The Fallacy of Customer Development

One of the lessons learned from my last post had to do with the Fallacy of Customer Development:

Unless you are building a business (like Enterprise software) where the primary channel to customers is through direct sales, customer development is NOT a scalable way to reach customers. Instead, Customer Development is a form of qualitative learning and while it’s the fastest way to learn from customers, that alone may not be enough. The biggest challenge most web applications face is building a significant path to customers. Rather than going through customer discovery, customer validation, and then tackle customer creation, which can take months, I find it critical to start building and testing “significant enough” paths to customers much sooner.

A few folks (including Sean Murphy and Patrick Vlaskovits) weighed in with insightful comments and questions which makes this a worthy topic for further elaboration. First I’ll explain the fallacy (hopefully more clearly) and then offer some solutions.

The Fallacy

In order to build a successful product, you have to eventually find a scalable and repeatable way to reach customers. There is an implicit expectation that customer development will uncover that path to customers. My experience (with web based products) has been that it’s not as much the uncovering of the path but the building of the path that is troublesome. Some paths are obvious but hard, such as those built on referrals (word of mouth), SEO, etc. It’s comparatively a lot easier to find 30-50 people, validate you have a problem worth solving, build a MVP, even get them to pay you – all of which is a false positive if it was predicated on a customer acquisition approach that won’t scale or more importantly be applicable to how you acquire customers in the future.

The fundamental cause for this fallacy is rooted in an attempt to literally apply Steve Blank’s Customer Development techniques without consideration for specific business and channel types (of which I’m just as guilty). “The Four Steps to the Epiphany” was written for a specific type of business – Enterprise Software, and a specific type of channel – Direct Sales. While Customer Development is still the fastest way to learn, you have to adapt the tactical techniques for your business and channel to avoid hitting a wall after Problem/Solution Fit.

I have been in search of one such adaptation for web businesses where the primary selling channel is through a marketing website.

Highlights follow:

Customer Development As Adapted for Web Apps

Customer Development needs to be a continuous learning process

Customer Development is about establishing a continuous feedback loop throughout the product development cycle. Enterprise selling necessitates constant contact with customers throughout the customer lifecycle. This is NOT the case with web software. To fix this, you have to build appropriate customer touch points in the form of ongoing usability tests, customer feedback calls, interviews, etc.

Start building and testing a path to customers from day one

Seasoned entrepreneurs know that building a significant enough path to customers is one of the hardest aspects of building a successful product. I always get at least one question from the audience during my talks/workshops on “finding prospects”. I used to answer this question with the canned response of “make a list of people you know, start there, ask for referrals, interviews will reveal the path to customers”. Now, I slant my response more heavily towards finding prospects by way of testing the actual channels you intend to use for reaching your future customers.

Don’t get me wrong. Talking to anyone is still way better than talking to no one. If you have no idea how to reach early prospects or if the channel takes time to build, start with your 1-degree network. But, don’t declare Problem/Solution Fit unless you’re able to recruit a fair number of your interviewees using an actual channel you will use.

It’s also equally important to point out not to fall in the trap of prematurely optimizing this channel. You may not have a problem worth solving or have to pivot to a different customer segment. Premature optimization is a form of waste. Your first objective should be driving just enough traffic to support learning. During Customer Discovery, that means enough traffic to yield 30-50 interviews.

Some examples –
1. Write a blog post that ends with a call for participation in an interview
2. Create a landing page and drive SEM traffic with a call-to-action ending in an interview (not a survey)

Start building and testing your selling process as soon as possible

While Customer Development does NOT magically scale even for Enterprise software, the learning from the interviews can more easily be applied towards building a repeatable and scalable direct sales process. Selling a product over 15-20 minutes in an interview is very different from selling a product in 5-8 seconds on a landing page. Again early interviews are helpful in identifying what’s important (your unique value proposition) but you need to start testing that in the right format (e.g. landing page) as soon as possible.

Some examples –
Run customer interviews in a usability test format.
1. Instead of verbalizing the unique value proposition, show them a landing page and test positioning.
2. Instead of getting a letter of intent, watch them sign-up to your service and note where they get stuck.

Retention, not Revenue is the Ultimate Validation

Getting paid is only the first form of validation. Providing ongoing value (as measured by customer retention) is the ultimate validation. While this is true for both Enterprise and web businesses, early revenue plays a bigger role towards customer validation in Enterprise software than web software. For one, the purchase order is a lot bigger. Also, because the sales process is a lot harder to close, the barriers to loosing a customer are higher compared to web based software that can be canceled at any time.

Here again, Enterprise Software has a natural customer feedback loop built-in – in the form of account managers whose job is to periodically ensure customer engagement and retention are healthy. In a web based business, you have to build and monitor that feedback loop yourself with potentially thousands of users.

Some examples –
1. Implement an automated lifecycle marketing system that helps drive retention and engagement.
2. Measure/optimize for retention
3. Get Community Managers

Customer Development HAS to scale

I believe that in order to realize Eric Ries’s full vision of the Lean Startup, Customer Development HAS to scale.

  • It can’t just be applied at the head and tail ends of the product development cycle but needs to be ingrained throughout the product development cycle.
  • It can’t just be applied to finding a problem worth solving and building a MVP but needs to be part of an ongoing process for how features are built and validated.
  • It can’t just be limited to interviews, but needs to incorporate other forms of qualitative learning (usability tests) and techniques for reaching customers (lifecycle messaging).

What do you think?

Become a Practice Trumps Theory member. It’s free.

Get exclusive updates, worksheets, and resources delivered to your inbox.

  • Pingback: Tweets that mention The Fallacy of Customer Development --

  • jonathanjaeger

    Great post, especially the part on premature optimization as a form of waste. No point spending time on fine-tuning something that will ultimately be scrapped altogether.

  • Michael P. Daugherty

    This was a very timely post for me. My New Year resolution is to find more sustainable ways of generating leads to my website ( 60-70% of my customers immediately reorder our products, which indicates they’re happy with the quality, but I’m still having a lot of trouble bringing new customers to the site.

    Last year I had one piece of marketing success in November with a Lifehacker about another small web tool of mine that ended up leading some people to my main site (e.g. where I can make money), but I found that that sort of novelty traffic disappears pretty quickly. This time around, I’m focusing on finding a sustainable path to customers, even if that path builds gradually, instead of with a big bang like last time.

    Since this is my first time posting here, I also want to say thanks for such a great blog. It’s refreshing to be able to read specific, immediately useful applications and tips of lean startup theory from someone who is using them in practice.

  • Scott L Caruso

    Excellent. Articulate and insightful. I started by selling enterprise hardware/software (Sun). Then moved to selling Enterprise services (ISP). Over the past 10 years I’ve been involved in business that sell SaaS to both consumer and professional (enterprise and small biz). Recognizing the distinction between customer profiles, engagement style (long term feedback loop) and scaleable acquisition & retention are key for any business.

  • Scott L Caruso

    Excellent. Articulate and insightful. I started by selling enterprise hardware/software (Sun). Then moved to selling Enterprise services (ISP). Over the past 10 years I’ve been involved in business that sell SaaS to both consumer and professional (enterprise and small biz). Recognizing the distinction between customer profiles, engagement style (long term feedback loop) and scaleable acquisition & retention are key for any business.

  • Pingback: Customer Development and Marketplaces | Market By Numbers()

  • Pingback: The Fallacy of a Literal Interpretation | Vlaskovits()

  • Nino Baraka

    I have been following your blog for a while now and I appreciate your reflections, I that you are absolutely right when it comes to your thesis and I will try to explain why.

    As a systems thinker and an executive MBA I started to try to find a system perspective on entrepreneurship and following that perspective. I found that the best tools to use have been the Customer Development method. However it is an meta implementation process, the codification of the operations differs from industry to industry. I think the Lean Startup method is the Customer Development method that has codified the operations for the web startup industry.

    If you want to understand the codification process I can recommend studying Ray Krocs codification of McDonalds or Taiichi Ohnos codification of Toyota. Whenever you are not active on the web there is no natural feedback tool (especially in the physical world) so you need to use systems thinking to connect a continuous feedback system. It is in the feedback the learning and your assumptions are perfected however what the Customer Development method is doing that differs them is connecting strategy to operations and making the company a natural part of its environment (market). The old lean concepts tend to focus only on the codification of operations.

    You can follow some of my thoughts on my blog

  • Zoli1702

    Hi Ash & all,

    I’ve started to develop my subscription based web product later last year, based on your book& blog and Steve Blank’s book, and just came to the very same conclusion as you have.

    The resolution I managed to figure out is based on roughly the same principles you have shared in this post.

    1, I’m going to pull up a website with a landing page where I will test my initial UVP, in form of questions, like “are you tired of…?” etc. to validate the initial problem hypothesis. It will be a kind of sales letter but only one page short, and with a call to action at the bottom.

    2, Then this link will lead to a clickable mockup that represents the core user scenarios (first assumptions).

    3, Finally there will be a link with some clever questions to get feedback on the mockup and the problem hypothesis itself. The last question will be a kind of “persmission for follow-up” question, something like “Are you interested in participating/etcetc…?”. And also one question asking for referrals like it would be a live product.

    4, Then I will drive some traffic to the site, based on channel assumptions, so that I can test the cannels as well. This includes the classical custdev method like personal referral, but also includes social media and SEM (in may case).

    I think it will be fine enough. I see the following advantages:

    a, the whole process is a kind of a funnel: you can see where the targeted traffic quits the process so you can adjust

    b, you can start testing your UVP, core features, problem hypothesis and channels at the same time. It is true, you need to be careful not to waste too much on premature optimization though!

    c, The visual stuff is more catchy for the prospects that “Can we have a chat I about your daily habits in the field of yx”.

    Obviously it is not as effective as the classical custdev approach but it is better than nothing and also better than chasing prospects for months without any result.

    What do you think, will it fly?


  • John Seiffer

    Great post – but I’m not sure the title is accurate. It’s more like the limitations of customer development, not the fallacy. I know there is this idea that start-ups are not like “big” companies and there is a lot of validity in that. You can’t scale till you’ve found your business model. However let’s look at a large company model for some clarity. One function every company has to perform is generating leads. This is usually done by marketing. Then they have to convert those leads into sales. This is done by the sales department or team. Then a product is produced and delivered (usually by manufacturing dept. or production or something similar)

    I think what you’re describing is that Customer Development as a methodology is better for learning what should be produced (Minimum Feature Set) than it is at learning how to generate leads and turn them into sales.

    In my reading of Steve Blank’s work the part you’re calling customer development is what he calls customer discovery. And the part you’re struggling with he calls customer validation. That’s where you learn how to sell to these people (and presumably that includes a scalable model for generating leads).

    I would also toss out there that the web and all it’s tools are a benefit to making and testing cheaper feature sets (at least in some industries) but the web might not be the best way to market (and/or sell) many products. It’s major limitation is that there is so much that is not under your control. Remember the first word in SEO and SEM is SEARCH. If no one is searching your EO and EM don’t accomplish much.

  • Ash Maurya

    Thanks for your comments – I agree the title can use some additional qualifiers.

    Yes I am saying customer development is qualitative learning from customer which is the fastest way to learn. But I am also additionally saying, that just as Eric Ries drew a 2-way feedback loop between customer development and product development (agile), we need to draw a 2-way feedback loop between customer development and marketing/sales/support.

    Much like customer development can’t build your product, it can’t generate sales. But it should be a key part driving both.


  • Michael Scepaniak

    > we need to draw a 2-way feedback loop between customer development and marketing/sales/support.

    Well said, Ash. And fantastic points here. Your views are refreshingly semi-contrarian to the “established” Lean Startup thinking. You’re questioning and challenging the theories in ways- and to an extent- that I don’t see anybody else doing. And I’m sure Eric and Steve are loving it.


  • Ricky Yean

    Great post, Ash. As an early stage company founder who have built a product with passionate users but less understanding of channels, I am working on improving my understanding there through constructing a feedback loop.

  • Nino Baraka

    In response to Ash. A 2-way feedback loop comes from the Cybernetician (systems thinker and anthropologist) Gregory Bateson (later applied by Argyris in organizational learning) and he called it “double loop learning”. In his book “Steps to ecology of mind” he describes it thoroughly.

    The benefits of using systems thinking are that the 2-way feedback loop needs to be actively deployed in all departments and the company and its environment for it to act as “one” system. As you probably know agile programming was derived from systems thinking so if you want some help in getting the full perspective, I can recommend Batesons book. However it is always a challenge to make the abstract practical but that’s what the skill of “execution” is all about and that’s why I think “doers” (like yourself) should be the ones interpreting the abstract.

  • Willis F Jackson III

    This is all very specific to the web software application, which is useful, but I think it would be pretty useful to generalize it out of software altogether. If I am understanding you correctly, you are saying that you need to have an extra “fit”, the two being problem/solution and product/market. It is ok to follow 4STTE through problem/solution, but then you must validate that your product fits the market, including the channel, by testing for it specifically.

  • Rahul Chakrabarty

    Very well articulated. I have a question when you are developing a B2C product/solution, for the start-up point of view there are 2 customers – consumer and the business/companies. That means the customer development has to be two faced – an MVP to reach out to consumers showing what they will finally see and an MVP to show to the business the value proposition.

    This is the approach I am taking for my start-up and once I have +eve data from consumers (in terms of registrations) and business (in terms of registration/intent) that I want to launch the BETA which truly is the MVP. Just want to know if this at all make any sense?

  • Pingback: Ash Maurya Reconciles Customer Development with Web Apps Business Realities » Process for the Enterprise()

  • Ash Maurya

    Rahul – Are you describing a marketplace of consumers and businesses or two different customer segments?

    If marketplace, often you want to interview both sides of the equation but start with an MVP targeted at one and demonstrating at a smaller scale first.

    If you are talking about an app with consumers and businesses as potential customers, I’d recommend starting with one first.

  • Pingback: Por que ter um bom blog na sua Startup (e logo!) | Manual da Startup()

  • Pingback: Build Your Lean Startup Through Conversations()

  • Pingback:

  • Dave Doolin

    “Talking to anyone is still way better than talking to no one. If you have no idea how to reach early prospects or if the channel takes time to build, start with your 1-degree network. But, don’t declare Problem/Solution Fit unless you’re able to recruit a fair number of your interviewees using an actual channel you will use.”

    This is exactly where I’m at right now: finding anyone to talk to.  What I’m taking from here is that this initial round is qualitative, and I should use it to help bootstrap my way to quantitative evaluation.

    I suspect some of these mistakes are necessary for beginners to make, as it’s much easier in hindsight to see how it fits together.

  • Pingback: Business Model Validationscurri blog()

  • Pingback: Tekedia » The Startup Customer Development Model()

  • Pingback: Tekedia » The Startup Customer Development Manifesto()

  • Andres Valencia

    I think that much of this process always refer to software products, but, this applies if someone creates a restaurant, a dance academy, or consulting office?

  • Pingback: Good Reading Of The Week – 07/20/2013 | Thomas Pichon's blog()

  • Pingback: Por que ter um bom blog na sua Startup (e logo!) | Startup SC()

  • Pingback: Lean Startups: 17 Tips To Build Your Inbound Channel | Thomas Pichon's blog()

  • Pingback: Lean Startups: 17 Tips To Build Your Inbound Channel | TheCollaborativeStartup()

  • Pingback: How To Build Your Startup Community From Scratch – Part 1 | TheCollaborativeStartup()

  • Pingback: TheCollaborativeStartup()

  • Pingback: SKMurphy, Inc. Customer Development and Channel Development()