Chapter 9: The Importance of Small Batches


Most of the decisions that startups make are uncertain regarding how often they should release a product and whether there is a reason to release weekly instead of daily, quarterly or annually. Product releases come with a lot of overheads; thus, from an efficiency viewpoint, product releases always leave less time to devote to product building. However, waiting too long to release a product can make something no one wants.

How much time and energy must companies invest in infrastructure and planning early to predict success? If you spend too much, you will waste valuable time that could have been devoted to learning. On the other hand, if you spend too little, you might fail to take advantage of early success and give up your market leadership to a fast follower.

What should employees do in their jobs? How do we hold employees responsible for learning at an organizational level? Departments create incentive structures that make people focus on their excellence based on their specialities like marketing, sales and product development etc. However, what happens if the company’s best interests are dealt with by cross-functional collaboration? Startups need organizational structures that fight against extreme uncertainty that happens to be a startup’s main enemy.

The important first question for any Lean Startup is which activities create value and which activities are a waste? Once you understand this difference, you can start using the Lean Startup techniques to take away waste and increase the efficiency of value-creating activities. To use these techniques in a startup, they should be used in the unique situations of entrepreneurship. In chapter 3, we saw that value in a startup is not through creating products but validated learning on how to build a sustainable business. What products do customers want? How will the business grow? Who is the customer? Which customers should be heard, and which should be ignored? These questions need to be answered immediately to increase a startup’s chances of being successful.

In part 3 of the Lean Startup series, we will explore techniques that allow Lean Startups to grow without sacrificing the speed and cleverness integral to every startup.

Chapter 9 – The Importance of Small Batches

Lean manufacturing uses the one envelope at a time approach called a “single piece flow” in lean manufacturing. Although it may seem inefficient to focus on stuffing letters one by one in envelopes, it has been confirmed in many studies that it is a faster approach to getting the job done. It works due to the power of small batches. When we do work that moves in stages, the batch size refers to how much work shifts from one stage to another. For example, if you were stuffing 100 envelopes, the instinctive way to do it would be to fold 100 letters at a time, and the batch size would be 100. Single piece flow is named as it has a batch size of one.

Why does stuffing one envelope at a time get your job done faster, although it appears slower? This is because our intuition doesn’t consider the extra time needed to sort, stack and rotate the large piles of partially completed envelopes when done the other way. It seems more efficient to repeat the same thing partly because we expect to get better at this simple job the more we do it. In process-oriented tasks like this, individual performance isn’t almost as important as the entire system’s performance.

Even if the amount of time each process takes is the same, the small batch production approach would still be better for even more illogical reasons. For example, let’s assume that the letters didn’t fit in the envelopes. We won’t discover this with the large batch approach until the end. With small batches, we’d immediately know. What if the envelopes are faulty and won’t seal? In the large batch approach, we’d have to unstuff all the envelopes, get new ones and restuff them. However, we’d immediately discover this in the small batch approach, and no rework needs to be done.

These issues need to be understood by any company regardless of the size, as there are real and more significant consequences. The small batch approach produces a finished product every few seconds, while the large batch approach should simultaneously deliver all the products. What if a customer decides that they don’t want the product? Which approach do you think would allow a company to find this out sooner? The most significant benefit of working with small batches is that quality issues can be identified faster. This is the birth of Toyota’s famous andon cord, which allows any worker to get help immediately if they notice any problem like a defect in a physical part and thereby stop the entire production line if it cannot be immediately rectified. However, this is another unreasonable practice. The andon cord can disturb the careful flow of the cars being produced as the line is stopped repeatedly. Nevertheless, the advantages of finding and fixing problems quickly outweigh this disadvantage. This process of repeatedly chasing out defects has been highly successful for Toyota and its customers. It’s the main reason for Toyota’s finest ratings and low cost.

9.0 Small batches in entrepreneurship

Just like Toyota discovered that small batches made their factories more efficient, the goal is not to produce more stuff efficiently in the Lean Startup but to learn how to build a sustainable business quickly.

When considering the envelope stuffing example, what if the customer doesn’t want the product you’re building? Although this is disappointing news to an entrepreneur, finding out sooner is better than later. Working in small batches, therefore, ensures that a startup can reduce the cost of time, money and effort that finally turns out to be wasted.

9.1 Repeated deployment beyond software

There are 3 ways in which most industries are witnessing that their design process is accelerated by the same fundamental forces that make rapid iteration possible in the software industry:

1. Hardware becoming software

Almost all consumer electronics’ values are decided by their software. Even ancient products like automobiles are witnessing ever larger parts of their value generated by the software they carry inside, which controls everything from the entertainment system to engine tuning to controlling the brakes. What can be built out of software can be changed faster than a physical or mechanical device can.

2. Fast production changes

Due to the success of the lean manufacturing movement, most assembly lines are built to ensure each new product comes off the line to be completely customized without sacrificing quality or cost-effectiveness. Traditionally, this has been used to provide the customer with many product choices, but in the future, this capability will enable product designers to get faster feedback about new versions. When the design changes, there is no more inventory of the old version to slow things down. As machines are made for quick changes, new versions can be quickly produced as soon as the new design is available.

3. 3D printing and quick prototyping tools

Most products and parts that are made of plastic are mass produced through a technique called injection molding. This process is very costly and time-consuming to set up, but once it’s functioning, it can reproduce millions of similar individual items at a very low cost. Unfortunately, it’s a typical large batch production process. It has disadvantaged entrepreneurs who want to develop a new product as only large companies can afford large production runs for a new product. However, new technologies like 3D printing and quick prototyping tools allow entrepreneurs to build small batches of products of similar quality to those made with injection molding but at a lower cost and faster. The important lesson to understand here is that not everyone should be shipping products 50 times a day but that by reducing batch size, you can get through the Build-Measure-Learn feedback loop quicker than your competitors. The capability to learn faster from customers is an important competitive advantage that startups should have.

9.2 Small batches in action

To see small batches in action, let’s consider the example of SGW Designworks. SGW’s speciality is quick production techniques for physical products, and startups are mainly its clients. SGW Designworks was engaged by a client who a military client asked to build a complicated field x-ray system to discover explosives and other destructive devices at border crossings and war zones.

The system ideally included an advanced head unit that read x-ray film, multiple x-ray film panels and the structure to hold the panels while the film was being exposed. The client already had the technology for the x-ray panels and the head unit, but to make the product work in rough military situations, SGW needed to design and produce the supporting structure to make the technology usable in the field. In addition, the framework had to be stable to ensure a quality x-ray image which was durable to be used in a war zone, easy to deploy with less training and small to collapse into a backpack. This is the type of product we are used to thinking that it takes months or years to build, but new techniques are reducing that period. SGW quickly began to produce the visual prototypes by using 3D computer-aided design (CAD) software. The 3D models were a quick communication tool between the client and the SGW team to make early design decisions.

The team and client decided on a design that used an advanced locking hinge to provide the collapsibility needed without compromising stability. The design integrated a pump mechanism to enable fast, continuous attachment to the x-ray panels, which is complicated. 3 days later, the SGW team provided the first physical prototypes to the client. The prototypes were refined from aluminum directly from the 3D model by using a technique called computer numerical control (CNC) and were hand-assembled by the SGW team. The client quickly took the prototypes to its military contact for review. Although there were many minor design changes, the concept was yet accepted. In the next 5 days, the client and SGW team finished another design iteration cycle, prototyping and design review. The first production run of 40 completed units was ready for delivery 3 ½ weeks after the development project was started.

SGW found this was a successful model as feedback on design decisions was almost immediate. SGW used the same procedure to design and produce 8 products, serving a broad range of functions in a year. Half of those products are currently generating revenue due to the power of working in small batches.

9.3 The large batch death spiral

Small batches impose a danger to managers of traditional notions of productivity and progress as they believe that functional specialization is more efficient for expert workers. For example, assume that you’re a product designer responsible for a new product, and you need to produce 30 separate individual design drawings. It seems like the most efficient way to work is peacefully on your own by producing the designs one by one. Then, once you’re done with all of them, you pass the drawings to the engineering team and let them work. In other words, this means you’re working in large batches.

From an individual efficiency perspective, working in large batches is reasonable, and it also has other benefits such as promoting skill building, making it convenient to hold individual contributors responsible and allowing experts to work without disturbance. Unfortunately, although this is the theory, it doesn’t work out this way in reality. Let’s take an example where after passing 40 design drawings to engineering, the designer has the opportunity to turn his attention to the next project. But engineering has many doubts about how the drawings are supposed to work, or they find the drawings unclear. These problems undoubtedly become interruptions for the designer and are now interfering with the next large batch that the designer is supposed to be working on. If the product designer has to redo the product design, the engineers might become idle while they wait for the rework to be done. If the designer isn’t available, the engineers might have to do the designs themselves.

Large batches eventually grow as moving the batch leads to more work, rework, delays and interruptions. Thus everyone is responsible for doing work in large batches by trying to reduce this cost. This is known as a large batch death spiral as there are no physical limits on the maximum size of a batch. Therefore, a batch size can keep growing. Eventually, one batch becomes the highest priority project as the company has taken long since the last product release. However, now the managers are motivated to increase batch size instead of shipping the product.

9.4 Pull, don’t push

Lean production solves the problem of stockouts with the pull technique. For example, if you bring a car into the repair shop to repair the dent of your new Toyota Camry, one blue 2011 Camry bumper gets used. This creates a hole in the repair shop’s inventory which immediately causes a signal to be sent to a local restocking facility known as the Toyota Parts Distribution Center (PDC). Next, the PDC sends the repair shop a new bumper, creating another inventory hole. This sends the same signal to a regional warehouse, Toyota Parts Redistribution Center (PRC), where all parts suppliers ship their products. That warehouse communicates with the factory where the bumpers are made to produce one more bumper, which is manufactured and shipped to the PRC.

The goal is to achieve small batches down to single-piece flow throughout the supply chain. Each step in the line pulls the parts it needs from the earlier step. This is the just-in-time production method of Toyota. When companies shift to this production method, their warehouses quickly decrease as the amount of JIT inventory is reduced. This decrease in JIT inventory is when a lean manufacturing gets its name.

In manufacturing, pull is used mainly to ensure production processes are tuned to levels of customer demand. Without this, factories can finish making more or less of a product than customers want. However, using this approach to develop new products isn’t clear-cut. Some people mistake the Lean Startup model as simply applying pull to customer requirements. This presumes that customers could tell us what products to make and that this would act as the pull signal to product development to make them.

As we saw earlier, this isn’t how the Lean Startup model works, as customers don’t always know what they want. The goal in building products is to run experiments to help you learn how to build a sustainable business. Therefore, the correct way to view the product development process in a Lean Startup is that it responds to pull requests in the form of experiments that need to be conducted. As soon as you develop a hypothesis that you want to test, the product development team should be engineered to design and run the experiment fast by using the smallest batch size that will complete the job. Although the feedback loop goes as the order in Build-Measure-Learn, the planning goes the other way by figuring out what you need to learn and then working backwards to see what product will work as an experiment to acquire that learning. Therefore, it’s not the customer but the customer hypothesis that pulls work from product development and other functions. Apart from this, additional work is a waste.


Toyota’s production system is likely the most advanced management system globally. It has shown an ability to reveal its employees’ creativity, achieve continuous growth, and produce new products for nearly a century. This is precisely the long-term success that entrepreneurs should dream of achieving. Although the lean production techniques are strong, they are simply a demonstration of a high-functioning organization that wants to achieve maximum performance by engaging the correct measures of progress over the long run. The Lean Startup model only works if you can build an organization as adaptable and quick as the challenges it faces. This demands solving the human challenges fundamental in this new way of working, which will be the topic of the subsequent chapters.