Chapter 12: Innovation In a Startup

Innovation teams should be correctly structured to succeed. Venture-backed startups usually have some structural features because of being small, independent businesses. Internal startup teams need support from senior management to create these structures. According to Eric Ries, regardless of having internal or external teams, startups need three structural features:

1. Scarce but secure resources.
2. Independent authority to build their business.
3. Personal risk in the result.

We will now go through each of these features one by one.

12.0 Scarce but secure resources

Division leaders in large established companies are experts in using politics to increase their budgets but know that those budgets are a bit unstable. They often gain as large a budget as possible and prepare to save it from other departments. If a crisis occurs in another department, their budget might reduce by 10%.

On the other hand, with startups, too much budget is as harmful as too little. It’s very rare for a startup alone to lose 10% of its cash in hand suddenly. This is dangerous in many cases as startups are usually not open to errors. Therefore, startups are easier and more demanding to run than traditional divisions as they require less capital, but that capital must be secure from tampering.

12.1 Independent authority to build the business

Startup teams need complete freedom to build and market new products in their limited time. They should conceive and execute experiments without having to acquire an excessive number of approvals. Handoffs and approvals slow down the feedback loop and inhibit learning and accountability.

12.2 Personal risk in the result

Entrepreneurs need a personal risk in the outcome of their creations. In stand-alone startups, this is achieved by stock options or other types of equity ownership. Where a bonus system should be used, the best incentives are connected to the long-term performance of the innovation. However, Eric Ries doesn’t feel that a personal risk should be financial. This is important in non-profit and government organizations where the innovation isn’t connected to financial objectives. In such cases, teams can have a personal risk. The parent company should clarify who the innovator is and ensure that the innovator gets credit for bringing the new product to life if it’s successful.

12.3 Creating a platform for experiments

Next, it’s essential to focus on creating the principles on which autonomous startup teams function. That is, how to protect the parent organization, make entrepreneurial managers accountable and recombine an innovation into the parent organization if it’s successful.

12.4 The dangers of concealing innovation inside the black box

Without the ability to experiment more smartly, a company would suffer the ‘Innovators Dilemma’, which is ever-higher profits and margins every year until the business suddenly dies. We mainly place internal innovation challenges by asking how we can protect the internal startup from the parent organization. But Eric Ries reverses the question and asks how we can protect the parent organization from the startup. In his experience, people defend themselves when they feel threatened, and innovation cannot succeed if defensiveness is given freedom. This is why the suggestion to conceal the innovation team is misunderstood. There are examples of one-time success stories of using a secret offsite innovation team to build the original IBM PC in Boca Raton, Florida, utterly separate from mainline IBM. However, these examples should be cautionary stories as they have hardly led to sustainable innovation. Hiding from the parent organization can cause adverse outcomes in the long run.

When you consider this from the managers’ point of view, they may feel betrayed and a little paranoid for something like this to be hidden from them, and thus, more things could be hidden behind their backs. However, the fact that innovation was a success doesn’t justify the dishonesty. It’s, therefore, unfair to criticize these managers for their response. Instead, the criticism should be aimed at senior executives who failed to build a supportive system to operate and innovate. This is a potential reason IBM lost their leadership position in the new markets it developed using a black box like the PC business – they could not recreate and preserve the culture that led to the innovation.

12.5 Making an innovation sandbox

In this case, the challenge is creating a process that clearly empowers innovation teams. This is the path to a sustainable innovation culture, as companies face continuous experimental threats. Therefore, Eric Ries has suggested creating a sandbox that will include the impact of the new innovation but not limit the methods of the startup team. It works like this:

1. Any team can make an actual split test experiment that impacts only the sandboxed parts of the product or service for a multipart product, only particular customer segments or territories for a new product. However,

2. One team should observe the whole experiment throughout the entire process.

3. No experiment can run longer than a certain amount of time.

4. No experiment can impact more than a certain amount of customers.

5. Each experiment needs to be evaluated based on a single standard report of five to ten actionable metrics.

6. Each team that works inside the sandbox and each product being developed should use the same metrics to evaluate success.

7. Any team that experiments should observe the metrics and customer reactions while the experiment is in progress and cancel it if something fatal happens.

Initially, the sandbox should be small. Based on the types of products the company makes, the sandbox size can be defined in many ways. For example, an online service might limit the sandbox to specific pages or user flows. A retail operation might restrict it to particular stores or locations.

Unlike in a concept test or market test, customers in the sandbox are real. The innovation team can create a long-term relationship with them because they might experiment with those early adopters long before their learning milestones are achieved. The innovation team should be cross-functional and maintain a fixed team leader whenever possible. It should be allowed to build, market and release products or features in the sandbox without early approval. It should report the success and failures of those efforts by using standard actionable metrics and innovation accounting.

This approach can be used by teams that have never worked cross-functionally before. The first few changes, like price change, might not need excellent engineering effort but require coordination from the engineering, marketing and customer service departments. Teams like this are more productive if measured by their ability to create customer value.

Real experiments are easy to divide as successes or failures as top-level metrics move or they don’t. In either case, the team learns quickly if its assumptions on how customers will behave are correct. By using the same metrics every time, the team becomes educated about the metrics across the company. As the innovation team reports its progress using innovation accounting, anyone who reads these reports learns a lesson about the power of actionable metrics. This impact is very powerful. Even if someone wants to ruin the innovation team’s efforts, they must learn all about actionable metrics and learning milestones to do it.

The sandbox promotes quick iteration as well. When people have an opportunity to see a project from top to bottom, and the work is done in small batches and provides a clear decision quickly, they benefit from the power of feedback. Every time they fail to move the numbers, they have a good chance of quickly acting on their findings. By making the batch size small, the sandbox approach helps teams to make small mistakes quickly and start learning. Next, we will see that these small initial experiments can show that a team has a viable new business that can be merged into the parent company.

12.6 Accountability of internal teams

With an internal startup, the sequence of accountability is similar to building an ideal model of the expected disruption that is based on customer archetypes, launching an MVP to create a baseline and then trying to tune the engine to get it closer to the ideal.

By operating in this framework, internal teams naturally act as startups. As they succeed, they must integrate into the company’s whole portfolio of products and services.

12.7 Nurturing the management portfolio

When an internal startup grows, the entrepreneurs who created the initial concept should overcome the challenge of scale. Then, as new mainstream customers are gained, and new markets are won, the product becomes a part of the company’s public face with important significance for PR, marketing, sales and business development. In addition, the product will likely gain competitors like copycats and fast followers.

Once the market for the new product is made, procedures become routine. To fight against the inevitable commoditization of the product in its market, line extensions, incremental upgrades, and new types of marketing are essential. This may need a new kind of manager who excels in optimization, delegation, control and execution. Company stock prices depend on this type of predictable growth.

Then there is the domain of outsourcing, automation and cost reduction which is the fourth phase in addition to the Build-Measure-Learn phases. Nevertheless, infrastructure is yet mission critical. Failure of facilities, essential infrastructure, or the lack of loyal customers could ruin the whole company. However, unlike the growth and optimization phase, investments in this area won’t help the company achieve growth. Managers in this kind of organization are criticized when something goes wrong or unappreciated when something goes well.

We prefer speaking about these four phases from the point of view of large companies in which they may represent whole divisions and thousands of people. That’s practical as business growth in these extreme cases is the easiest to watch. However, companies engage in all four phases of work every time. As soon as the product is introduced to the market, people work hard to progress it to the next phase. Each successful product or feature began life in R&D and later became a part of the company’s strategy, was subject to optimization and ultimately became old news. This is one of the reasons why companies struggle to find creative managers to foster innovation. Each new innovation competes for resources with established projects, and talent is one of the scarcest resources.

12.8 Entrepreneur is a designation

When products shift between phases, they are handed off between teams. Employees can select to move with the product as part of the handoff or stay back and begin to work on something new. Neither choice is correct or wrong as it depends on the temperament and skills of the particular person.

Some people are natural investors who prefer to work without pressure and expect the business phases. Others are ambitious and view innovation as a path to senior management. People should therefore search for jobs that suit them best. Entrepreneurship should be viewed as a viable career path for innovators in large organizations. Managers who lead teams by adopting the Lean Startup method shouldn’t leave the company to achieve rewards for their skills or pretend to fit in the strict hierarchies of their departments. Instead, they should have a business card with “entrepreneur” under their names. They should be accountable under innovation accounting and be promoted and rewarded.

Once an entrepreneur has nurtured a product in the sandbox, it needs to be reintegrated into the parent organization. A larger team will eventually be required to grow, commercialize, and scale. At first, this team will need the continued leadership of the innovators who worked in the sandbox. This will positively impact the process as it allows innovators to train new team members in the new way of working that they mastered in the original sandbox.

Ideally, the sandbox will eventually grow without moving the team out of the sandbox and into the company’s usual routines. There may be opportunities to widen the sandbox. For example, the online service example we saw could be achieved by starting with a sandbox encompassing the product pricing page. When those experiments succeeded, the company could add the website’s home page to the sandbox. It might later add the search functionality or the whole web design. If certain customers or a certain number of customers were targeted at first, the product’s reach could be increased. When such changes are expected, the senior management must consider if the teams working in the sandbox can take care of themselves politically in the parent company. The sandbox was made to protect them, and the parent company and any expansion need to keep this in mind.

12.9 Becoming the status quo

This last transition is tough for innovators to admit – their transformation from radical outsiders to the incorporation of the status quo. When Eric Ries was running his product development team at IMVU, he had to hire new people, and they had to be brainwashed into the Lean Startup culture. Split testing, repeated deployments and customer testing were all usual practices. He had to be a strong advocate for his Lean Startup ideas by ensuring that new employees were willing to give them a try. However, those ideas had become a part of the status quo for those working there for a while.

Like many entrepreneurs, he had to balance advocating his ideas and to listen to his employees’ suggestions on how his ideas could be improved. It took him a lot of effort to bring these suggestions seriously. Some of the suggestions that were considered contributed to the success of IMVU. Those who consider adopting the Lean Startup as a defined set of steps won’t succeed. In a startup, things can constantly go wrong. If we’re in the middle of adopting a new way of working, we tend to always blame the new system for the problems that arise. Sometimes that tendency is correct, and sometimes it’s wrong. Learning to tell the difference needs theory. You need to predict the outcome of the changes you make to know if the problems that occur are real problems.

For example, changing the definition of productivity for a team from functional excellence to validated learning will cause problems. As we saw earlier, functional specialists are used to measuring efficiency by considering the proportion of time they are busy doing their work. This is why most traditional work environments frustrate such experts with endless meetings and handoffs that affect their efficiency. However, the individual efficiency of these specialists isn’t the goal of a Lean Startup. Instead of forcing them to work cross-functionally with other teams to achieve validated learning. It doesn’t matter how fast we can build and measure. What’s important is how quickly we can get through the whole feedback loop.


Startups need three structural features, regardless of being an internal or external startup:

1. Scarce but secure resources.
2. Independent authority to build their business.
3. Personal risk in the result.