One of my favorite interviews with Masayoshi Son on the start of Softbank:
How did you actually start your company?
I started Softbank in 1981, a year and a half after I came back from the United States, after graduating from Berkeley. I wanted to start my own company when I came back to Japan. I thought of 40 different businesses I could start. It was like thinking of an invention. As a student, I had a hobby of inventing new ideas for products. For me, thinking of new businesses is like inventing new products.
For a year and a half, I did research and made business plans. While I prepared, I had no income. I spent money, I had a new baby. My wife was worried. All my friends, my father, my mother, everybody was worried. They asked me, what are you going to do? You spent years studying in the United States, and now you aren’t doing anything. I spent all my time just thinking and thinking, studying what to do. I went to the library and bookstores. I bought books, I read all kinds of materials to prepare for what I would do for the next 50 years.
I came up with 40 new business ideas—everything from creating software to setting up hospital chains, since my wife’s father is a doctor and has a hospital. Then I had about 25 success measures that I used to decide which idea to pursue. One success measure was that I should fall in love with a particular business for the next 50 years at least. Very often, people get excited for the first few years, and then, after they see the reality, they get tired of the business. I wanted to choose one that I would feel more and more excited about as the years passed.
Another factor was that the business should be unique. That was very important to me. I didn’t want anyone else doing exactly the same thing. A third was that within 10 years I wanted to be number one in that particular business, at least in Japan. And I wanted to pick a business where the business category itself would be growing for the next 30 to 50 years. I didn’t want to choose a sinking ship.
I had all those measurements, about 25 in all, and 40 new ideas. I took a big sheet of paper, and I drew a matrix and put down scores and comments for each. Then I picked the best one, which turned out to be the personal computer software business. That was the start of Softbank.
In Andy Grove’s book Only the Paranoid Survive, he begins by describing the strategic inflection point the 10x change facing Intel when the Japanese entered the memory market with superior cost advantages, destroying Intel’s core business and forcing them to accelerate development of processors – 8008, Pentium, Xeon, Core, etc. – which they are now widely known for today, having survived the market shift and a direct attack.
As Grove describes it, this was part of a broader transformation from a vertically-integrated computer industry to a market with horizontal segmentation, with specialization in each layer, which enabled a host of new companies to enter the market destroying incumbents like DEC, Wang and Univac.
During the period from 1980 to 2000, Grove’s observations were realized and became generally accepted. Horizontal specialization was a superior strategy that paid off in spades.
Dell with lean inventory management, superior logistics and sales & marketing, and Microsoft with Windows 95 / 98 and Microsoft Office were able to divide up IBM’s empire.
Apple – at least during this period – was in a state of decline having persisted with its vertically integrated strategy – albeit with very poor management and execution i.e. refusing to license Apple OS, blindly diversifying into printers & peripherals, building vertical desktop software e.g. Powerschool, FilemakerPro, etc.
But to go back to the beginning. The shift from a vertically integrated market to a horizontal market always began with the commoditization of one layer first.
In the computer industry, that was the Japanese entry into the memory market, which enabled Compaq and Dell to enter the PC market, which then enabled Microsoft to have a diverse enough customer base (besides IBM) to sell MS-DOS, then Windows 3.1 and eventually Office as Microsoft moved up market.
There was a great deal of path dependency, one layer had to be in place first (e.g. Operating Systems) for the next layer to shift (e.g. Microsoft Office, Adobe Photoshop).
That same shift is occurring now in the ERP market – or more broadly business systems. Accounting systems are the equivalent of chips above, and Xero is the Japanese entrant attacking Intuit.
Keybridge is our operating system built on top of Xero to provide a rapidly implementable, turnkey business system with an API to enable custom vertical development.
What do we mean by custom vertical development?
In the old days, if you wanted to build vertical software for any business or industry. You would always start with accounting and CRM, since every organization whether for-profit or not, would require a customer list, P&L, balance sheet, etc. to manage their business.
This resulted in wasteful R&D since every large vertical software company would be duplicating development of their own accounting and contact management layer. All of them mediocre and with a 1990s design and style.
By the time R&D got done with accounting, they didn’t have the energy left to polish their vertical functions to perfection. This is why so many enterprise applications look like CRUD because too much was done with too little R&D resources and because everyone stopped giving a damn about their product.
The Highway to Hell:
Yet these vertically-integrated businesses could survive commercially (despite charging egregiously high rates) in part because the vertical solutions were highly complex, hard to build, hard to sell, and because niches were tiny markets that were not subject to intense competition.
Market entry was expensive and if someone had 70%+ of the market, you had better be damn sure that you could beat them faster, than you could go bankrupt.
Keybridge changes everything.
It enables you to build on top of a strong base of generic functions: accounting, billing & invoicing, contact & relationship management, payment gateways and analytics, so that you can get straight into building the business logic for your vertical. Helping a property manager organize thousands of tenants, helping an auto dealer track their test drives, etc.
As Isaac Newton once said: If I have seen further it is by standing on the shoulders of giants.
Keybridge is your giant. It is our second step in accelerating the market shift into the CRM and vertical software layer.
What if accounting and CRM wasn’t internalized in the vertical software?
At any significant business volume e.g. > 500 customers / transactions per year, this would result in a violation of single-source of truth data principle.
Someone in the organization would be doing a lot of manual labor copy and pasting information between systems with a lot of errors.
Any organization with a focus on efficiency would consider this a broken workflow.
Certainly no business workflow is perfect, but brute force manual labor is a real, tried and true internalized cost.
Unless the margins are fantastic to support high levels of waste, this person’s time (and mistakes) are likely costing the business real money.
What about hooking up integrations like Zapier and connecting stand-alone CRMs with Xero, etc.?
While these integrations may work for smaller companies. This is akin to Apple vs. Microsoft with the choice between true and complete integration vs. patchwork integration.
Patchwork integration is a guarantee that things will break at some point in time. Do you want to build your business on stilts? Oops, Highrise just changed their API – sorry your contacts over the past week were not synced. Well now you and your team have a big mess to cleanup… The mistakes and bugs are not worth the business cost.
End users (i.e. your team) ultimately prefer consistency, they do not want to learn how to use another application or silo their work functions in 4-5 places – if they can use one system with a clean consistent interface that gets the job done, this is preferable and ultimately more stable, robust and cost effective for the business.
Apple’s approach was to control every aspect of the product. Our approach is similar, you cannot change the base interface of Keybridge, you must use our UI kit and controls. You cannot pick random colors like pink, etc. because these have no place in a real business.
While these rules may seem draconian, you can think of it like planning guidelines in the city. If you are a building a skyscraper, we are not going to let some guy build a shack next to you that looks like it should be demolished.
Furthermore, since we eat our own dog food and use Keybridge internally in our own business, you can be confident that the system & API will be maintained. The roads will be paved, flowers will be planted and that the general infrastructure will be supported.
I will be the first to admit that I am biased (heavily) on this topic, but I view consulting whether for development or in a product manager’s role as an intermediate and temporary step.
As a transient stage that reflects an earlier level of professional development and awareness about what is important in life and what is not.
I say this broadly because most senior developers and designers make the transition from consulting individually, then in groups, and then into building products, but almost no one goes in reverse from building products back to consulting.
There are certainly exceptions, security consultants perform an essential service to society by lending their expertise to help organizations secure their systems, and the more consulting work they do, the better the world is because their knowledge scales as it is applied. Similarly experts performing code reviews provide a quality assurance mechanism.
The distinction here is that consultants being hired to build products are like mercenaries being hired to fight a war, whereas security consulting & systems optimization involves transfer of expertise that can be compressed in a short amount of time.
How do the two worlds compare:
1. Building complete products takes a long time (years vs. weeks) and has no fixed end date. Good products last forever and provide infinite utility.
Consulting engagements have a fixed expiration e.g. 4 weeks, which is not a suitable timeframe for building products.
2. Responsibility and empathy for the customer’s problems is essential to building good products.
The results have little to no impact on consultants personally (or financially), so there is rarely empathy.
3. People building products have ownership. Their work and contributions are like those of an engineer and architect showing off their building with pride. They can point to it as tangible proof of their contribution.
Consultants may point to code that they have contributed, but there is rarely (in my experience) if ever passion, pride or excitement. Why?
4. People building products understand sacrifice and have a high level of pain tolerance. They are stronger and better for it.
Consultants generally have boxed lines that they do not cross i.e. oh, it’s 5 pm, I am going home, sorry.
5. Products require responsibility, accountability and alignment of success with the customer.
Consultants do not have to be responsible, they just have to log their hours accurately enough to get paid.
6. Building products is like marriage. There must be true love and patience.
Consulting is like genetic engineering. It will produce mutations, Frankensteins, and defects.
7. Product developers, designers and product managers all have significant freedom and control to shape the product’s direction.
Consultants have very little control over the general direction of their work. There is an illusion of control.
There was a long 4-hour TV mini-series Carlos: The Jackal that was produced several years ago.
After watching it, I went into a deep reflection because Carlos was a terrorist-for-hire with no particular goal – just a leftist bent, who committed several atrocities over the course of two decades: OPEC hostage crisis, killing Israelis at the Olympics, etc.
Carlos would hop from country to country at the bidding of the PLO, Colonel Qaddafi in Libya, etc. but each attack Carlos carried out involved a burst of activity, the same rapid flight & departure, and getting paid from his financial sponsor / master at the end of the process with long bouts of intoxication in between each of his attacks.
It might have been that I had been running around too much with jet-lag, but my conclusion from the film was that here was a guy – clearly very resourceful and intelligent – who had gotten mixed up and confused in the wrong causes, who was starting so many tiny battles, but never actually winning a war – not one permanent victory in his entire life.
And I wondered that must be how life is for a consultant whose projects never see production, who have no dreams and who are just doing their work to get paid – a lot of jobs without permanent victory.
It is sad to me, but this is a process that is repeated daily on ODesk.
When you are building products, you may face death like Dr. Braz in the film The Core:
But you can die with honor like Braz knowing that what you have dreamed and imagined has become reality built with your blood:
“Look, for twenty damn years I’ve done nothing but Virgil. Twenty years. Virgil belongs to me, and I will not let her fail. I will not. Now, if you want to know what’s worth dying for–this ship. Building it, instead of imagining it. If Virgil needs blood, it will be my blood.”