Peter Armstrong

The parallels between writing a book and doing a startup are numerous. I’m going to start by going over 3 of these parallels, and then introduce why the Lean Startup approach should apply to the process of writing a book. Specific technical details about how this can be done will be discussed in the next post.

Parallel #1: There are market risks, technical risks and a very low probability of success.

Doing a startup is a highly risky endeavour. There are market risks (is there a market for the product? will some other company beat us to market?) and technical risks (is the product even possible? will it be any good?). The founders take all these risks because of the potential for a huge payday and to change the world–this has been referred to as “changing the world, one million dollars at a time.”

Writing a book has a similar set of risks. There are market risks (will anyone want to buy the book? will someone else write a similar book first?) and technical risks (do I have what it takes to finish a book? can I write a book that is entertaining and engaging?). The author takes these risks for the possibility of fame and fortune–albeit a smaller fortune than can be acquired by building a successful startup.

Also, both startups and books can be derailed by personal issues. These can include interpersonal issues between co-founders or co-authors, or personal problems (health, relationship) that can crush a founder or author. These problems seem to happen more frequently than with normal 9-5 jobs, since doing a startup or writing a book is so much more demanding than a normal job.

All things considered, it’s completely unsurprising that the probability of success for both startups and books is so low. That people attempt either, and do so on a regular basis, is a testament to the entrepreneurial spirit. (Speaking as a Canadian who has lived and worked in Silicon Valley for a number of years before settling in Vancouver, the ability to take huge risks without a stigma attached to failure is one of the things I admire most about the culture of the United States, and of Silicon Valley in particular: this spirit is more alive there than anywhere else I have been…)

Parallel #2: Both writing a book and creating a startup are highly creative processes undertaken by one or a few people working closely together.

There is a reason that there are no startups founded or books written by 100 people: doing anything creative with that number of people produces a product in which any brilliance is diluted. (This is the reason why the phrase “designed by committee” is not a compliment.) A book or a startup is best created by 1 or 2 people, who are the authors or founders. You can create a book with 3 or 4 authors, but essentially all the great books have been written by one author. In fact, if you have more than 4 authors, you’re not even really producing a book–you’re really producing an anthology of individual essays. Similarly, a startup typically has 2 co-founders (Apple had Steve Jobs and Steve Wozniak, Microsoft had Bill Gates and Paul Allen, etc.). You can create a startup with 3 or more founders (BEA had three: B, E and A), but that’s the exception–and it’s extremely rare for any successful startups to have more than 4 founders.

Despite the similarities, there is actually an interesting difference between the ideal number of authors and founders. To achieve great results, it seems the ideal number of authors of a book is 1 and the ideal number of founders of a startup is 2. My personal take on the reason for the difference, having written two books and started one startup, is that writing a book is essentialy the technical portion of creating a startup–a solitary endeavour which requires long periods of sustained thought. However, creating a startup is a more multi-dimensional activity: besides the solitary time spent at your computer creating the product, there are a vast number of activities (raising funds, reacting to the market, acquiring customers and evolving along with their needs, etc) which are too much for almost all individuals to sustain over the time it takes to develop a product.

Parallel #3: Historically it has taken about a year, often spent in isolation or “stealth mode” and funded by outside investors (angels or VCs, publishers), to develop and release the first version.

Startups have traditionally operated in secrecy from the moment of being founded until the “first customer ship” or “product launch.” This period of time is spent raising money from angel investors (accredited wealthy individuals) or VCs, possibly hiring a few early employees (depending on the amount of money raised) and then rapidly building the product envisioned by the founders in order to get it to market before anyone else does. Startups that don’t raise money support themselves by bootstrapping, either on the savings of the founders or by doing other jobs (typically consulting) until the startup is earning enough money to feed the founders. (This achievement is called being “ramen profitable” by Paul Graham, and it is an important milestone since it means that the startup founders can focus more on their business and less on scrounging up money from investors or consulting.)

A similar situation exists for authors. One risk is that someone else will write essentially the same book first. After all, while there are an infinite number of book ideas to be had, it seems that there is a finite number of good ones at any given moment in time. Assuming this is the case, since there are millions of potential competitors, authors are justifiably paranoid that someone else will publish first–especially if the subject is topical, or is a Ph. D. thesis. Another parallel with authors and startup founders is the role of the publisher. Just as startups are often funded by angels or VCs, books are often funded by publishers. A publisher typically pays an advance against royalties, which the author typically uses to replace some of the lost income that s/he would have earned during the time spent writing the book. Since the prospect of repaying an already-spent advance is distasteful, this typically gives the publisher a lot of leverage over the author in terms of the content of the book. This can be a good thing, since it incents authors to actually finish their books, but it also can lead to a book being finished prematurely or at a lower quality.

Coupled with the focus on speed to launch and secrecy is the fact that both publishing and venture capital are hit-driven businesses. Venture capitalists (VCs) build their businesses on the assumption that only 1 of 10 startups they fund will be a hit, 2 of 10 will limp along, and 7 of 10 will completely fail, losing all the money invested. Similarly, publishers produce many books but get the bulk of their profits from only a handful. Once these books are developed and begin to be marketed, the publisher gets a better sense of which books (if any) have the potential to be breakout successes, and focus their marketing efforts on those ones. Since publishers want to produce hits, they tend to force all books to seek as wide an audience as possible. This presumably isn’t much of a problem for fiction, but it is a problem for technical books such as computer books, since it forces much introductory material (that the author doesn’t want to write) into books, in order to theoretically broaden the potential market for the book as much as possible. (Of course, the risk is that this material will bore and alienate some readers, but publishers seem willing to take this risk.) This doesn’t just apply to computer books: how many otherwise-decent business books about startups seem compelled to explain (badly) the history of Silicon Valley or who Marc Andreessen is?

So, when you put combine all these factors, the outcome is that with a startup or book, what typically happens is that the founder(s) or author(s) get some money, lock themselves in a room and slave away to produce something targeted at a very broad market (it has to be a hit, remember?), which the market may or may not want. And if they get the product slightly (or completely) wrong the first time, the advance or investment money is mostly (or entirely!) gone and there’s not enough money left to buy the time to iterate enough times to actually produce a product that people want. Historically it has been extremely difficult to iterate on books: the only iterations are second and subsequent editions, and those editions only get produced for successful books. It has been very difficult to iterate on an unfinished or unsuccessful book, in a way that distributed these iterations to its readers. (Full disclosure: my company Ruboss is producing a product called Leanpub that makes it much easier to sell and iterate on unfinished books. So I’m completely biased in thinking that this is a real problem, and a problem worth solving, since I’m hawking a solution.)

Now that we’ve considered the parallels in the problems of writing a book and doing a startup, we can start looking at the parallels in the solutions.

Solution: Use Lean Startup Principles

One movement which has spread over the past year throughout the startup community is Eric Ries’s concept of the Lean Startup. This concept originated with applying Steve Blank’s concept of Customer Development to the problems of creating web 2.0 startups like Eric Ries’s IMVU, of which he was the CTO and co-founder. Briefly, the concept of a Lean Startup involves launching the product extremely early (so early that you’re embarrassed by it) and iterating rapidly in response to customer feedback. These iterations follow the OODA loop (Observe, Orient, Decide and Act) of USAF Colonel John Boyd, applied to consultation with customers using a process of Customer Development. Both Eric Ries’s blog and Steve Blank’s blog and book The Four Steps to the Epiphany are essential reading for anyone interested in Lean Startups (or Lean Publishing), so I’m not going to summarize their thoughts more here. Instead, I’m going to urge you to read their blogs and Steve Blank’s book.

The Lean Startup approach of doing Customer Development and getting meaningful feedback from early customers is very similar to the process of releasing a book very early in the writing process and getting meaningful feedback from readers. In both cases, if the Lean approach is done well, the startup or book will evolve in ways that are not necessarily planned from the outset. (This is similar to the concept of Agile software development, but the scope is actually broader in that the entire direction of the startup or book can change.) This isn’t armchair speculation for me: I have written 2 books (Flexible Rails was written using a Lean approach, while Hello! Flex 4 was written using a traditional approach) and I have been part of 2 startups before founding my own Lean startup. Our Leanpub product is very much evolving in response to early customers.

Continuing an analysis of the Lean approach, both startups and books used to have a much higher barrier to entry than today. One aspect of this barrier to entry was high capital requirements and lack of access to distribution channels for individuals. This led to much of the infrastructure which is less necessary today. You can’t launch a hardware startup without outside investment (unless you’re a multi-millionaire), but you can launch a web 2.0 startup with a few thousand dollars for the lawyers and a few weeks of work. Similarly, in the past, you couldn’t publish a book and get it in distribution channels without a publisher, but now anyone can write a book and stick it on sites like Leanpub (which is designed for selling in-progress ebooks) or Lulu (which is designed for selling print books).

Since you can sell in-progress books today using sites like Leanpub, you can do the “launch fast and iterate” approach recommended by Eric Ries for building a Lean Startup, but for a book. I did this the hard way with my book Flexible Rails, selling it on Lulu as an unfinished PDF–and getting in the top 100 books all-time by revenue–using my own techniques which I evolved over 18 months. These techniques were remarkably similar to Steve Blank’s process of Customer Development, which is why when I encountered them and Eric Ries’s Lean Startup concepts I realized the parallels and my company created Leanpub to embody them.

The techniques I created to sell an unfinished book on Lulu evolved out of necessity, since the process of selling an unfinished book on Lulu was very cumbersome. For example, Lulu does not tell you who bought your book. So, you have no way of knowing who your customers are! So if you don’t know who your customers are, how do you give them free updates to your book?

How did I use Lulu to do the Customer Development with Flexible Rails? How did I build a community using freely available tools? How did I sell an unfinished book and distribute updates to people whose identities I could not confirm? These topics are the subject of my next post…

Note: This principle was called “A Book is a Lean Startup” in my top-ten list I launched this blog with. However, I’m evolving my thinking as I write the principles, so I’m renaming it here. My guess is I’ll probably rename some of the other principles too…