Lean Start-up Principles

Covers basic understanding. The basic starting principle is: Invention $\neq$ Innovation. It is not yet commercialized. An Idea $\neq$ Business opportunity. There are many ideas, but that doesn't translate to a business opportunity. A problem $\neq$ market.

What we are actually interested in finding is a Product or Service that creates value for customers. They see the value and purchase it as it solves a need or a problem. In short, There are no facts inside your building, so get the hell outside. ~ Steve Blank

Do less thinking, sitting in a chair and reading but do more of getting out and talking to people, customers, understand what is going on. Actually, with open ended questions, infer what a potential pain point could be and what a potential problem could look like. Don't push your ideas onto them!

You need a Fundamental mindset shift from:

The Engineer Mindset: Can this be built?
- Is the tech feasible?
- Product Development Process
  - Sophisticated Product
  - High up-front Investment
  - Developed in Lab
- Periodic product launches
Hypothesis & Experiment Mindset: Should this be built?
- Is it needed? Will people use it? Are people Interested?
- Customer Development Process
  - Scrappy product (MVP) - have people sign up
  - Low investment (e.g., for ads)
- Measurement (incomplete purchase)
- Updates and iterative feedback cycles

What's a Startup All About?

In the Lean Startup Methodology, startups are learning institutions. You turn into a guru in your field by really diving deep into customer interactions. It's all about continuously experimenting, learning from real-world feedback, and adapting your approach. Progress $=$ validated learning - gathering data directly from your experiences, not just market reports. This is the YODA Principle, create your own data. Marketing reports are okay for foundation, but try to rely most on your own data, that you know where it is coming from, understand etc. But always, this is an iterative process.

Lean Methodology

Big Picture: Lean Approach

Consider the Lean Canvas as your business's blueprint. It’s a living document that helps you figure out what you’re doing right and where you might be going off track. Here’s the lean approach in a nutshell:

Lean Canvas: Sketch out the core components of your business and formalize your ideas.
Customer Interaction: Hit the pavement and talk to your potential customers. Figure out what makes them tick.
Pretotype/MVP: Make a basic version of your product to see if people are willing to pay for it.
Prototype: Refine your product based on real user feedback.

For each of these steps, create hypotheses that are clear and testable. Gather data, check for biases, and really understand your customer. Then, validation occurs as you do testing, data collection, ruling out biases.

A lot of people skip these steps. Don’t build until you have done the first couple steps. What are your assumptions? How do you formalize them into a falsifiable hypothesis that you can measure and test? What type of method do you use to collect data? How do you make sure this data.

Customer Discovery and validation is important. Think, who are the people interested in our solution. What are their problems, pain points etc. Then go out there and actually validate as much as what we think their problems and pain points are.

After this you do your customer creation and build out your company.

Steps of Lean Methodology

Actionable Steps

So how do we actually do this?

Identify pain points: Use customer interviews to discover what they need, not what you think they need.
- Vitamin or Antibiotic. Is it solving a pain? Is it core to their need?
Hypothesize: Translate into falsifiable hypotheses that you can test.
Prioritize tests: Figure out which hypotheses to test first.
Test and learn: Put your ideas into action and see what happens.
Validate or pivot: Decide if you're on the right track or if you need to change course.
- Preserve solution or falsify (i.e. pivot)
Generate new ideas: Use what you learn to come up with better ideas.
Repeat: Keep learning and keep iterating.

What methods can you use for validated learning?

Interviews ➡️ teach you verbalizable opinions; Mom test
- Try not to get the opinions I was fishing for
- Observations ➡️ teach you the un-articulatable, the hidden; IDEO; Thick Data
- Sit and observe
- Not everything valuable is measurable / quantifiable
- Careful with how you use data. Thick vs Big Data
- When human behaviour enters the equation, data is less reliable.
Focus Groups ➡️ diverse views
Surveys ➡️ larger $n$ for broader validation
Experiments ➡️ Prompt [Non-]Action of earlyvangelists

Examples of this are:

H: People will buy shoes online.
MVP: Make simple homepage & hand deliver without logistics i.e. scrappy and simple
Measure: No. of successful orders

Or alternatively, as Dropbox (opens in a new tab) did:

H: Simple storage and file sharing needed
MVP: Make a Short-Video (demo with jokes)
Measure: Waiting list sign-ups ( $5k \rightarrow 75k$ )

All we do at the start is to see, "Is there interest?", "Who will be interested?". You want to get people as close to a purchasing decision as possible. That is fuller validation.

Go in with the mentality that you are most probably wrong. See how wrong you are. Inform and make substantial changes to your lean canvas, and pivot.

Slip-Ups

Distinguish between customers and users. If you have multiple user roles in your product, identify your customers. A customer is someone who pays for your product. A user does not.

Make sure you split broad customer segments into smaller ones. Start-ups that feel the problems they are solving are so universal, they apply to everyone, have a massive problem starting up. You can't effectively build, design, and position a product for everyone. While you might be aiming to build a mainstream product, you need to start with a specific customer in mind. Even Facebook, with its now 500 million+ users, started with a very specific user in mind: Harvard University students.

Sketch a Lean Canvas for each customer segment. The elements of your business model can and will vary greatly by customer segment. Start with the top two or three customer segments you feel you understand the best or find most promising.

Quantifying is addictive. It is easy to neglect data that can't be measured i.e. Qualitative data. Thick vs Big Data. You need this customer discovery and validation and delivers real depth and meaning. Unlike big data. Grounds business questions in human questions. Helps rescue the context loss that comes from making big data useable. Then you can scale it. Quantification bias.

Lean canvas vs. Business model

The Lean Canvas is format for brainstorming possible business models, prioritizing where to start, and tracking ongoing learning. At this stage you would only have an inkling of a problem, a solution, and maybe a customer segment.

Rushing to build a solution can cause a lot of harm; so, can prematurely picking a customer segment or business model. Here, this "selection bias" is untested and may result in a suboptimal business model or local maxima. As with most optimization techniques, the caveat is that these improving/iterative searches only use local information about the objective function. Therefore, they cannot distinguish between local and global optima. Therefore, it is very very important to initialize it well. It is the same with the iterative process of business development, and a lean canvas helps to initialize.

Lean canvases are less rigid than a business plan. Business plans may take several weeks or months to write a 60-page business plan (largely built on untested hypotheses) and could have, after all that, been initialized wrong. Lean Canvases, on the other hand, should take anywhere between 15 mins and one afternoon. But both are, fundamentally, building on the same material i.e. Problem, UVP etc.

Your business model is in fact your "product". Your product is not actually the product. Lean Canvas' help deconstruct your business model into nine distinct subparts that you can then systematically test, in order of highest to lowest risk.

A potential downside to the lean canvas approach of experimentation, learning and appropriability at an early stage is that by releasing early versions of products, start-ups risk imitation by a competitor, or even replicas.

Before moving on, Complete a Lean Canvas. ~ Ash Mayura

Customer Interview Guidelines

Pre-planning:

Make a list of 25-50 (100 for B2B) interviewees and several places around customer segment.
Develop an interview script
Pre-test interview script with friends
Make adaptations

Guidelines include:

B2B or B2C type or business. B2B must be more technical, less of an informal conversation. Here, often times the people you speak to are not the ones that experience the problem, so you need to talk to multiple people. So, who do you talk to? Middle management, COO, CEO etc.? The lower down the ladder you try, the more likely you are to get an interview. Can use a "Co-creation lens" where you work on a solution together and work in specific problems and present it to the CEO together. B2C for smaller about, the person you are talking to is the one making the decision. Also, some Venture Capitalists also only fund B2B as it is much easier to validate.
What mindset do you have? Are you selling or are you actively listening? The mom test. You want to find their unbiased way of thinking, you don't want to lead them to your solution at all. Don't sell and talk about the problem but listen to their, the customers', problem. Develop line of questions close to customer experience of the world and have the customers tell you stories. Use indirect questions to learn about problem context, stay quiet and avoid talking about product features and what you think the problem is. Remember, how wrong are we? We want the counter intuitive insights.
Make it two per interview (where it makes sense). One person interviews, the other takes notes. Don't want the customer to feel overwhelmed.
Get excited when you hear things that you didn't want to hear. "This is how we were wrong....", "This is what I learnt...." will develop your learning. This will help you a lot more. The further waya you go from your own world, the better.
Ask open ended questions that focus on actual behaviour (past/present). "What is the hardest part about....", "What do you do when...", "Can you tell me about...", "How was the experience with...", "Why did you buy device?", "How do you currently solve the problem...". Don't ask 'yes' or 'no' questions (do you do X) and don't ask what-if hypothetical questions about potential behaviour (no would, no could). Really try to understand how it can be solved, and the mentality around it.
Don't focus on technical aspects (like the engineer you are). Listen for market, industry related markers! "I wish this existed" is the best thing to hear. "I don't know why they can't offer X....", "I wish this existed, too bad no one is doing that....".
Play stupid and keep asking for clarifications (why and how questions). "Can you help me understand and explain why...", "How exactly does this work...", "Can you talk me through an actual example when that happened to you", "How often do you experience that situation/problem?". Get them to explain tings in a stupid way.
Ask for 2-3 referrals and introductions. "Can you introduce me to somebody else who faces this challenge?", "Do you know someone else who is in a similar position as you?".
You are done with interviewing when you know what the interviewee will answer; After 5-10 interviews repetition should kick in. This holds true when you are speaking to similar customers. If you variation, then it is worth inquiring why.
Record and make notes. Compare to tease out insights/patterns immediately. Transcription may not be necessary unless thick transcript. Regularly, at times daily, team calls should happen to compare insights may be necessary. Slight adaptations to interview script might be needed to further verify insight.

Additional Resources:

Uncertainty Motivation