Running Lean Summary (Review & Book Notes)

Tomas Laurinavicius
Updated on April 24, 2024

If you liked The Lean Startup, you’ll love this one. It’s the perfect accompanying book, although more action-oriented than the former. It will teach you how to really start a startup. Highly recommended, especially to the first-time entrepreneur.

Authors: Ash Maurya

Originally published: 2012

Pages: 240

Genre: Business

Goodreads rating: ⭐️ 4.06/5

👉 Buy Running Lean on Amazon

🎧 Listen for free on Everand (plus 1+ million other books)

Three Stages of a Startup

Stage 1: Problem/Solution Fit

Key question: Do I have a problem worth solving? The first stage is about determining whether you have a problem worth solving before investing months or years of effort into building a solution. While ideas are cheap, acting on them is quite expensive. A problem worth solving boils down to three questions:

  • Is it something customers want? (must-have)
  • Will they pay for it? If not, who will? (viable)
  • Can it be solved? (feasible)

Stage 2: Product/Market Fit

Key question: Have I built something people want? Once you have a problem worth solving and your MVP has been built, you then test how well your solution solves the problem. In other words, you measure whether you have built something people want.

Stage 3: Scale

Key question: How do I accelerate growth? After product/market fit, some level of success is almost always guaranteed. Your focus at this stage shifts toward growth, or scaling your business model.

Pivot versus Optimizing

Achieving product/market fit is the first significant milestone of a startup and greatly influences both strategy and tactics. For this reason, it is helpful to further delineate the stages of a startup as “before product/market fit” and “after product/market fit.”

Before product/market fit, the focus of the startup centers on learning and pivots. After product/market fit, the focus shifts toward growth and optimizations.

This may sound like a subtle distinction, but it has a significant impact on both strategic and tactical execution. Before product/market fit, a startup needs to be architected to maximize learning.

You stand to learn the most when the probability of the expected outcome is 50%; that is, when you don’t know what to expect.

In order to maximize learning, you have to pick bold outcomes instead of chasing incremental improvements. So, rather than changing the color of your call-to-action button, change your entire landing page. Rather than tweaking your unique value proposition (UVP) for a single customer segment, experiment with different UVPs for different customer segments.

The Iteration Meta-Pattern

While an experiment helps you validate or invalidate a specific business model hypothesis, an iteration strings multiple experiments together toward achieving a specific goal, such as getting to product/market fit.

The first two stages (Understand Problem and Define Solution) are about getting to problem/solution fit or finding a problem worth solving.

Then you iterate toward product/market fit by testing whether you’ve built something people want using a two-stage approach: first qualitative (microscale), then quantitative (macro-scale).

Brainstorming Possible Customers

When you first start out, all you have is an inkling of a problem, a solution, and maybe a customer segment. Just as rushing to build a solution can lead to waste, so can prematurely picking a customer segment or business model. The danger here is that this “selection bias” is untested and may result in a suboptimal business model or local maxima.

While there is no way to completely avoid the local maxima problem, you raise your odds for finding a better solution when you are initially open to exploring and even testing multiple models in parallel. Start by brainstorming the list of possible customers for your product:

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. Split broad customer segments into smaller ones.

I’ve worked with startups that felt the problems they are solving are so universal, they apply to everyone.

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.

Put everyone on the same canvas at first

If you are building a multisided business, you might find it necessary to outline different problems, channels, and value propositions for each side of the market. I recommend starting with a single canvas first and using a different color or tag to identify each customer segment. This helps you visualize everything on a single page. Then split if needed.

Sketch a Lean Canvas for each customer segment

As you’ll find shortly, the elements of your business model can and will vary greatly by customer segment. I recommend starting with the top two or three customer segments you feel you understand the best or find most promising.

How to Craft a Unique Value Proposition

Here are some of my tips on how to craft a UVP:

Be different, but make sure your difference matters

The key to unlocking what’s different about your product is deriving your UVP directly from the number-one problem you are solving. If that problem is indeed worth solving, you’re more than halfway there already.

Target early adopters

Too many marketers try to target the “middle” in the hopes of reaching mainstream customers, and in the process they water down their message. Your product is not ready for mainstream customers yet. Your sole job should be to find and target early adopters, which requires bold, clear, and specific messaging.

Focus on finished story benefits

You’ve probably heard about the importance of highlighting benefits over features. But benefits still require your customers to translate them to their worldview. A good UVP gets inside the head of your customers and focuses on the benefits your customers derive after using your product.

So, for instance, if you are creating a résumé-building service:

  • A feature might be “professionally designed templates.”
  • The benefit would be an “eye-catching résumé that stands out.”
  • But the finished story benefit would be “landing your dream job.”

A good formula for crafting an effective UVP (by way of Dane Maxwell) is:

Instant Clarity Headline = End Result Customer Wants + Specific Period of Time + Address the Objections

Pick your words carefully and own them

Words are key to any great marketing and branding campaign. Look at how the top luxury car brands have used a single word to define themselves:

  • Performance: BMW
  • Design: Audi
  • Prestige: Mercedes

Picking a few “key” words that you consistently use also drives your search engine optimization (SEO) ranking.

Answer: what, who, and why

A good UVP needs to clearly answer the first two questions—what is your product and who is your customer. The “why” is sometimes hard to fit in the same statement, and I’ll frequently use a subheading for that.

Study other good UVPs

The best way to craft a good UVP is to study the UVPs of the brands you admire. Visit their landing pages and deconstruct how and why their messaging works. Some of my best teachers have been Apple, 37signals, and FreshBooks.

Create a high-concept pitch

Another useful exercise is creating a high-concept pitch. High-concept pitches are used heavily by Hollywood producers to distill the general plot of a movie to a memorable sound bite. The high-concept pitch was also popularized as an effective pitching tool by Venture Hacks in its ebook, Pitching Hacks.


  • YouTube: “Flickr for video”
  • Aliens (movie): “Jaws in space”
  • Dogster: “Friendster for dogs”

The high-concept pitch should not be confused with a UVP and is not intended to be used on your landing page. There is a danger that the concepts the pitch is based on might be unfamiliar to your audience. For this reason, the high-concept pitch is more effective when used to quickly get your idea across and make it easy to spread, such as after a customer interview.


The initial goal of a startup is to learn, not to scale. So, at first it’s OK to rely on any channels that get you in front of potential customers.

The good news is that following a “customer discovery1/interview” process forces you to build a path to “enough” customers early. However, if your business model relies on acquiring large numbers of customers to work, that path may not scale beyond the initial stages, and it’s quite possible you’ll get stuck later.

For this reason, it’s equally important to think about your scalable channels from day one so that you may start building and testing them early.

While there are a plethora of channel options available, some channels may be outright inapplicable to your startup, while others may be more viable during later stages of your startup. I typically look for the following characteristics in my early channels.

Freer versus paid

First, there is no such thing as a free channel. Channels we normally associate as being free, like SEO, social media, and blogging, have a nonzero human capital cost associated with them. Calculating their ROI is complicated because, unlike a paid channel that is used up after you pay for it, these channels keep working for you over time.

Inbound versus outbound

Inbound channels use “pull messaging” to let customers find you organically, while outbound channels rely on “push messaging” to reach customers. Examples of inbound channels:

  • Blogs
  • SEO
  • Ebooks
  • White papers
  • Webinars

Examples of outbound channels:

  • SEM
  • Print/TV ads
  • Trade shows
  • Cold calling

Direct versus automated

As a scalable channel, direct sales only make sense in businesses where the aggregate lifetime value of the customers exceeds the total compensation of your direct sales people, such as in certain B2B and enterprise products. But as a learning channel, direct selling is one of the most effective, since you interact face to face with the customer. First sell manually, then automate.

Direct versus indirect

Another area where startups waste energy is prematurely trying to establish strategic partnerships. The idea is to partner with a larger company to leverage its channels and credibility. The problem is that until you have a proven product, you won’t get the right level of attention from the bigger company’s sales reps to make this work. Imagine you are a sales rep at the bigger company. Given the choice of selling what you know or selling an unproven product to make your quota, which would you choose?

You have to first sell your product yourself, before letting others do it.

Retention before referral

Many startups are obsessed with building virality and referral/affiliate programs into their product from day one. While referral programs can be very effective in spreading the word about your product, you need to have a product worth spreading first.

The Key Metrics

Every business has a few key numbers that can be used to measure how well it is performing. These numbers are key for both measuring progress and identifying hot spots in your customer lifecycle.


Acquisition describes the point when you turn an unaware visitor into an interested prospect.

In the case of the flower shop, getting someone walking by your window to stop and come in to your shop is an acquisition event.

On a product website, getting someone to do anything other than leave your website (abandon) is a measure of acquisition. I specifically measure successful acquisition as getting my visitors to view my signup page.


Activation describes the point when the interested customer has his first gratifying user experience.

In the case of the flower shop, if the prospect found the shop in disarray once he comes inside, there would be a disconnect with the promise made at the front of the store. That wouldn’t be a gratifying first user experience.

On the product site, once the prospect signs up, you have to make sure you get the customer to a point where he can connect the promise you made on your landing page (your UVP) with your product.


Retention measures “repeated use” and/or engagement with your product.

So, in the case of the flower shop, the action of coming back to the store—and in the case of the product website, the act of logging back in to use the product again—would count toward retention.


Revenue measures the events that get you paid.

These could be buying flowers or buying a subscription for your product. These events may or may not occur on the first visit.


Referral is a more advanced form of a user acquisition channel where your happy customers refer or drive potential prospects into your conversion funnel.

In the case of the flower shop, this could be as simple as telling another friend about the store.

For the software product, this could range from implicit viral or social sharing features (like Share with a friend), to explicit affiliate referral programs or Net Promoter Score.

The Unfair Advantage

A real unfair advantage is something that cannot be easily copied or bought. Here are some examples of real unfair advantages that fit this definition:

  • Insider information
  • The right “expert” endorsements
  • A dream team
  • Personal authority
  • Large network effects
  • Community
  • Existing customers
  • SEO ranking

Guidelines for Running Business Model Interviews

Another effective technique for further calibrating your risks is getting out of the building and validating them with people other than yourself.

It is imperative that you share your model with at least one other person.

I used to advocate jumping right into customer interviews after documenting my initial models, but now I prefer to first spend a little additional time prioritizing risks and brainstorming alternative models with people other than customers—e.g., advisors.

The main reason I do this is to maximize speed and learning. Customers cannot directly give you all the answers, and due to the iterative and qualitative nature of early learning, validating hypotheses takes time. Furthermore, you might still be targeting too broad a customer segment, too small a customer segment, or the wrong customer segment altogether.

The “right” advisors, on the other hand, can help you identify risks on the “total plan” and help you to further refine and/or outright eliminate some models.

I use the term advisor rather loosely. An early advisor might be a prototypical customer, a potential investor, or another entrepreneur with specific expertise, domain knowledge, or experiential knowledge that applies to you.

Here are some guidelines for running business model interviews:

Avoid the 10-slide deck

I completely avoid a traditional “10-slide deck” because the point of the interview is learning versus pitching. The other extreme, no slides, although most natural, requires practice and may not lead to as many actionable insights because it may be hard for the other person to retain everything you tell her.

My tool of choice is an incremental build of the Lean Canvas delivered in an iPad (or paper). I start with a blank canvas and incrementally reveal parts of the business model as I walk through it.

Devote 20% of your time to setup, 80% to conversation

The stacked flow allows me to pace the conversation and leave all the information on the screen. It usually takes me three to five minutes to walk through my model; then I shut up and listen.

I have found that leaving the complete canvas open in front of people always evokes a reaction because people can visualize the entire model and they always have an opinion.

Ask specific questions. I specifically want to know:

  • What do they consider to be the riskiest aspect of this plan?
  • Have they overcome similar risks? How?
  • How would they go about testing these risks?
  • Are there other people I should speak with?

Be wary of the “advisor paradox”

As we’ll see shortly, just as customer interviews aren’t about asking customers what they want, these interviews aren’t about asking advisors what to do.

The key is not to take this feedback as either “judgment” or “validation,” but rather as a means of identifying and prioritizing risk.

It is still your job to own your business model. But because you don’t have all the answers, you need to build your startup through a series of conversations—with advisors, customers, investors, and even competitors.

Success is unlocked at the intersection of these conversations, and it’s your job as the entrepreneur to synthesize it into a coherent whole.

Recruit visionary advisors

Much like early adopters want to help when you nail their problems, visionary advisors will want to help when you present them with interesting problems that trigger their strengths and passion.

You’ll know if there’s a fit based on their answers and body language. If so, consider bringing them on as formal advisors.

Running Effective Experiments

Identify a Single Key Metric or Goal

When formulating an experiment, stay focused on the key learning or key metric you need to achieve, which will vary by the type and stage of your product. While it’s possible to tackle multiple metrics and goals simultaneously, I’ve always found it most effective to stay singularly focused.

Do the Smallest Thing Possible to Learn

Challenge yourself to find the simplest thing you can do to test a hypothesis. This is an underappreciated skill. Once you truly understand what’s riskiest about your product, it’s often possible to build something other than the product to test it.

Formulate a Falsifiable Hypothesis

What most people write down for their business model is really not yet in a form that is testable. The Lean Startup methodology is heavily rooted in the scientific method and requires that you convert these assumptions into falsifiable hypotheses.

A falsifiable hypothesis is a statement that can be clearly proven wrong.

When you skip this step, you can easily fall into the trap of accumulating just enough evidence to convince yourself that your hypothesis is correct.

A formula for crafting a falsifiable hypothesis is:

Falsifiable Hypothesis = [Specific Repeatable Action] will [Expected Measurable Outcome]

Validate Qualitatively, Verify Quantitatively

Before product/market fit, the terrain is riddled with extreme uncertainty. The good news is when you have a lot of uncertainty, you don’t need a lot of data to learn.

Your initial goal is to get a strong signal (positive or negative) that typically doesn’t require a large sample size. You might be able to do this with as few as five customer interviews.

A strong negative signal indicates that your bold hypothesis most likely won’t work and lets you quickly refine or abandon it. However, a strong positive signal doesn’t necessarily mean your hypothesis will scale up to statistical significance; nevertheless, it gives you permission to move forward on the hypothesis until it can be verified later through quantitative data.

Make Sure You Can Correlate Results Back to Specific Actions

One of the harder things to do is to correlate measured results back to specific and repeatable actions, as your product is always changing. When running qualitative experiments (like interviews), it’s important to run them the same way until certain repeatable patterns emerge. For quantitative experiments, techniques like cohort analysis and split testing allow you to achieve this. We’ll cover this in more detail a bit later.

Applying the Iteration Meta-pattern to Risk

While the top three starting risks serve as a quick diagnostic for prioritizing your canvases, here is how you systematically tackle them in stages.

Stage 1: Understand the problem

Conduct formal customer interviews or use other customer observational techniques to understand whether you have a problem worth solving. Who has the problem, what is the top problem, and how is it solved today?

Stage 2: Define the solution

Armed with knowledge from Stage 1, take a stab at defining the solution, build a demo that helps the customer visualize the solution, and then test it with customers. Will the solution work? Who is the early adopter? Does the pricing model work?

Stage 3: Validate qualitatively

Build your MVP and then soft-launch it to your early adopters. Do they realize the unique value proposition (UVP)? How will you find enough early adopters to support learning? Are you getting paid?

Stage 4: Verify quantitatively

Launch your refined product to a larger audience. Have you built something people want? How will you reach customers at scale? Do you have a viable business?

Here is how you view them based on risks:

Product risk: Getting the product right

  1. First make sure you have a problem worth solving.
  2. Then define the smallest possible solution (MVP).
  3. Build and validate your MVP at small scale (demonstrate UVP).
  4. Then verify it at large scale.

Customer risk: Building a path to customers

  1. First identify who has the pain.
  2. Then narrow this down to early adopters who really want your product now.
  3. It’s OK to start with outbound channels.
  4. But gradually build/develop scalable inbound channels—the earlier the better.

Market risk: Building a viable business

  1. Identify competition through existing alternatives and pick a price for your solution.
  2. Test pricing first by measuring what customers say (verbal commitments).
  3. Then test pricing by what customers do.
  4. Optimize your cost structure to make the business model work.

Tactics for Interviewing Prospects

Build a frame around learning, not pitching

In a pitch, since you’re doing most of the talking, it’s very easy for customers to pretend to go along with what you’re saying, or to outright lie to you.

The problem with starting with a pitch is that it is predicated on having knowledge about the “right” product for the customer (Problem/Solution Fit).

Before you can pitch the “right” solution, you have to understand the “right” customer problem.

In a learning frame, the roles are reversed: you set the context, but then you let the customer do most of the talking. You don’t have to know all the answers, and every customer interaction (interview, tech support, feature request, etc.) turns into an opportunity for learning. Plus, people are generally willing to help if you set the right expectation of seeking their advice over trying to pitch to them.

Don’t ask customers what they want. Measure what they do

It’s fairly common to find customers lying in interviews—sometimes out of politeness and sometimes because they really don’t know or don’t care enough. Your job shouldn’t be to call out their lies, but rather to find ways to validate what they say with what they do, preferably during the interview.

Another tactic is to use strong calls to action. If a customer says he would pay for your product, instead of getting just a verbal commitment, ask for an advance payment or partial payment and provide him with a money-back guarantee.

Stick to a script

While exploration is a critical aspect of talking to customers, you need to bind the conversation around specific learning goals.

Otherwise, you can easily blow off a lot of time and end up with an overwhelming amount of unactionable information.

Unlike a pitch, it doesn’t help to tweak your story after every interview. You need consistency and repeatability to instill some method to the process. Scripts help you do that.

Cast a wider net initially

Even though your first objective will be to home in on the defining attributes of early adopters, not all of your prospects will (or should) be early adopters. It’s better to start with a broader sweep of initial prospects at this stage (to avoid running into a local-maxima problem), and refine from there. You will have ample opportunity to narrow down your filter in the next round of interviews.

Prefer face-to-face interviews

Earlier, I stressed the importance of being able to see your interviewees. In addition to picking up on body language cues, I find that meeting someone in person instills a sense of closeness that you can’t re-create virtually. This is critical in customer relationship building.

Start with people you know

Finding people to interview can be challenging at first. Start with people you know who fit your target customer profile. Then use them to get two or three degrees out to find other people to interview. Not only does this help you practice and get comfortable with your script, but it’s an effective way to get warm introductions to other prospects.

Pick a neutral location

I prefer to conduct the first interview in a coffee shop to create a more casual atmosphere. Doing it at a prospect’s office makes it more “business-like” and makes it feel more like a sales pitch—which it shouldn’t be. That being said, I’ll agree to meet the prospect wherever she chooses.

Ask for sufficient time

My interviews typically run between 20 and 30 minutes, without feeling rushed. Make sure you set the right time expectations up front and are respectful of the interviewees’ time.

Don’t pay prospects or provide other incentives

Unlike usability testing, where it is acceptable to provide incentives for participation, your goal here is to find customers who will pay you, not the other way around.

Avoid recording the interviewees

I tried recording interviewees early on (with their permission), but found that it made some people self-aware during the interview— another example of observer bias. That, coupled with the fact that I never really went back to listen to an interview, made it a nonstarter for me. Your mileage may vary.

Document results immediately after the interview

I recommend spending five minutes immediately following an interview to document the results while your thoughts are fresh. Debrief with others later.

Prepare yourself to interview 30 to 60 people

As a rule of thumb, prepare to interview 30 to 60 people over a fourto six-week period, which means talking to two or three customers a day, with some time built in for iteration.

The actual numbers could vary based on the strength of the signal you receive, your specific path to customers, and your business model. You know when you are done: when you stop learning anything new from the interviews. In other words, when you can accurately predict what the customer is going to say just by asking a few qualifying questions, you are done.

Finding Prospects

Whenever possible, you want to prioritize finding prospects through a channel you will actually use to acquire future customers. Unless you already have a path to customers, this may not be possible at this stage. Here is a list of other techniques you can use to find and recruit interviewees:

Start with your first-degree contacts

The first place to start is with your immediate contacts that meet your target customer demographic. Some are wary that feedback received from close contacts may be biased. My view is that talking to anyone is better than talking to no one.

Ask for introductions

The next step is to ask your first-degree contacts for introductions to people who meet your customer demographic. It’s a good idea to include a message template that your contacts can simply cut and paste and forward to save them time.

Play the local card

People are generally willing to meet if they can identify with you. The email in the preceding list item emphasizes “Austin” in the body and was quite effective in setting up meetings with local photographers.

Create an email list from the teaser page

If the Web is a viable channel for your product, setting up a teaser page early is a great way to find people to interview. See the Appendix for detailed steps on crafting a teaser page.

While you may not know whether people here meet your target customer demographic, they do represent people who were motivated enough to act on your unique value proposition (UVP). Reach out to them and ask if they’d be willing to spend 20 to 30 minutes with you on a call.

Give something back

Turn the interview into a “real interview” and offer a write-up, blog post, or video in exchange.

Use techniques such as cold calling, emailing, and LinkedIn

The secret to getting a prospect (cold or warm) to agree to an interview is to “nail their problem.” You may not be able to do that out of the gate, which is why I typically rely on the other techniques in this list to run a few interviews first.

What you Need to Learn on the Problem Interview

The Problem interview is all about validating your hypotheses around the “problem-customer segment” pair. In the Problem interview, you are specifically looking to tackle the following risks:

Product risk: What are you solving? (Problem)

How do customers rank the top three problems?

Market risk: Who is the competition? (Existing Alternatives)

How do customers solve these problems today?

Customer risk: Who has the pain? (Customer Segments)

Is this a viable customer segment?

What you Need to Learn on the Solution Interview

Armed with a prioritized problem list and an understanding of existing alternatives, you are now ready to formulate and test a solution. You will start by double-checking your learning from the Problem interview, then look to test the following additional risks:

Customer risk: Who has the pain? (Early Adopters)

How do you identify early adopters?

Product risk: How will you solve these problems? (Solution)

What is the minimum feature set needed to launch?

Market risk: What is the pricing model? (Revenue Streams)

Will customers pay for a solution? What price will they bear?

Testing the Price

Don’t Ask Customers What They’ll Pay, Tell Them

You can’t (and shouldn’t) convince a customer that she has a must-have problem, but you often can (and should) convince a customer to pay a “fair” price for your product that is usually higher than what both you and the customer think it is.

The mindset most of us have during Solution interviews is one of “lowering signup friction.” We want to make it as easy as possible for customers to say yes and agree to take a chance on our product, hoping the value we deliver over time will earn us the privilege of their business.

Not only does this approach delay validation because it’s too easy to say yes, but a lack of strong customer “commitment” can also be detrimental to optimal learning.

Your job is to find early adopters who are at least as passionate about the problems you’re addressing as you are, and if you’re charging, who are willing to pay your fair price. As we covered earlier, your pricing not only is part of your product, but it also defines the customer segment you attract.

The solution interview as AIDA

  • Attention: Get the customer’s attention with your UVP—derived from the numberone problem you uncovered during earlier Problem interviews. The most effective way to get noticed is to nail a customer problem.
  • Interest: Use the demo to show how you will deliver your UVP and generate interest.
  • Desire: Then take it up a notch. When you lower signup friction, you make it too easy for the customer to say yes, but you are not necessarily setting yourself up to learn effectively. You need to instead secure strong customer commitments by triggering on desire. The earlier pricing conversation generated desire through scarcity and prizing.
  • Action: Get a verbal, written, or prepayment commitment that is appropriate for your product.

The Solution Interview Exit Criteria

You are done when you are confident that you:

  • Can identify the demographics of an early adopter
  • Have a must-have problem
  • Can define the minimum features needed to solve this problem
  • Have a price the customer is willing to pay
  • Can build a business around it (using a back-of-the-envelope calculation)

Defining the Activation Flow

Once you have distilled your features list, you are ready to start defining your activation flow. Your activation flow describes the path customers take from signing up for your service to having a gratifying first experience.

While the ultimate objective of your activation flow is to get your customers to experience your UVP as quickly as possible, most of what goes wrong right after you launch happens here. For this reason, it is far more critical to architect your activation flow for learning over optimization. Here are some ways to do that:

Reduce signup friction, but not at the expense of learning

It is generally a good practice to keep your signup forms short and only collect what you absolutely need, but don’t shy away from asking for critical contact information (like an email address) up front.

Reduce the number of steps, but not at the expense of learning

The same principle of architecting for learning over optimization also applies to the number of steps in your activation flow. While it is important to reduce the number of steps, it is far more important to keep critical steps separate so that you can troubleshoot where people drop off when things go wrong.

Deliver on your UVP

A good activation flow needs to deliver on the promise established on your landing page. When you map out your activation flow, make sure it demonstrates your UVP—preferably in one sitting. You only get one chance to make a good first impression.

Be prepared for when things go wrong

Offer inline troubleshooting and provide multiple ways customers can reach out for help: email, a 1-800 number, and so forth.

Building a Marketing Website

The purpose of your marketing website is simple: to sell your product.

Your marketing website is critical in driving the acquisition trigger in your customer lifecycle.

Acquisition describes the path a customer takes from first landing on your website as an unaware visitor to becoming an interested prospect.

Following the principle of architecting for learning over optimization, I recommend starting with explicit pages for each step. Each page should have a primary call to action and a secondary call to action. My primary call to action directs visitors to my pricing page (acquisition subgoal), while my secondary call to action offers a link to more information (e.g., product tour).

The landing page is by far the hardest of the three. Its job is to make a case for your product to an unaware visitor in fewer than eight seconds. I’ll list a few other pages you probably should also include:

About page

While the job of your landing page is to provide a compelling reason to buy your product, the job of your About page is to provide a compelling reason to buy from your company. This is your opportunity to put a face on your product, to tell your story, and to connect with your customers.

Terms of Service and Privacy Policy pages

Both of these pages are basic requirements for offering a service on the Web. They are also fairly standard, with lots of good examples online to model after. That said, the Terms Of Service and Privacy Policy can create legal headaches if they are not adequate. Put in at least enough time researching this part of your site to satisfy yourself that you are relying on decent models. If you have doubts, you should get some competent professional advice.

Tour page (video/screenshots)

I usually defer this page to later and start with just the landing page. But if your customers are more analytical or research-oriented, you might need to provide a separate page with more details, technical specifications, and so forth. It comes down to fundamentally understanding your customers and their motivations.

Cohort Analysis

A cohort is a group of people who share a common characteristic or experience within a defined period (e.g., are born, are exposed to a drug or a vaccine). Thus a group of people who were born on a day or in a particular period, say 1948, form a birth cohort. The comparison group may be the general population from which the cohort is drawn, or it may be another cohort of persons thought to have had little or no exposure to the substance under investigation, but otherwise similar. Alternatively, subgroups within the cohort may be compared with each other.

For our purposes, a cohort is any property that can be attributed to a user. The most common cohort used is “join date,” but as we’ll see, this could just as easily be the user’s “plan type,” “operating system,” “gender,” or something else.

What to Learn on the MVP Interview

With your MVP, marketing website, and conversion dashboard ready, you are all set to pay your prospects another visit. Your objective is to sign them up to use your service and, in the process, test out your messaging, pricing, and activation flow.

If you can’t convert a warm prospect in a 20-minute face-to-face interview, it will be much harder to convert a visitor in less than eight seconds on your landing page.

During the MVP interview, you are specifically looking to answer the following questions:

Product risk: What is compelling about the product? (Unique Value Proposition or UVP)

Does your landing page get noticed? Do customers make it all the way through your activation flow? What are the usability hot spots? Does your MVP demonstrate and deliver on your UVP?

Customer risk: Do you have enough customers? (Channels)

Can you bring on more customers using your existing channels?

Market risk: Is the price right? (Revenue Streams)

Do customers pay for your solution?


Your first objective during trials is to reduce user abandonment on your acquisition and activation paths. Your next objective is to increase retention and engagement, get paid (if that applies), and collect favorable customer testimonials.

Your goal should be to get 80% of your early adopters through the complete cycle. Because you’ve manually qualified your early adopters until now, this number needs to be higher than what you might typically expect after you publicly launch your product.

Acquisition and Activation

Priority: Ensure that you are driving enough traffic to support learning.

Drill into your subfunnels

Explore your acquisition and activation subfunnels to see where users are dropping off.

Start with the leakiest bucket first

Are you losing them on a particular page, such as the landing page or pricing page?

Look for patterns

Do certain types of users (e.g., Mac versus Windows users) experience higher failure rates than others?

Reach out to your users

You should be able to extract the list of users that failed at a particular step in your funnel. If you know what went wrong, correct it, and ask those users to come back. If you don’t know what went wrong, reach out with an offer for help (more like a call for help).

Catch and report unexpected errors

When early users run into problems, they don’t turn into testers. They leave. To be able to still learn from their experience, catch and report unexpected errors so that you can troubleshoot the problem without them.


Priority: Get users to come back and use your product during the trial.

Send gentle email reminders

Email is a very effective (and often underutilized) medium for engaging your customers. Everyone has an email address. Email can be automated, tracked, and measured.

A common technique used by email marketers is drip marketing, where you schedule a set of predetermined messages to your users over time. Even interested users get busy and distracted, and gentle reminders can help bring them back to your product.

But even better than drip marketing is lifecycle marketing. Lifecycle marketing additionally considers the user’s stage in the customer lifecycle. So, for instance, if a user gets stuck during activation, instead of educating him about your advanced features, you would know to send him timely and appropriate troubleshooting help.

Follow up with your interviewees

During the MVP interview, you asked for permission to follow up with your early adopters. Follow through. Call them up or meet with them and get their feedback.


Priority: Get paid.

Implement a payment system

Now is the time to implement a payment system for customers to pay you.

Get paying customers to talk to you

Get them on the phone, thank them for upgrading, and ask them:

  • How they heard about you (if you don’t know)
  • Why they bought from you
  • What could be improved

Get “lost sales” prospects to talk to you

You stand to learn as much (if not more) from your lost sales as you do from your sales. While some people are happy to provide honest feedback if you make a sincere request at the end of the trial, others might need a small incentive. Offer a $25–$50 gift card or donation to charity in exchange for 15 minutes of their time.


Priority: Get testimonials.

Ask for customer testimonials

Get happy customers to write a short paragraph on your product’s value proposition.

Getting Ready to Launch

You are ready when at least 80% of your early adopters consistently make it through your conversion funnel. Specifically, they should:

  • Be able to clearly articulate your unique value proposition (UVP)
  • Be primed to sign up for your service
  • Accept your pricing model
  • Make it through your activation flow
  • Provide positive testimonials

Once you have a minimum viable product (MVP) that works, your final step is to revisit your acquisition channel(s) to ensure that you have a steady stream of prospects entering your funnel. However, be wary of spending a lot of effort prematurely optimizing your acquisition channels at this stage.

Strive to drive this traffic through the actual channels you’ve identified for your product (e.g., content marketing), but supplement with other means if needed (e.g., search engine marketing).

Your goal is to establish “just enough” traffic to support learning.

If you have a large list of “warm” prospects from your earlier efforts (teaser page, referrals from interviewees), consider exhausting that list first in the form of more “early access” signups before doing a public launch.

Constraint the Features Pipeline

A good practice for keeping your features pipeline in check is to limit the number of features that can be concurrently worked on and only work on new features after you’ve validated that the features you just deployed had a positive or negative impact (i.e., yielded learning).

A great way to do this is to use a Kanban board (or visual board).

A Kanban board is to feature tracking as a Conversion Dashboard is to metrics tracking. Both let you focus on the macro.

Feature Requests Process

The first determination involves checking the request against your product’s immediate needs and priorities: is it “Right action, right time?”. So, for instance, if you have serious problems with your signup flow, all other downstream requests should take a backseat to that.

After that, you need to consider whether this is a small feature/bug fix or a larger MMF.

If this is a small work item and something that is needed immediately, fix it right away (i.e., code-test-deploy using your Continuous Deployment process). Otherwise, add it to your task board’s Backlog bucket. I recommend also keeping the task board Backlog in priority order. That way, anyone on the team can simply pull off a small work item and push it all the way through deployment when she has some idle time.

If this is a larger MMF, it goes on your Kanban board’s Backlog bucket.

The Feature Lifecycle Process

Understand problem

1. Backlog: Once you have identified a feature, the first step is to test to see if the problem is worth solving. If you can’t justify building the feature, kill it immediately.

  • Customer-pulled requests: If the feature is a customer-pulled request, arrange a call or meeting with the customer. Even though the customer might be asking for a specific solution, get to the root of the problem. Try to talk the customer out of wanting the feature. Have the customer sell you on why you should add the feature. At the end of the call, you should be able to assess whether this is a nice-to-have or a must-have problem, whether it is worth solving, and which macro it will affect.
  • Internal requests: If the feature was generated internally, review the same criteria as shown earlier with other team members, and similarly get to an “Is this worth solving?” determination for this feature.

Define solution

  1. Mock-up: Once you have a feature worth building, build a mock-up. Start with paper sketches, but quickly get to HTML/CSS views that are ideally accessible from within your application.
  2. Demo: With the mock-up ready, conduct an interview similar in structure to the Solution interview that tests your solution with customers. Iterate as needed on the mock-up until you have a strong signal to move forward.
  3. Code: With the mock-up validated, you can now start to build the functionality behind the feature. It will most likely make sense to break the feature into a number of smaller work items that you can track using your task board and deploy incrementally using your Continuous Deployment system.

Validate qualitatively

  1. Partial rollout: Once the feature is coded and ready for use, partially deploy it to just a few customers first.
  2. Validate qualitatively: Conduct usability interviews similar to the MVP interview. Iterate as needed to correct issues.

Verify quantitatively

  1. Full rollout: You are ready to do a full rollout. Once your feature is rolled out, it is marked “Done,” and the lock of the work-in-progress limit is released. This allows you to start working on the next feature in the backlog queue.
  2. Verify quantitatively: With the feature fully live, you should now be able to compare your conversion cohorts for the week the feature went live against the previous week to verify the expected macro impact. Depending on the type of feature, you might additionally need to set up a split-test. Split-testing is a matter of judgment at this stage. The more concurrent split-tests you have going, the longer the verification time window. Longunning experiments can also start interfering with other experiments and complicate your cohorts. For these reasons, it is best to use your judgment to decide when to split-test and when not to.

The Sean Ellis Test

How would you feel if you could no longer use ?

  1. Very disappointed
  2. Somewhat disappointed
  3. Not disappointed (it isn’t really that useful)
  4. N/A – I no longer use

If you find that over 40% of your users are saying that they would be “very disappointed” without your product, there is a great chance you can build sustainable, scalable customer acquisition growth on this “must have” product. This 40% benchmark was determined by comparing results across hundreds of startups. Those that were above 40% are generally able to sustainably scale the businesses; those significantly below 40% always seem to struggle.

Focusing on the Right Macro

The first is primarily driven by the experience of the service, which can be effectively measured using the activation metric. The second also relies on a good first experience (so good activation is still important), but success is driven through repeat usage—making retention the more indicative measure of “building something people want”.

Identifying the Key Engine of Growth

Here are some general guidelines to make the selection process easier:

1. Start with validating your value metrics.

Every product has to start by demonstrating and delivering a basic value proposition to customers.

2. Understand how customers behave with your product.

Study your baseline customer lifecycle to identify any particular usage patterns:

  • If you have implicit virality built into your product—that is, users repeatedly bring in other users as a natural side effect of using your service (e.g., Facebook and Twitter)—you might consider investing in a viral engine of growth. Often, that also drives the lowering of signup friction, such as making the service free to maximize user growth.
  • If you have a recurring use model—for example, a Software as a Service product—it might be worthwhile to invest your effort initially to drive up the lifetime value of your customers by reducing your churn rate. At some point, you will hit a ceiling of diminishing returns, which might be your cue to switch to another engine of growth, like paid. In these types of products, even though you might have some referrals, the referrals do not repeat beyond one or two degrees (i.e., the viral coefficient is less than 1).
  • If you have a one-time-use product that isn’t also viral, such as the wedding photographer and divorce attorney examples, your only bet is to invest in the paid engine of growth. Again, your product might exhibit word-of-mouth referrals, and you may even have repeat customers, but neither of these are key to driving sustainable growth.

3. Pick an engine to tune.

Once you’ve selected your key engine of growth, put a stake in the ground: Declare the key metric and improvement you want to achieve. Then, align your next set of experiments toward that goal.

Life after Product/Market Fit

Along with continually tuning and resetting your engine of growth to meet customer adoption challenges as you attempt to “cross the chasm” between early adopters and mainstream customers, you will inevitably also be faced with new challenges as you grow your company.

Every process works well until you add people.

The key is to build a continuous learning culture of experimenters versus specialists, where it’s everyone’s job to be accountable toward creating and capturing customer value.

How to Achieve Flow in a Lean Startup

Activities that flow typically have the following attributes:

  • They have a clear objective.
  • They need your full concentration.
  • They lack interruptions and distractions.
  • They provide clear and immediate feedback on progress toward the objective.
  • They offer a sense of challenge.

Creating Daily Flow

Work Hack 1: Establish uninterruptible time blocks for maker work

My planned maker activities are typically coding and writing tasks I’ve previously identified. Because these activities need an uninterruptible block of time, I schedule these very early in the morning (6:00 a.m.– 8:00 a.m.). I usually schedule this task the night before, and it is the first and only thing I do. I don’t check email or Twitter or look at anything else. No one is calling at that hour, so distractions are at a minimum. I find two-hour blocks work best for me.

Work Hack 2: Achieve maker goals as early in the day as possible

I’ve tried both staying up late and waking up early, and I prefer the latter as it isn’t interrupted by sleep, which allows the day’s activities to flow better. I also find that accomplishing something tangible that early in the day sets the tone for the rest of the day. Depending on the day of the week, I might allocate more two-hour blocks later in the morning or afternoon, but they aren’t as intense as the first one and can be interrupted by something more urgent.

Work Hack 3: Schedule manager activities as late in the day as possible

Planned manager activities, like customer meetings, are easier to schedule because they are clearly time-boxed and calendar-driven. Unless there is an unworkable schedule conflict, I prefer to schedule these for the afternoon so as not to interrupt my morning flow.

Work Hack 4: Always be ready for unplanned activities like customer support

Unexpected interruptions can surface from anywhere throughout the day—server issues, customer support calls, and so on. You have to be prepared for interruptions, especially from customers. Both server alerts and customer calls (1-800 number) are routed directly to my mobile phone. This is also a good place to apply a Five Whys process to ensure that unexpected incidents don’t become recurring incidents (I will discuss this process in more detail shortly).

Creating Weekly Flow

Work Hack 5: Identify the best days for planned Customer Development

For instance, Mondays and Fridays are usually bad days for initiating new customer contact, as people are generally either recovering from the weekend or getting ready for it. I plan these types of customer development activities for Tuesday through Thursday.

Work Hack 6: Take advantage of customer downtime

Since Mondays and Fridays are usually slower from a customer perspective, I use these days for larger maker tasks, like writing blog osts. My blog posts are usually identified on Friday, outlined over the weekend, written/proofed on Monday, and published on Tuesday.

Work Hack 7: Balance face time with customers

Not all customer development activities require face time. Beyond the initial customer discovery stage, there is a strong tendency to rely more heavily on asynchronous communication using tools like email, forums, and online usability testing. While all these tools are great for lowering real-time distractions and achieving scale, I find it important to still create opportunities for face time with existing and new customers. Unscripted conversations are the best way for learning about unscripted problems.

I put our 1-800 number on all pages and encourage customers to pick up the phone versus emailing whenever possible.

Eliminating Sofware Waste

Work Hack 8: Avoid overproduction by making customers pull for features

Customer pull is another concept from “Lean,” and it requires that no product or service be produced until a customer asks for it. Eighty percent of your effort should be spent toward optimizing existing features versus building new ones. The whole point of Customer Development is to identify an MVP that resonates with customers, and the whole point of customer validation is to test whether that resonance will scale. If it doesn’t, the answer is not adding features, but possibly pivoting and going back to Step 1: customer discovery.

Work Hack 9: Iterate around only three to five actionable metrics

A few actionable metrics are all you need to identify and prioritize the most critical issues to tackle.

Work Hack 10: Build software to flow

You might have noticed that I don’t have days or tasks identified for building, testing, or releasing software. That is because I follow a continuous deployment process (also popularized by Eric Ries) where software is built, tested, and packaged automatically at the end of every maker task, with no effort on my part other than checking in code. One click, and the code is released to customers. Manufacturing processes have traditionally been arranged around machine time-breaking tasks into batches and queues. “Lean” challenges this approach and calls for arranging around human timeorganizing tasks so that they flow.

How to Set Pricing for a SaaS Product

Start with a single pricing plan

Starting with multiple plans that cover everyone under the sun is a form of waste. I’ve seen startups launch with plan options targeting one-person startups to enterprises composed of more than a thousand people.

Not only does supporting multiple plans require you to write more code to support plan/feature segmentation, but the return on learning is diluted when you attempt to target multiple customer segments all at once. In the example in the preceding paragraph, the business models and tactics vary greatly when selling to startups as opposed to selling to enterprises.

The bigger point here, though, is that when you’re starting out, you don’t yet have enough information to know how to correctly price or segment the feature set into multiple plans.

Use a “Free Trial” plan

Time-based trials help time-box your pricing experiments so that you can force a conversion decision, which allows you to learn and iterate faster.

Pick a price to test

Existing alternatives create “reference points” in the minds of customers that they will use to rank your solution, so it’s important to understand and position your price against them.

In the rare case that you are actually solving a brand-new problem or don’t have clear reference points (more common in enterprise-based products), you might have to pick a starting price out of thin air and refine from there.

Take your costs into account

The ultimate goal is to find a scalable business model, so it should go without saying that you also need to keep an eye on what it would cost you to deliver your solution and ensure that you have a healthy margin built in.

One rule of thumb for building a successful business (by way of David Skok, Matrix Partners) is to ensure that the lifetime value of your customers exceeds the cost of customer acquisition by at least a factor of three.

It’s hard to accurately calculate these at this stage, so instead, do a back-of-the-envelope calculation based on your people/hardware costs and subscription revenue to find your break-even point.

How to Build a Teaser Page

The number-one way to get a prospect (cold or warm) to agree to an interview is to “nail his problem.”

One of the best exercises for crafting such a message involves spending an afternoon writing a shorter version of a long-form sales letter—no matter what type of business you’re building.

You will not be sending this letter to any prospects. The point of the exercise is to get you to explain your product in narrative form, which will be helpful when requesting interviews, when conducting interviews, and while building a marketing landing page.

How to Write a Sales Letter

While there’s more that goes into a complete sales letter, I recommend starting with just your unique value proposition (UVP), problem, and solution.

Make a large promise (UVP)

This is a short headline that summarizes what your product will do for the customer (i.e., the finished story benefit).

A good formula we used earlier (by way of Dane Maxwell) is:

Instant Clarity Headline = End Result Customer Wants + Specific Period of Time + Address the Objections.

Psychological principle in play: Attention through surprise, clarity, and bold promise.

Connect with the customer (Problem)

This is a short paragraph that explains the problem from the customer’s worldview. You want to visualize the customer nodding his head in agreement. During your interviews, check for this. Psychological principle in play: Empathy, by showing you understand the customer.

Generate interest/desire (Solution)

Then, state what your product does in another short paragraph (i.e., how it solves the problem) and list your top three features written as benefits.

How to Build a Conversion Dashboard

A key design principle is to decouple data collection from data visualization.

This lets you minimize waste by allowing you to build your conversion dashboard incrementally.

Here’s how to get started with data collection:

Map metrics to events

The first step is to identify all the key events (user actions) that map back to your metrics. You should already have all your steps for your acquisition and activation funnels clearly defined.

It is helpful to also identify any key events for the other macro metrics.

Track raw events

I recommend tracking your raw events in a separate events table/database or using a third-party system like Google Analytics, KISSmetrics, or Mixpanel.

Log everything

A good practice to complement tracking raw events is to log every “potentially interesting” property along with each event. An example of a property could be your user’s browser, operating system, or referrer. While you may not use a particular property today (or think you’ll ever use it in the future), it’s fairly inexpensive to log a few extra bytes of information that could end up saving you time later, and more important, could provide a treasure trove of historical data.

How to Visualize Your Conversion Dashboard

Now you’re ready to start visualizing your data:

Build a weekly cohort report

The first report I use on my conversion dashboard is the weekly cohort report by join date I showed earlier.

Be able to drill into your subfunnels

You should be able to drill into your detailed subfunnels and visualize all the steps, which is valuable for troubleshooting problems.

Be able to go behind the numbers

At any given subfunnel event, you should be able to go behind the numbers and get to the list of people.

👉 Buy Running Lean on Amazon

🎧 Listen for free on Everand (plus 1+ million other books)

Tomas Laurinavicius

Hi! I'm Tomas, a writer and growth marketer from Lithuania, living in Spain. I'm always involved in multiple projects driven by my curiosity. Currently, I'm a marketing advisor at Devsolutely and a partner at Craftled, building Best Writing and Marketful. Let's connect on X and LinkedIn.