operator does not know how the subsystem operates.The Lean Startup
- Why is there a system error?
The pump is worn out. The worker was not trained.The Lean Startup
- Why is there a possibility of such a system error (what was overlooked)?
The pump does not have a gasket. This department does not train employees.The Lean Startup
With the help of five “why” you can get to the root cause of the problem – that’s what needs to be eliminated. This method serves as a speed regulator: on the one hand, a startup should speed up the “create-evaluate-learn” cycle as much as possible, on the other hand, spend time-solving problems, and most of the time is spent on solving problems just at the beginning of the startup life cycle when the especially great temptation to hurry.
How not to spend too much time experimenting and solving problems
- Delegate an experiment to a cross-functional team that does it completely, from start to finish, without asking minute-by-minute permission from superiors.
- Allocate the experiment exactly at the agreed time.
- Limit the group of clients covered by the experiment.
- Evaluate each experiment against a single report containing less than 10 standardized scores.
- Track consumer reaction: if customers are strongly opposed to new versions, immediately stop the experiment.
Innovation Incorporation is a method that enables cross-functional teams to act independently, quickly, and efficiently, acquiring and applying fact-based knowledge in a constant feedback loop. At the moment, this method is the basis for a lean startup, but it is possible that new methods will appear in the learning process. Moreover, the emergence of new methods should be expected, because the essence of a lean startup is not to create a successful business but to learn how to learn, to quickly separate what works well from ineffective ones.
10 main thoughts
1. Startup – an enterprise offering a fundamentally new product or service, mastering new production processes or new markets. A startup operates in conditions of high uncertainty, so traditional management methods in relation to a startup are obviously losing.
2. A Lean Startup produces a huge amount of experimentation with minimal investment of money and time, with a fast feedback loop. Thrifty does not mean “cheap”, but “non-wasteful”.
3. To avoid wastefulness, you need to experiment with a minimum work product, avoid unnecessary costs, maintain a feedback loop of “create-evaluate-learn”, and apply an innovation evaluation method – an accurate quantitative calculation of the return from each innovation.
4. The minimum working product is used in the base experiment, all options are added after potential customers show real interest in the minimum working product.
5. The method of assessing innovations is aimed at accelerating the “create – evaluate – learn” feedback. It involves three stages of learning:
- Creation of a minimum working product for the first experiments.
- Numerous micro-optimizations were followed by split testing.
- Frequent check of achieved parameters and appropriate course correction.
6. At the stage of creating a minimum work product, the archetype of the client and the factors influencing his behavior are specified. The experiment should focus on real indicators of customer behavior, and not on a survey of intent.
7. At the optimization stage, growth mechanisms are tested. At the course correction stage, a decision is made on the need for a turn.
8. Following the rules of Toyota’s lean manufacturing, the startup works with small batches and a minimum set of options, striving to do everything “just in time”.
9. Although the speed of feedback cycles is extremely important for a startup, it is equally important to stop and think through the reasons for each failure, applying the five “why” method.
10. These principles can be applied not only in business, but also in public-private partnerships, politics, and in schools.