Build, measure, learn
Ideas are the starting point of innovation, but they must be transformed into reality to be valuable. Often times what seems like a great idea in one’s mind turns out to be impossible, impractical, or illogical after attempting to reduce it to practice. The build-measure-learn loop is simply the process of transforming an idea into a valuable product or service.
First, build a model to represent the idea, i.e. a cardboard model, a computer simulation, etc. Next, get feedback from potential users–to collect data about the problem, not to sell a solution. Finally, transform this information into an improved concept and build another model or a working prototype. Repeat the cycle until the solution aligns with the problem.
It’s a long journey from concept to creation, but one that’s necessary to build something customers are willing to pay for.
Identify a problem
There are countless examples of products and services that need improvement, or problems that could be solved by a new product or service. It’s easy to find these opportunities by observing human behavior, i.e. what people do and why they do it. Be curious and ask people lots of questions, especially ‘why.’
Self-observation can be a great source of invention ideas, but it can also be a source of delusion. Inventors frequently fall in love with their ideas and refuse to listen to feedback from others. The best problems to solve are those shared among many people, and similarly, the best solutions are those accepted by many people.
Study current solutions
Unless the identified problem addresses an unarticulated customer need–meaning people aren’t even aware of the problem or haven’t considered possibilities–there’s likely a number of current solutions on the market. Before reinventing the wheel, study alternative and competitive solutions. This will provide insights about the problem and ultimately save time and money.
|(Create a quality and compact skateboard multitool.)||
Sketch a solution
After the problem has been identified, it’s time to unleash the creative energy. Start sketching different concepts of how to solve the problem. Don’t worry about artistic talent–the purpose of sketching is to visually think through multiple solution scenarios. Throughout the inventive process it is helpful to get feedback, so don’t be afraid to collaborate with others.
Build a model and get feedback
A model is something tangible and nonfunctional that represents an idea, concept, or solution. For physical products, models are usually made out of foam, cardboard, wood, and other materials found around the house or at the hardware store; for services or software, models are usually computer simulations or storyboard mock-ups. A model isn’t a tool to sell a future product, rather it’s a tool to collect feedback about the problem being solved so that the future product will be accepted be the market.
|(Cut and shape more copper and PVC pipes.)||
Improve the model and get more feedback
Keep collecting feedback and building new models until the problem is well understood and the solution being modeled addresses that problem.
|(Create dimensional drawings with CAD.)||
Design a prototype
For many physical products, engineering drawings and circuit schematics are required to create a prototype. Drawings can be done by hand, but more commonly are done with Computer-Aided Design (CAD) programs. There are many CAD programs available for 3D modeling or circuit schematic capture; some are free and some cost thousands of dollars.
For many software products or service, functional code is required to create a prototype. Determine a suitable programming language, for example Java to program mobile devices or C to program embedded systems, and download or purchase an Integrated Development Environment (IDE) and compilers.
|(Contract a machinist or manufacturer to build a few.)||
Build the prototype and get feedback
The purpose of a mode is to address the solution’s desirability (do customer’s desire the solution), whereas the purpose of a prototype is address the solution’s feasibility (can the solution be built). A prototype is used by a customer so that functional requirements can be determined, for example dimensions, material properties, etc. Encourage users to abuse the prototype, because it’s drastically better to discover design flaws now rather than during mass production.
|(Contract a machinist or manufacturer to build some more.)||
Build another model and get more feedback
It is common to discover fundamental design issues from the first few prototypes, or even that the problem wasn’t completely understood. If this happens, take a step back and rethink the solution concept.
|(Create more dimensional drawings with CAD.)||
Design another prototype
Building multiple prototypes can be expensive and time-consuming. Some manufacturers are willing to help develop prototypes for free with the expectation of receiving mass production business in the future. However, unless a manufacturer has a prototyping business, use this method only when the design is nearly finalized.
|(Contract a machinist or manufacturer to build some more.)||
Build another prototype and get feedback
Keep collecting feedback and building new prototypes until the design meets customers’ expectations, i.e. it solves the original problem without creating new ones.
|(Contract a manufacturer to make a few hundred.)||
Produce a small quantity and get even more feedback
It is important to collect feedback and take care of design issues during this process, because each step from idea to production becomes more costly. Before spending tens of thousands of dollars on production tooling, build a few hundred pre-production units to test. Don’t be lured by the low per-unit cost of mass production compared to the high per-unit cost of pre-production; money spent now to find any lingering bugs is well worth it.
|(Contract a patent agent to file an application.)||
File a patent application
At this point most of the features and attributes of the invention have been determined in sufficient detail to file a nonprovisional patent application. Nonprovisionals can be filed any time, but early applications are more likely to require follow-up applications to protect features and attributes discovered from models and prototypes. A nonprovisional must be filed within 12 months of filing a provisional patent application.
|(Contract a manufacturer to produce thousands.)||
Finalize and mass produce (and keep getting feedback)
Mass production is expensive, especially for products that require custom tooling like plastic molds or metal stamps. A final product rarely resemble its initial concept, so it’s important to go through several build-measure-learn iterations before spending lots of money on tooling, materials, or packaging. For most mass production, it’s necessary to work with a manufacturer. to protect intellectual property if a patent application hasn’t been filed yet, execute a Non-Disclosure Agreement (NDA) with the manufacturer before discussing too many technical details or providing prototypes.