25 October 2010

Hypotheses and testing

Customer development for publishers

I’m continuing my thoughts and research notes on electronic publishing here, obviously written mainly as personal notes, unedited, disorganised and possibly quite incoherent.

Since my earlier research notes were comprehensively poo-pooed by the few that read them, take all of the following with a grain of salt.

Most useful theories and models of entrepreneurship are subsets of Saras Sarasvathy’s Effectual Reasoning model. The core of her idea is to use the dichotomy of causal reasoning versus effectual reasoning to analyse the differences between the established business decision making process and the entrepreneurial decision making process.

The short version is that entrepreneurs don’t begin with the causes or what they’d like to specifically accomplish and make (causal reasoning) but with what they can effect with their resources and skills and work from there (effectual reasoning).

Even shorter version: The reason why I have no interest whatsoever in print publishing is that I’ve been making websites for 16 years, did my PhD on ebook interactivity and have been working for several years now in software sales and marketing (ebooks are, IMO, for the purposes of sales and marketing, a subset of software).

Ebooks are the only subject that seem to tie together my interests, skills, experience and training. The reason why I don’t write about the role of print in these research notes isn’t because I don’t believe print hasn’t a future but that these notes are personal, foundational research on what possibilities there are for me to go in life.

Might not amount to anything. Consider this me thinking out loud.


One well-tested and currently popular software entrepreneurship model is Steve Blank’s Customer Development, built around four steps in a company’s buildup.

These definitions are taken from Brant Cooper’s “What is Customer Development?”

Each step is supposed to test and validate a business assumption.

Customer Discovery. Assumption: “A specific product solves a known problem for an identifiable group of users.”

Customer Validation. Assumption: “The market is saleable and large enough that a viable business might be built.”

Company Creation. Assumption: “The business is scalable through a repeatable sales and marketing roadmap.”

Company Building. Assumption: “Company departments and operational processes are created to support scale.”

You only exit each stage once you have proven and tested the assumption for that step. You don’t leave Customer Discovery unless you have some empirical evidence that your product “solves a known problem for an identifiable group of users”, for example.


Most people don’t see how Customer Discovery is relevant to electronic publishing but that becomes clearer if you rephrase the steps:

Step One: Formulate your hypotheses about what readers want to read, in what formats, at what prices, and where they’re likely to find them. Or: In what format and at what price is the book you have written (or about to write) likely to solve a problem for the largest number of readers.

Boredom, under-stimulation, and skills development are three problems books are good at solving. Throughout step one you test your hypotheses about the variables above.

This can be done by designing a landing page with a book preview that collects the e-mail addresses of those interested to be notified when the book is released (useful for establishing contact with readers for interviews, there’s a lot you can only find out from people in a phone interview). This test has the advantage of being easily iterated, given the right circumstances you should be able to test for interest in varying formats, prices, covers and titles.

An alternative test is to set up a simple ecommerce site and sell an early draft of the ebook (Minimum Viable Product, in Customer Development lingo) using e-junkie, cartloom or Big Cartel’s upcoming pulley app (if you’re lucky enough to be a beta tester). This has the benefit of giving you a large degree of certainty in your findings (since they’re willing to part with their money) but possibly has a lower conversion rate because a lot of readers aren’t willing to buy from an unknown ecommerce site.

The third test is identical to the second, but instead of using your own ecommerce cart, you do a low-key release of your book on the Kindle and link to that and rely on affiliate data to make up for not being able to track the checkout/sales process. Benefit: Possibly a more realistic test of actual market performance. No need to set up an ecommerce chart on your own site.

You don’t exit step one until you have gathered empirical evidence with a series of tests that leaves you reasonably certain that your format/price/book combination has an identifiable readership (if you don’t know who they are, you can’t sell to them). Surveys are also useful once you’ve started to form a picture of what your readership looks like.

Identifying the readership is, in my opinion, more important than tailoring the book for a specific readership (which is just causal reasoning). That’s why I’m not talking about testing variations in the book itself. The goal is to identify who actually wants to read the book and what formats, platforms and prices suit them, and, just as importantly, what sort of features or services your website needs to support.

The way to maximise the value of all of this research is, obviously, to gear it around testing for a series rather than a single book, but that’s a whole other research note to write, I think.


Step Two: Referring back to Brant Cooper’s overview of Custer Development linked above, this step consists of three things. 1. You have identified the target readership. Those of your identified readership that have encountered it are passionate about your book/series and you understand the core value that you are providing to them. 2. You understand how to market and sell to that readership and “have proven that every dollar that you put into marketing and selling your [book] results in more than one dollar back.” 3. Your target readership is large enough for the sales and marketing processes to be scaled up to a level where they can sustain a publishing business (which might just be yourself and freelance help, or might involve something larger).

In Steve Blank’s own words, step two or Customer Validation, is “where you develop a sales model that can be replicated and scaled.”

He also likes to say that “facts reside outside your building” so the only way to really find out what’s going on is to get out of the building and talk to customers.

Step one and step two are also iterative. You often find out in step two that you’ve got something wrong and that you need to go back to step one and reiterate the discovery/hypothesis testing process. That is good. Every time that happens you’ve discovered something important.

The core of step two is to figure out how to sell your book or series. It doesn’t matter whether it’s word of mouth, search engine ads, Facebook ads, direct sales/marketing tactics, the various possibilities need to be tested and measured and repeated. The only thing that matters is that for every dollar you spend, more than one dollar returns in sales and that the difference is enough for a sustainable and scalable publishing business.

Only when you have developed a process that can be repeated and scaled do you exit step two.

Also note that wide distribution immediately sabotages any hope you have of exiting step two. It’s hard enough to try develop a sales and marketing process for one platform. The Smashwords tactic of uploading one file and automatically distributing it to a dozen large platform is foolish, short-sighted and completely devoid of any sort of understanding of where and how people are supposed to find your book.

For step one and two, stick to only a couple or even just one target distribution platform. You can always scale up to wider distribution once you have tested and validated the basics.


Step Three: This is where you spread out to other distribution platforms. This is when you roll out support for the other ebook stores you hadn’t supported before. This is when you start investing, properly, in sales and marketing.

This is actually the first time you think about launching, pitching stories to the press, give interviews, etc.


Step Four: This is when the publishing house is actually born, when you shift your research and development efforts from discovery to providing the most value to your readership.

Everything on this site is written by . In case you hadn't guessed already.