Learning from ARM

Long time no blogging, right? Here is the new stuff. I am writing in the conference room at Goldman Sachs Private Internet Conference at Bellagio Hotel in Las Vegas.

I just finished a great talk with Tom Lantzsch, from ARM. It was a 1:1 – that type of session you register and spend about 30 minutes with people you want to meet alone. I regret a little bit to request that session when it is approaching because I suddenly realized that I don’t have any intersection between my business and chip design business, and I worried that it may end up with nothing essential and waste both party’s time, and it turned out to be exactly the opposite. Here is a brief note about what we can learn from ARM.

1. Long Term

ARM is a 22 year company. In its first 20 years, 20 billion chips was using ARM architect, and they got 6 billion on the 19th years, and about 10 billion after that. It was then ARM become widely known. They charge license fee, and royalty fee, and in the first many many years, they only run the company to be covered by the immediate license fee. Royalty fee is the big part (0.07 USD per chip) but that came much later. This is not common even in semi-conductor field, and their success is really postponed by focusing on long term.

2. Keep Small

ARM is a 17 billion USD company now. However, it only has 2000 full time employees. They intentionally keep the company small.

3. Run it Cheaply.

ARM is about cheap. They build a culture of cost. As executive committee members, Tom still travels on economic class, and most of ARM’s clients sitting in the section before in flights. Although it does not impact the company financially for CEO to fly business, it does affect the culture. We talked about Google, and talked about the expansion in people, and the luxurious travel package. We all agreed that it is one way road, and cannot come back. ARM charges 0.07 USD per chip (this includes the more expensive chips), and that is the core of its business.

4. Run it Different.

ARM is not a design firm. They feel lucky that several of their principles worked so well in the last 20 years. Thanks to the slow advancement of battery technology, ARM’s core competence around low power consumption chips are stronger over time. I assumed a senario that battery life goes 10x, whatever ARM does not matter that much. Tom agreed.

ARM is also about partnership. Since they don’t design chips, they set standard, and create the architect, so they have few thousand partners, and they used that as a big network effect component: More partners on the equipment side, the more partners on the chip designer side. They played it in a really big and long term way.

Inspiration for me

I was impressed. Many great companies share some common ground. That is very different from the majority. “Luke 13:24: Strive to enter the strait gate”. Follow what most companies are doing and suggesting, and you are doomed.

By talking with great leaders, I can understand how to keep the peace inside, and keep doing what we believe to be true.

Different View on New Technology Adoption

In Silicon Valley, obviously there are a lot different views on technology. Although generally they are on the same direction, the minor difference are huge enough. I had conversation with three different people, and their view comes from aggressive, neutral and conservative. Here is the story (without revealing the name, while sorry for not giving the credit).

New Technology! Always! As Fast As You Can!

I visited a startup, and they are not officially launching their product yet. They are using the latest technology. Here are some:

  • Meteor.com
  • Underscore.js
  • Backbone.js
  • MongoDB
  • Node.js
  • Python & Ruby

Although some of them are not that new, they claim PHP is dead many years ago, and don’t tell interviewer that you are using PHP. MySQL is dead…

They claim that most silicon valley startsup are not afraid of leveraging the latest technology. And they even doing their best to keep up with the latest builds of the technology. They update the servers every several days to patch it and upgrade it to the latest version, with the risk of something broken. If it does, fix it. They found out that consistent integration is easier to do than holding for a long time.

With the new technology, every is fast. Most of the frameworks claim to do what people are doing at fraction of the time. They does!

Technology is the Layer in the Stack that does not Matter That Mach

I had another conversation with a technical CEO who just sold out his company at great valuation. He mentioned that 5-10 years ago, the core of a technical startup is technology. There is a clear line about what can be done, and what can’t. For example, the variety of drivers of Microsoft Windows system, and the file system Google built. They have to build their own servers, their own storage, and their own almost everything to handle the challenges of the huge volume of traffic and data. The 5 mil-second counts.

Currently, however, shift to a new model. The technology on Internet has developed that storage, and coding is not the hard part. Technology is more hybrided with art. Just as Apple demonstrated, a lot of new startup CEOs are not technical background but can control the messaging, the vision, the valuation proposition so well. Instagram and Foursquare are those type of company.

Technical company is so strong in technical and they can build everything so fast, and they don’t need to think that carefully before what to built. The not-so-technical company has the constrain that they have to think more carefully and decide what to give their users. This value is higher in the value stack than technical value, just like the lightness of an electronic bulb is not that important to the massive audience as the Addison time.

Be Careful of Engineers who uses New Technology

Another senior technology guy suggested this. He said that most of the startups failed because of leveraging new technology to quickly, the most recent example being Digg, by jumping into Cassandra too quickly to kill the company.

He claimed that every framework is promising to look at surface, but it needs a lot time to validate the feasibility. There are HIVE, and PIG camp before, and Facebook and Yahoo! were at the back. Now, PIG is gone, and a whole batch of companies, and engineers were left there with a lot of codes that does not work.

Their philosophy is to choose a technology that is at least used by a bigger company for at least one year to adapt to it. There is a consistent technology framework in the company that everyone use, and not many. Any new technology adapted at production server needs to be reviewed by himself.


I believe it is completely another Horse Crossing River problem. The three point of view are all valid. They are just the pro and con of new technologies. I would say, the companies in China needs more mentality in the first bucket, not the third at the current stage. After technology is over-used, we need to remind us for the third one.


Spam Enters China

I am not talking about email spam, or comment spam. I am talking about Spam from Hormel – the lunch meat.

I am not sure if it has been in China market for a while or not. I just saw it on the table of my dining room few days ago.

I am happy that people are rich enough not to have to each spam every day.

Big Surprise: How Much Apple Product I Bought

One of the biggest surprise in 2011 was how many Apple product I bought. I had never imagined that before. Here is a list:

1. iMac as the main desktop computer in our reading room.

2. MacBook Air 13′ as my main laptop.

3. MacBook Air 11′ as Wendy’s main laptop

4. iPhone 4 for my main phone

5. iPhone 4 for Wendy’s main phone

6. iPad first generation for me

7. iPad 2 for Wendy

8. Time Capsule as main wireless router, and main NAS

Apple has a much higher revenue / employee in IT industry.

P.S. Corrected my typo – should be MacBook Air, not Pro.

P.S.2 Another surprise. I now have Apple friends, and have many friends joining Apple, in Beijing and Shanghai.

Pretotype – Inspired by Alberto Savoia

Yutong sent me the prototype book by Alberto Savoia. In the first page I saw this quote:

Reid Hoffman, founder of LinkedIn once said: “If you are not embar- rassed by the first version of your product, you’ve launched too late.”

Very nicely said. Thanks Yutong.

Obviously, anything that sounds right must be put into context, and the tough part is, the author often ignore that part, and only tell you the conclusion. Just as the “Little Horse Crossing the River” story, the Ox, and the squirrel told completely different idea but they forgot to tell the horse that the depth of the river was relative to THEIR heights.

Let me put a little context to what Reid Hoffman said.

That means, it has to be 1) very easy to release. 2) very easy to change, 3) free to try out.

You cannot use the philosophy on hardware product (Well, I am still thinking about the iPhone 1.0 case), and you cannot use it on something too expensive to try. Image you asked your user to buy a $1000 product that is really not ready, and you expect the customers to buy it again few months later. (Well, iPhone 1.0 case?)

I am still thinking about the other context Reid Hoffman didn’t tell in the simple sentence.

Here is the link:

Video Like: http://vimeo.com/19443634

Slide: http://www.slideshare.net/LuisFelipeFernandes/innovation-at-google

Try to be a Good CEO

Whenever I felt frustrated, and felt there were so many balls falling and so little time, my friend just smiled and told me: “Welcome to the world of a CEO”.

Yes. That is the experience to be a CEO. As an entrepreneur, you are facing the challenge that only the people who were in that position can understand – the loneliness, and the excitement – oh, boy, nothing in this world is comparable to that. I show my full respect to every CEO in this world – even my competitors, even former Apple CEO, former Yahoo! CEO, and former whatever company CEO who ended their career in a way of public failure – they have all the due respects, not to mention the greatest CEOs of this century – Steve Jobs, Bill Gates, and Jack Welch. We are just the same kind.

Few weeks ago, I had breakfast with Sun Cheng-Yaw, and Jack Gao (one was the former HP Greater China GM, and the other was former Microsoft China GM). They shared their experience of running a company as the real CEO and a big country subsidiary. Basically the difference is, you know you can screw up a company single-handedly while screwing up China does not kill HP or Microsoft. Pressure really comes from responsibility.

I am just trying to be a better CEO – a journey that needs a lot of work, and needs a lot help. The most important thing is to surround yourself with the right group of people, and share the same dream. That is so important, and I will name it as the third learning in my startup, besides the two I already shared with Stanford students few weeks ago (focus and cost).

Soul Searching – The Deep Trace of the Reason

The term soul searching comes from the book The Facebook Effect, on page 180 of the chapter 9, 2006.

The astonishing success of Facebook’s photos application led to a bout of soul-searching at the company. What was it, Zuckerberg and his colleagues asked themselves, that made photos so successful?

Soul Searching

Soul searching means the deep trace of the reason why something worked. It is easy to be happy about a great feature, and a successful campaign, but it is way to easy to just stop tracing the deeper reason of the product. Just like the photos application of Facebook. It is a simple application without most of the features other photo sites have, but it is soon becoming the most successful photo application on the Internet. What is the driver for that? Why is that? Why, Why and Why?

With the deep trace of the whys, they learnt the secret of social graph, and how application can be built on top of social graph, and that leads to the iconic Facebook homepage – the News Feed (a feature that was originally called Timesorting).

That is what soul searching is about.

Asking more Why’s

If something happens, and it is a good one, don’t let it go. Push ourselves to do a deep soul searching and understand the deeper reason behind it. Thank God we are in Internet space, and we have all the data needed to understand the reason. Just like Facebook can dig into the data and understand every photo change leads to 25 new page views, there must be some link between the reason and the result. Go get it.

10 Years Later

So many things dramatically changes over 10 years, but it happens so gradually that we didn’t pay any attention to it when it happens. Here are some examples.

  • The IT support in my office was gray haired. When I just started my career, IT support is a profession of young people. Now, when the outsourced IT person appeared in my office, I realized that he is also in his 30-40s. That changed.
  • Credit card processing is so quick. Now, I don’t bother to use cash since most of the time, the credit card processing is as quick as the same time as swiping card – it is supposed to work that way, but when I just started to use credit card, it easily takes few minutes, waiting for the printing machine to start printing. Sometimes, I have to make some phone calls to the bank and let the bank call center person to teach the cashier how to swipe the card.
  • The business meal in the Xujiahui area gradually raised to a level higher than Hong Kong, and most places in U.S. The set dish has reached to 58 RMB (8-9 USD) or higher. The 10 RMB box meal completely disappeared from this area.
  • The CPI has raised to a level that people in Shenzhen started to buy home supplies in Hong Kong, and more people discuss about pork price these days than 10 years ago.

10 Years. Many things changed.

People Matters

People are always the core of any company. Enough focus on its people is the essential to the success.

Keep the talent density high enough is to the key.

Be sensitive to the environment and act quick enough.

Always find the right people to do the right thing.

How Startups Enter China?

My friend asked me a question:

It is stupid to ignore China market for any startup. What is your suggest for startups to enter China market?

My quick answer is: Don’t.

My alternative suggestion is, enter China market by putting your headquarter in China.

I have many successful entrepreneur friends in the Silicon Valley, who don’t have China presence yet (and I suggested them not to do it), and I also know some great American in Shanghai who have amazing business in China. In contrary, I also know many of my local friends who were laid off after working for the Shanghai branch offices for startups in US, or Europe.

Why is that?

Startups are constrained by resources. With resources, I mean in terms of people and money. It is NOT constrained by market size. Either US or China is big enough, way to big for a startup to explore. Concentrate in one market, and setup the model before expansion.

The more important question to ask is, whether you want to start up the company in US, or in China.

Integrity of Resume

Today, a candidate came to interview at our company. The experience, and the answers to questions looks very promising, but there is one problem: he intentionally changed the date of his employment history by few years. We immediately terminated the interview when we found out.

Integrity is the key to any person, and company. It is by trust that this society works together. No matter how good a person is, if there is integrity problem, that is a big problem. In Microsoft, for example, integrity issue is the biggest mistake someone can make, and will result in immediate termination of employment. The first value of the seven key values of Microsoft is Integrity and Honesty. At least in the few years I was in Microsoft, it was really enforced. I know it is a big deal.

So, don’t play trick with your resume, and don’t play tricks in most of the things.

Xindanwei is Cool

I was in Xindanwei this afternoon to chat with @isaac, @popoever, and @aaajiao.

I heard about Xindanwei from time to time, but never related it to Isaac – who is an angle investor to this idea.

The idea is simple: you pay as low as 15 RMB per hour to get an office space there – by hour, or by month, with nice meeting room.

I put it this way: It is a real estate company + a property management company. Well. Isaac put it different: it is a place where people share space, share connection and share knowledge… I admit his version looks nicer.

I will be at Xindanwei from 16:30 – 18:30 this Friday to attend the Friday Chitchat with topic: Social Media and Me by @popoever. If you are also interested, see you there.

Xindanwei is a nice experiment to leverage social media to run a pretty traditional (but obviously something brand new) business. Anyone who visited there become a “big mouth” to broadcast the message. I am acting as a promoter of Xindanwei there. Enjoy your time in Xindanwei.

P.S. They offer 1 hour of free office space (with Wifi) there.

Bet is Greatest Tool to Increase Visability

I just made a bet with college in my company. The bet is like this:

The loser will travel to a city in the middle of Hunan province named Loudi (Loudi on Wikipedia), and take a photo before the Loudi Train Station at the end of this month.

Here is the transition plan:

K859 4:54pm – 9:08AM Shanghai South Railway Station – Loudi

K80 – 2:02pm – 4:53am Loudi – Shanghai South Railway Station

This is an interesting practice. We all have goals. Most of the goals are easier to accomplish with the peer pressure. Run Liu put it this way: “The best way to accomplish something (like passing an exam, or losing weight) is to tell all your friends that you are going to do it.” By increase invisibility, you setup some pressure for you that you cannot remove by yourself. The only way out is to work hard on it.

A bet is like this. It draws excitement for everyone involved, and a great way to show case that you are serious about it.


Hanting, the chain hotel group will go IPO tomorrow (March 26, 2010). The ticker will be HTHT – China Lodging Group.

I am a frequent Hanting customer. I know their style. It is a very cost concious and cost effective company. My gut feeling tells me that they should be doing well after IPO.

I will wait and see the final result, and do some reflection after comparing the final results with my thoughts, and learn more about how to valuate a company better.

Disclaimer: This is not an investment advice, and I don’t take any responsibility related to this article.

In Market We Trust – Part II

xge made a very good point under my entry In Market We Trust. That is exactly what I am trying to say about the market – the difference is, xge just gave better examples. Let me continue the argument here.


The government decision makers are very likely to make decisions based on intuition, and the planned economy mindset. I, myself, made even more frequent mistakes when I make decisions about the market. I admit that I did several things wrong in my business without respecting the market laws. Here are some examples about the actions people take (basically using xge’s examples).

Hainan’s Land

The recent sharp increase in the real estate price in Hainan was a typical example. When the Hainan announced the “International Tourism Island” strategy, the house price almost doubled over night by the stimulate.

The Hainan government was scared because of the jump in price. Then, in order to control the price, they immediately announced to pause supply of all commercial residential lands. Not surprisingly, the price doubled over night again.

Train Tickets and Taxi

Train tickets in Spring Festival times are extremely hard to get because of the huge demand (2.5 billion people/time needs to travel in the 40 days), and limited supply (not enough trains). The price for the train tickets went as high as 4x of the original price in the escort market.

In order to keep the price low, the government requires all train tickets keep the cheap price, and crack down all the escort market, and implement Real Name system for train tickets – none of them touch the supply and demand. Not surprisingly, it is still hard to find a ticket.

For taxi, people complain that the taxi is expensive, and hard to hire in rush hours, so “illegal taxi” started to emerge to help them out. In order to keep the price of taxi low, the government tried everything they can to crackdown the illegal taxis, greatly reduced the supply. Not surprisingly, taxi is harder to get, and the government enforced low price (lower than fair market value) keeps taxi drivers out of the market. (Well. for the taxi business, the real problem is, the monopoly of government owned taxi companies get too high fee out of the pocket of taxi drivers, and use the power to keep private sector entering this market).

Supply and Demand

Finally, we started to understand the simple market rule: price is determined by the supply and demand. The only way to drive price down is to increase supply, and the only way to drive price up is either to decrease supply or increase demand.

In Market We Trust

The main take away from my Silicon Valley trip is the understanding of “market“. I know, the recent financial crisis can be used as a way to argue the other way, I am still talking about the relative older style of market (like the flea market, and fish market), not the financial innovation stuff.

Hotwire and Priceline

Websites like Hotwire.com and Priceline.com (how I found it) let me think more about how to use the market to allocate resource more efficiently. Without services like Hotwire, hotel suffers from empty rooms, and visitors suffers from high price. They are the bridge to allow those unsold rooms to be evaluated and priced – the price is lower than retail price, which is very likely to be the fair market value of the “unsolde” rooms.

“Market allow any tradable item to be evaluated and priced.”

Rental Contract

Another interesting thing is the rental contract I saw in my friend’s house. When their last lease period is going to expire, the landlord gives them a list of renew price. It is something like this (I mocked up the numbers)

1 month – $1710

2 months – $1700

3 months – $1670

4 months – $1640

5 months – $1610

6 months – $1690 <– Attention! Price raises!

7 months – $1710

8 months – $1720

9 months – $1840

10 months – $1700

11 months – $1640

12 months – $1600

The idea is, it is not the longer you lease, the cheaper it is. It is obvious that the 9 months lease is the most expensive lease. My guess is, it is the time most of the tenant in this area planned to move out (based on the existing contract and prediction), or the season where many graduate are moving in. The price of the house is decided preciously in a bid-offer fashion, not by some ideology (the longer you rent, the cheaper it is)

That is a vivid lesson for me to understand the market.


A market is the PLACE for buyers and sellers to trade.

Its function include pricing – market finds out the fair market value of any tradable item every minute.

The liquidity of the market is important. Liquidity means a seller can always sell something if they are willing to decrease the price a little bit above the fair market value, or the buyer can always buy something if they are willing to offer a little bit higher price than the fair market value.

“The essential characteristic of a liquid market is that there are ready and willing buyers and sellers at all times” – Wikipedia on Market Liquidity.

The market rule is only the price and item – for the same item, it is just the price.

The only way to decrease the price of an item is to increase supply.

The God of Market

It seems I need to spend more time to understand how the God of Market works, and its role in price discovery, and resource allocation.

China is a market economy, but in daily life, it is still not very market driven.

My Advice to Entrepreneurs

Occasionally, friends ask me for advices of their startup. I am flattered but shy away from giving any advice, since I just started the journey about 5 years ago myself, and I have more lessons from mistakes than tips for success to share. Now, let me give it a try (shamelessly pretending that I know this topic).

Try to Use Bad Words to Describe Your Business

It is not rare to see people describe their products, or services using many good words – the best, the cheapest, the highest in quality, or the newest… In my opinion, the chance to success with any thing more than one good word is small. Let me give you an example.

A friend of mine just started a business few months ago. He told me that they were going to deliver better service than CTRIP, and the ticket price is lower.

I could imagine a business to have worse service, but cost less, or a service with better service, and cost more. There are good opportunity to win for either model because customers’ needs are diversified, and differentiation in position can be a winning strategy. Some people prefer service (and willing to pay more), and some just care more about cost.

But for a startup with very limited resource, to optimize on two or more dimensions does not seem wise. Li Song, CEO of Zhenai.com, put it this way: “We can assume that our IQ is above average (which is often not true), but we can never assume that we are genius.” I would add that we also need to assume the competitors are at least above average, not idiots. In my friend vs CTRIP example, it is not safe to bet that CTRIP guys are stupid and everyone who have an idea can beat them in BOTH service and price.

If I hear people describe their business with one good word, but many bad words, that seems to be a winning pattern for me. For example,

“I am going to build the fastest websites, although it is ugly, lack of function, and cost more than competitor.”

I understand he/she is really serious about site speed. What do you feel if he says:

“We are going to build the fastest sites, with much more functions than any competitors, and have the best visual design, and meanwhile, it cost just 1/2 of other sites…”

If I hear that, I would comment: “You don’t know what you are building yet”.

There are many good things in this world, but you can only pick one or two at most at the same time.

A Business Model Fits Revenue and Cost

A friend of mine is doing an online content business by hiring people to translate and edit the contents. The cost is similar with an offline magazine, but revenue, even in the best case, is much lower than offline. The business model just does not fit into the current cost and revenue structure. There is no good or bad, right or wrong about how much the cost should be. The key is, the revenue has to fit to the scenario of the cost. Never create a business that cost more to deliver the same, or cost the same but deliver less, or even worth, to cost more and deliver less. At the end of the day, profit is always revenue minus cost.

Align Product and Business

Here is another type. They want to build a good function to attract users, and then leverage that to sell something else. The product and services they are building actually does not align with the business / money making efforts. If you visualize it as two arrows, they are pointing to different directions with 30 to 90 degree of angle in between.

Let me give you some examples. A friend told me that they are going to provide some useful tools like map and estimated price of real estate to attract users, then they can charge those people who list house classified on the site. It seems to me that they under estimate the difficulty to do either one really well. A real estate listing service itself is a full-time job of a large company, if you want to do it well, and house price estimation is at least equally demanding for focus, patience, and resources. In this highly competitive market, there are specialized company doing on either direction with 10x of resources. Any startup with more than one direction to go does not seem to be a good pattern to succeed.

There is Demand, Does not Mean You can Do it Well

I heard of another business model. It is like the previous “doing one thing, and monetize on the other” type. They said there are big demand on English version of map in China, and they want to build it first, and then deliver value added service (something irrelevant to map) to make money. The reason they started with map is, there is huge demand for it.

Well. There is demand does not mean that you can do it well, or there is a business around it. One extreme example is, there is huge demand to predict the up and downs of stock market, or there is even bigger demand to buy $100 bank note with $50. Everyone wants it, but there is no such a service because of a reason.

I am not discourage people to try something that no one did before, my point is, if you don’t know what makes you unique in this area, think harder.

Lego 3001 (Brick 2 x 4)

We started technical days with a great name – Lego Day. Yes. The Lego means the Denmark toy manufacture Lego.

We bought Lego 66284 Build and Play Value Pack that consists of 5573 and 5576.

Among all the bricks, obviously, LEGO 3001 is the star. A simple brick that is manufactured the most, and most easily found in every corner of this world. It is the spirit of create something and make it really good, instead of creating many different functions but none of them is as classic as Lego 3001 (I am talking about website design and almost all business).

Coins or Paper Money

We talked about the network effects of a real society many days ago. Hopefully, that explains why Chinese don’t use voice mail, calendar, or classified. We talked about it in YLF 2009 conference, and there are people who don’t agree with what I said.

Then I have another great example. When I was in Chengdu, no one uses coins. They always use the paper money of 1 RMB. Everyone hates coins. Why? Not for a rational reason. It is just because when you have a 1 RMB coin, you cannot spend it. People won’t accept it, and they ask you to change it to 1 RMB paper money. Just because you cannot spend it, you don’t accept it, and because you don’t accept it, others cannot spend it. The stronger the current feedback loop is, the harder it is to break.

On the contary, people in Shanghai LOVE coins. They don’t want paper money, although they unwillingly accept it. People will be surprised to get 7 coins as change to buy a 3 RMB ticket with 10 RMB note.

That is the network effect. Before debating about whether coin money is better than paper money or not, we need to look at where you are going to spend it, Chengdu, or Shanghai.

Ten Characteristics of Great Companies

Feld posted an article about Ten Characteristics of Great Companies. Nice article. Among the ten, if you ask me to pick three, here it is:

3) Great companies make lots of money but leave even more money on the table for their users and partners.

4) Great companies don’t look elsewhere for ideas. They develop their ideas internally and are copied by others.

6) Great companies are led by entrepreneurs who own a meaningful piece of the business. As such, they make decisions based on long term business needs and objectives not short term goals.

There are some hard lessons on this to run great company.