Ever tried to race a child at a simple task like stuffing envelopes? Eric Ries did, and he lost because he used the "efficient" large-batch method while his daughter used single-piece flow. This concept is the practice of moving a single unit of work through an entire process before starting the next one. Most people think doing work in big groups is faster, but it's actually the primary cause of waste in business.
Working in small batches reduces the time it takes to get feedback. It allows you to find mistakes immediately rather than at the end of a long cycle. By focusing on finishing one complete item, you eliminate the hidden costs of sorting, stacking, and moving unfinished work.
Single-piece flow is a core pillar of the toyota production system that Eric Ries adapts for innovation in his book The Lean Startup. In a manufacturing context, it means a car moves through the assembly line one station at a time without piles of parts sitting between workers. In a business or startup context, it means finishing a single feature, document, or sale before moving to the next.
This method is counterintuitive because it feels like you're doing more work by switching tasks. However, the goal of a startup isn't just to stay busy. The goal is to learn how to build a sustainable business as quickly as possible. Small batches are the most effective way to reach that goal.
In the envelope-stuffing example, the large-batch approach involves folding every letter first, then stuffing them all, then stamping them all. It seems fast because you're in a rhythm. But you're actually spending a massive amount of time managing piles of paper. Batch size optimization shows that when you do one at a time, you produce a finished product every few seconds.
This matters because you can find defects early. If the letter doesn't fit the envelope, the single-piece flow worker finds out on the first try. The large-batch worker finds out after folding a hundred letters. One person has zero rework; the other has a disaster on their hands.
Most organizations suffer from a "large-batch death spiral." The longer a team waits to ship a product, the more they fear it might fail. To mitigate this fear, they add more features to make the release "perfect." This increases the batch size and pushes the launch date even further back. A lean workflow breaks this by forcing the team to release small changes frequently.
Data from the book shows that Toyota became the world’s largest automaker in 2008 by using these small batches. They didn't have the capital for the massive machines used by American factories. Instead, they built smaller, general-purpose machines that could be reconfigured in minutes. This allowed them to produce a wider variety of cars faster than their competitors.
Functional specialization is a productivity trap. We think a designer should only design and an engineer should only code in big, separate chunks. This creates hand-offs that are filled with miscommunication and delays. When you move to a one-feature-at-a-time model, these experts must collaborate in real-time.
This shift moves the focus from individual efficiency to system efficiency. It doesn't matter how many lines of code an engineer writes if the feature doesn't solve a customer problem. By finishing one item at a time, the entire team stays aligned on the only thing that matters: delivering value to the customer.
In The Lean Startup, the author describes a race between his daughters and himself. He folded all the newsletters first, then attached the seals, then the stamps. His daughters completed each envelope one at a time. The daughters won easily. The father spent too much time sorting and moving the growing piles of unfinished paper, which added no value to the final product.
This design firm was tasked with building a complex x-ray system for the military. Instead of spending months on a master plan, they used 3D CAD software to build virtual prototypes in one day. They then used CNC machines to create physical aluminum prototypes in three days. By working in tiny batches, they completed three full design iterations in just nine days. They delivered forty production units less than a month after starting.
Pick a single feature or task that you can finish by the end of the day. Don't worry about the "full version" or the long-term roadmap. Your only goal is to get one complete piece of work into a state where it can be tested by a real user.
If you usually send a document to a different department for approval, bring that person into your room. Solve the problems together in real-time. This removes the "waiting time" that usually makes up 90% of a project's duration.
Don't wait for a weekly or monthly release window. Once that single feature is done and tested, put it live. This provides a clean break and allows you to start the next task with fresh data on how the first one performed.
Critics often argue that this approach is inefficient for specialists who need to be "in the zone." They feel that switching between different types of work for one feature is distracting. This is a common complaint in software engineering where deep focus is highly valued. However, the cost of that focus is often building the wrong thing for weeks at a time.
There are also regulatory and safety concerns in industries like healthcare or aerospace. You can't always "ship" a new heart monitor feature every hour. In these cases, the goal is still to reduce the batch size as much as possible within the constraints. Even if you can't release to a customer, you can still finish the work and test it internally in a single-piece flow to catch errors early.
Focusing on individual speed is a vanity metric that hides system-wide waste. Productivity is measured by how quickly you can move a complete idea through the entire loop. Push one small, finished change to a live user before five o'clock today.
While it feels slower because you're switching tasks more often, it removes the 'hidden' time spent sorting, moving, and fixing large-scale errors. In the envelope example, the father lost the race because the overhead of managing piles of paper exceeded the time saved by repetitive motion. Small batches reveal defects immediately, preventing massive rework later.
In software, this means finishing one user story or feature and pushing it all the way to production before starting the next. This requires automated testing and a robust 'immune system' for your code. It prevents the 'large-batch death spiral' where teams spend months building features that no one actually wants or that don't work together.
Yes. If you run a restaurant, it's making one plate at a time rather than prepping 50 salads that might sit and wilt. If you are in sales, it's following up with one lead fully before opening 20 new ones. Any process that has multiple steps can be optimized by reducing the number of items in progress.
Smaller batch sizes lead to higher quality because the feedback loop is shorter. If you make a mistake in a batch of one, you find it immediately. If you make a mistake in a batch of 1,000, you have to throw away or fix 1,000 items. This is why Toyota uses the 'andon cord' to stop the whole line for one defect.
Single-piece Flow Finish One Thing Before Starting the Next
The Magic Mix Preserving Your Core While Stimulating Progress
Zero to One vs One to n Horizontal vs. Vertical Progress Explained
Relentless Improvement How to Move the Needle on Existing Products
Lean Startup vs. Intelligent Design Why Iteration Won't Get You to 1
The Alchemy of Greatness Combining Discipline with Entrepreneurship
Learning Milestones An Alternative to Traditional Business Goals
Learning from Failure How to Conduct Autopsies Without Blame
Why We Stopped Trying to Cure Death The Shift to Indefinite Life Extension
Gillette’s Shaving Systems Technology as a Flywheel Accelerator