RFS.102: The Startup Stack

The Startup Stack does for starting a business what Heroku does for starting a cloud app.

JFDI invites startup founders to apply with ideas for building the startup stack.

This is one of a series of JFDI’s Requests for Startup. If this is the first you’ve read, you might like some context first.

Problem: It’s hard to start a business, for many reasons. Some are fundamental (you’ve made something nobody wants.) But others are mere annoyances that deserve to be engineered away.

Solution: make it possible, reliable, convenient, and cheap to do the parts of starting a business that have, until now, been hard – but needn’t be.

Example: a “launching soon” landing page: See launchrock.com.

Example: demand estimation: an easy way to gauge market demand for your startup idea.

Example: automated deal screening involving assessment of founder/idea fit, and other aspects of startup likelihood of success. JFDI is taking this on with the JFDI Startup Score.

Example: usability testing as described in Krug’s Rocket Surgery Made Easy. See usertesting.com.

Example: i18n and l10n. Many of these examples work best for Americans. Adapt the underlying models to work across Southeast Asia, across multiple languages, cultures, and currencies. Make it easy to port a software service across same.

Example: incorporation: In some jurisdictions (India) it takes months or years to register a business. In others (Delaware, Singapore), it takes hours or days. Why shouldn’t it take seconds or minutes? The domain registration process is pretty quick. Why shouldn’t incorporation be just as quick?

Example: legals: look at Clerky and LegalZoom. The world needs one-click, Delaware or Singapore incorporation, vesting plans, ESOP, term sheets, employment agreements, development of financial pro formas.

Example: fundraising: AngelList is making it easy for founders to meet funders. When they’re ready for term sheets, you can use the WSGR Term Sheet Generator or the Perkins Coie NVCA Term Sheet Wizard.

Example: payments. PayPal, Amazon Payments, Google Wallet. Stripe, Dwolla, Square.

Example: banking. Setting up a bank account takes several days, even in Singapore.

Example: corporate secretarial services. It takes days for a corporate secretary to generate paperwork around directors’ resolutions, AGMs, EGMs, and so on. These papers are fully understood and can be modeled as state transitions. I would like to see a startup service that fully models a business’s legal form and cap table in software, and spits out all the necessary paperwork in a fully automated fashion. “Just sign here”.

Example: finding a place to live and work. Apartment- and office-hunting is time-consuming and inefficient. Fix this.

Example: help with visas. While numerous visa experts stand ready to help you, all of those services involve high-touch human contact. Turn this into a high-volume API.

Example: outsource product development. There are many agile shops in Singapore and the surrounding region. Optimize the processes by which clients communicate briefs and take delivery of iterations.

Next, I’d like to see integration across these services, so that one can instantiate a new business with a minimum of human involvement, and have all the infrastructure services up and running, in the same way that one can integrate across the different offerings of the AWS suite through a scripted combination of command line API calls.

Right Motivation:

Scientists build in order to learn. Engineers learn in order to build. Hackers do both, all the time.

Suppose you’re a professor of computer science, and you’ve written a book about computer programming. But it’s the 1970s, during the phototypesetting interregnum following the decline of hot metal letterpress, so the publisher has sadly uglified your book. What do you do? If you’re Donald Knuth, you put the book to one side and you go off and pretty much invent digital typesetting. Then you use your invention (TeX) to produce your multi-volume magnum opus (The Art of Computer Programming). A hundred thousand people read the books. Millions use TeX.

Necessity really is the mother of invention.

Suppose you’re Richard Stallman and you’re not happy with the state of the Unix world. To write your essay about why you’re unhappy, you need a text editor. But you’re not happy with any of the text editors out there. So you write a new one (emacs) and then you write your manifesto. Emboldened by the manifesto, you leave the old Unix world behind and as you head into the brave GNU world, you naturally take the emacs source code with you. But you need a compiler to build emacs. So you sit down and write a C compiler (gcc) and a debugger (gdb) too. From scratch, naturally, and in emacs, naturally.

This is how seminal work is done: in answer to an organic need. The complete opposite of Schlep Blindness. I can think of no more Promethean act than to build whole layers of infrastructure beneath you, just so you can do what you actually want to do.

Aesthetic Insight:

We’re at a point in history where the parts of the stack that are most obviously wanting are not in hardware, not in software, but in business.

Trend and Premise:

Andreessen said: Software is eating the world. Our corner – startups – could use a nibble.

Whitehead said: Civilization advances by extending the number of important operations which we can perform without thinking about them. The infrastructure stack is how we organize those operations into modular layers. Every network engineer is familiar with the OSI stack. Twitter over HTTP over TCP over IP over 802.11.

With each advance the hard becomes easy, the impossible becomes possible, and the struggle moves to a new frontier. Hence Stephenson’s description of the technosphere: Underneath is technology that has already become free. Above is technology that has yet to be developed, or that is too crazy and speculative to be productized just yet.

But what is technology?

Brian Arthur says: technology becomes a complex of interactive processes – a complex of captured phenomena – supporting each other, using each other, “conversing” with each other, “calling” each other much as subroutines in computer programs call each other.

We’ve always thought of technology as software and hardware. We’re comfortable with that. We use it. We build it. We swim in it.

Hackers are now learning to think of business as a technology too.

The whole Lean Startup movement treats business as, basically, a branch of engineering – value engineering.

What is a business? A machine that produces value.

Seeing the world this way opens up the space of the adjacent possible. Growth hacking becomes a thing – even the thing.

Lots of businesses work as technologies. Consider UPS. They’re a service that encapsulates the complex of technologies involved in moving stuff from one place to another, reliably, conveniently, and affordably. Software can run API calls against UPS to insert a package into the system, track it, and get it back out. If that’s not a technology, I don’t know what is.

The more you rely on a technology, the more you treat it as infrastructure. Tap water. Electrical power. Roads, airports, ships. Law courts, arbitration, mediation. Safe food from public restaurants.

Today, it’s possible for an author to wake up, write a short story of 1000 words between breakfast and lunch, self-publish it over tea, and reach a million paying readers by dinner.

Tomorrow, I’d like to see a founding team explore their passions over breakfast, specify a new offering over lunch, get ten thousand signups by teatime, have the company incorporated by dinner with a bank account set up and all the paperwork signed, and spec an MVP for delivery overnight.

This is an RFS. Are you the right team to execute this concept? Stay tuned for details.