 |
 |
Monday
November 15, 2004
From: EDACafe
|
|
|
|
|
|
About This Issue
|
|
Interview With A CEO
|
|
November 8 - 12, 2004 By Dr. Jack Horgan
|
|
Read business product alliance news and analysis of weekly happenings
|
|
|
Introduction
The title is a takeoff on the Ann Rice novel and Tom Cruise movie “Interview with a Vampire”. Of course, most CEOs are not as scary as a vampire. Just to be on the safe side, I did the interview by phone. I did not want to ask a CEO the usual questions about the last quarter or guidance for the upcoming quarter or about the company's latest product release. I wanted to ask questions more about what it is like being a CEO. I was fortunate enough to get George Janac to cooperate in this experiment.
J. George Janac is the CEO and Founder of Silicon Navigator. A self professed serial entrepreneur, George also founded High Level Design Systems, Giga Scale IC and InTime Software. In addition he has been a director and investor in numerous EDA and chip startup companies.
George was founder, president and CEO of High Level Design Systems (HLDS), which was acquired by Cadence in 1996. He served as CTO of Deep Submicron Business Unit at Cadence Design Systems, post acquisition. George's chip design and EDA expertise dates back to SDA Systems, and Bell Laboratories in the early 1980's.
What is the role of a CEO?
My sort of strange definition is the person who at the end of the day is responsible for whatever happens to the company. It is his responsibility to fix it or make it right. The CEO has the P&L responsibility and the responsibility to make sure that the company succeeds. He is responsible to the investors and stakeholders in the company. At the operational level you are everything from the top cheerleader to the ultimate pessimist. You have to be optimistic but still make sure that everything that could go wrong doesn't go wrong.
You mentioned success. How do you define success?
At the end of the day people invest in startups looking for ROI. Success is building a product that leads to profitable company that returns money to its investors. Along with that the company produces something useful and is valued by its customers.
Role of the Board of Directors (BOD)?
They help you make decisions. One of the problems with these companies is that you have so many day to day activities going on that there are times when you need someone to say “You know, you really need to look at this way”. So really the Board of Directors oversees and helps people make the right decisions. The Board of Directors for many companies is very crucial. A lot of times especially with startups the boards are made up of investors who have different goals for the company. At the end of the day they are there to remind you of their goals and also help you to succeed. The corporate goal is to have investors that can help you with different areas. Someone an help with
technology, someone with sales, one director with detailed execution and another with global vision. Having the right mix of directors is really important.
How is the strategic direction of a company set?
In many ways with a startup the strategic directions is really set at its inception. Assumptions about the strategic direction are constantly tested against market and development reality. The direction is really with the original business plan. It is very unlikely that a company will start off in one direction and be successful in a different direction. It may be successful along a 30° or 40° arc from its initial conception but very unlikely if it goes in a completely different direction. By the way, if you are waffling on the direction in which you are going, you have a problem.
Sources of Innovation?
They really come in a couple of categories. In EDA ideas come from individuals. This is an industry whose great ideas come from individual people. That is a huge source of new ideas that come up; individuals who think up a new idea and are willing too put it on the line to make it a reality. The other source of ideas is the customers. Customers have problems, customers have needs. Very often you can figure out from the way they are working on or trying to solve a problem something which is applicable industry wide. Interactions with customers may either lead you to a new idea or very much help you fine tune the idea you will be successful on..
Role of Standards in the EDA industry?
Standards are progressively becoming more important. EDA standards very rarely come from a bunch of people getting together and writing a standard. I would say that Might makes Right. He who basically goes out and gets a standard adopted by a large number of customers then can get it adopted as an industry standard. You can look at Verilog which is now IEEE blessed. A standard always requires someone to push it. I think that this is a critical aspect of any standard. We are working with OpenAccess. Part of what we see is necessary is not just that Cadence and us but a lot of people are pushing it. The standard stuff is just becoming more and more important. Otherwise everyone needs to build everything from scratch.
How to company's make versus buy decisions?
You have to be incredibly disciplined because it's so easy to have the engineers choose the make decision. When you have a standard what you are trading off is the development time to build the code you know versus no development time for the unknown, the code you have to learn. So it is always easy to find some little flaw that says we should develop it. The difficulty comes from realizing the time, the money it's going to cost you. If you are in a startup, someone might say that this will take a man-year to develop this versus buying it. And you go that's not too bad. But I am really paying 15 people so if I wait a year I am really wasting 15 man-years of expense. The key is to be
realistic, pragmatic about the make decision. Look at the decision in the overall context of what you are trying to achieve. What kind of product and exit value.
Companies have only a certain amount of cash to get something done. It is really a question of how long it takes to establish a business. If you build everything, it will take a long time to get it right and build a business. There are some very good standards and people are suing them more and more.
There is some thought, some process, some algorithm that is the gem that makes a company unique. But 80% of the effort is not spent in building the algorithm but in building everything around it to make a product. That's the part where standards can help. If you can plug your algorithm into someone else's system, you have just saved yourself a pile of effort and time. You have made it better for the customers that can now use you along with somebody else.
Importance of Alliances?
Alliances are good things but in an alliance everyone has to gain or it doesn't work. They get published but don't lead to anything significant because there is not enough alignment, not enough gain for both parties. Align yourself with someone who has as much to gain from alignment as you do. It is very important for small companies to get alignment with the big companies because they have access to customers and support structure. On the other hand the little guy has a lot more to gain in the relation with the big guy. It is difficult to say that I'm going to be an island to myself, just something out there. You have to be able to interface and partner with people.
Major challenge facing a CEO?
This is a difficult climate to operate in EDA. Part of this is the economy; a lot of it is macroeconomic changes around the semiconductor industry. I think it is finding a niche that customers are willing to buy into and building that up into a high value position, while dealing with the fact that everyone is becoming very cost conscious. In many ways the mainstream flow is reasonably well set in terms of tools. There are no huge gaps in the design flow. It is really dealing with the whole economic situation to build your product and be successful with the customer.
Most Effective Marketing Tool?
The press is still very important but progressively on-line is becoming more and more important, especially in EDA the technical forum for information. I believe that more and more distribution will move on-line certainly for low end products. EDA customers are much more likely to purchase and have a relationship with you, if they believe in your technical credibility. They see the technical forum as being important: white papers, SMUG, EDACafe, technical articles, articles by customer.
How do you recruit and retain talent?
Engineering folks want to produce a product that is successful. They need to buy into the vision of the as company as being successful. That is probably the single biggest recruiting tool; people believing in a vision and wanting to participate. That's not to say that they will work for you for free or for no stock.
In this climate it is not particularly difficult to recruit even very good talent. Connections are the best way. I have some luxury of having been at this for 20+ years. I know half the people in the industry at this point.
Future of EDA?
The future of EDA is really tied to the evolution of semiconductor industry. People in the past looking at EDA saw chips as being the implementation platform. Where you do the backend is the difference between ASIC and customized tooling. I think we will be move moving progressively towards structured ASICs which have a low amount of recurring engineering. We are going to be looking at .18 micron fabs as being able to do analog and mixed signal at very low NRE. The high volume stuff will stay at 130 nm to 90 nm. So there is going to have to be specialization along the implementation flow. FPGAs will have to be dealt with by everyone. It is the vanilla digital ASIC implementation of
choice right now unless it can't meet you goals. The future of EDA is really tied to these varied implementation platforms and its ability to build solutions that target those effectively rather than slick Swiss army knives. If you look at a processing in the fab, you see these backed by slow leakage processes that people go through incredible pain to produce so that they can service handheld market which is completely different from those servicing the microprocessor market - they can attach a fan on the top. We will see specialization along these lines. If EDA can adjust to that then we will be okay.
Licensing models?
Time based license mode is here to stay. That is going to be the predominant business model. Some customers may try to negotiate back to the perpetual but TBL will be the dominant model. There is a lot more acceptance of low end tools being sold through the web without a sales force. Customers have gotten more comfortable buying their own stuff on-line and this will eventually move into the EDA world.
Design Outsourcing?
There will be a flow overseas. It will come in waves. The stuff that is well characterized, easy to understand and well communicated will move offshore just because of cost. All of the sophisticated design and architectural work will remain here. There will be many overseas choices as to where you take different products. Four to five years ago every startup chip company wanted to own the backend tools, do everything themselves, total control. Today, it's lets contract out this layout stuff. If you look the cost of the layout stuff it isn't labor cost, it is really tool cost. If the tool cost were to be at a reasonable price, there would be less incentive to outsource. There will be
structural changes with respect to who is going to be doing what. Some of those are driven not so much by cost as by risk. Today you can spend $1M on your backend system and if you hire the wrong person to run this, your chip will not work. That has to get a lot easier to do for the EDA industry to be able to sell its entire tool set to the entire design flow
If you look at the cost of place&route seats, DRC, extraction, LVS, timing and so forth for a 3-6 month project to layout a chip, the tools are a much higher price tag than the labor costs. EDA tools especially the backend tools are not an insignificant factor in a lot of these startups willing to spend the money. Tools are a significant factor with startups. It is much more efficient for someone who does this as a full time business who can keep those tools running a high percentage of the time and can amortize the cost of the tools over a much larger number of chips. This is why you see the growth in the e-Silicon business.
How do you spend your time?
I have been running companies for a long time, so I have developed a system to it. I get up very early in the morning and get to work at 5-6 AM. My goal is not to work too late into the day. I have a family and nobody cares if you are home for breakfast but they do care that you are home for dinner. I skew my hours earlier to deal with the family issue. I do technical work or what I call “content generation”, i.e. marketing, writing, specifications, until 9 AM. Six to nine is solid work. I spend the rest of the day in human interaction, working with customer and people in the company.
Geography?
The model is to establish in the US then go simultaneously attack a region in Europe and ASIA, usually ending up with distributors all over Asia and Europe. The key strategy when you go over seas is you need people that can efficiently run these operations, who are used to it otherwise you will be up 24 hours a day.
Micro versus macro management?
I believe in micro management at the beginning of the company because it is all about details. But as it grows you want to delegate and have people take over key pieces.
The founders have to establish the initial product and business. Once you have a formula it is much easier to evolve the product and to bring up a channel. At the beginning it is the people who really understand the product and the market, the founders, who really need to bear the brunt of establishing the initial wave.
How do startups succeed in a market dominant by giants?
The big three's revenue is largely derived from acquisition. Effectively you have to be able to have a unique idea, sell around them and either the business model becomes profitable enough so that you can go public (Verisity) or you have a highly valued point solution that needs to be incorporated into the big players' business and those companies basically get bought.
The key for startups, especially those doing replacement products is if you go into large accounts to sell a replacement product, it is very difficult top sell against the big guys because of the size and the discounting of the deal. The dominance of the three players really does effects how long it takes to proliferate against the big large corporate deal, what is called the fam or family of product deal.
If you look at successful startups, a lot of time they had the idea that nobody believed in, they had the vision, they saw the opportunity. They were able to establish it, developed it, and eventually what seemed a very fringe idea became mainstream. You have those kinds of plays.
Sometimes you have plays where there is an established product selling in the millions and then some engineer comes along with a better idea, a better mousetrap. That's why I say that it is individuals who matter. If you have the right person, someone else can put 10 people on it and they won't succeed against him.
Editor - The fist group brings to mind the book “Innovator's Dilemma” subtitled “Why successful companies fail”.
I was fortunate to get an hour from George Janac at a very busy time. His company, Silicon Navigator, just announced $6.0 million in Series-A funding. Investors include Jim Solomon, Chairman and founder of Cadence Design Systems. The company plans to build interoperable software components for optimization of multi-million gate design flows for power, performance and yield. The company's first Library Smart tools will be announced during the first quarter of 2005. Key SiNavigator technologies are now being licensed to partners and customers at the API level. The product strategy will leverage the OpenAccess data base now managed by the Silicon Integration Initiative (Si2).
Weekly Headlines
EDA
Cadence Sponsors Industry Forums; Series Addresses Key Challenges Affecting the Design Chain
Mentor Graphics and Xilinx Collaboration Reduces Design Time and Optimizes Performance for Integrated FPGA-on-Board Designs
Altium takes "Low-RISC" approach to 32-bit system development on FPGAs
Altium adds Verilog HDL support to Nexar
IBSystems Adds FreeCADApps.com to Its Growing On-Line Engineering Network to Offer More Than 6,000 Free CAD Software Downloads: Acquires This Asset from IMSI
Cadence Senior Vice President & CFO, Bill Porter, To Present at Deutsche Bank's 2004 Global Technology Sector Conference for Semiconductors & Capital Equipment
Accelerated Technology's Nucleus RTOS Now Available for Networked Applications Built Using the Freescale Semiconductor MCF5282
Thirteen Third-Party IP, Design Services and EDA Tool Suppliers Announce Support of LSI Logic's RapidChip(R)
Atmel Extends and Expands Partnership With Mentor Graphics, OEM Agreement Spans Synthesis, Simulation and Verification
Synopsys Announces Earnings Release Date and Conference Call Information for Fourth Quarter and Full Year Fiscal 2004
True Circuits Selects Amos Technologies to Distribute Its Silicon-Proven Timing Intellectual Property
MatrixOne Releases Synchronicity Developer Suite V4.1, Increases Productivity and Reduces Cost of Developing Semiconductors
IP & SoC News
Intel Board Elects Paul Otellini Intel CEO; Craig Barrett to Become Intel Chairman
Microchip Technology Introduces Low-Cost, 8-Gain-Step SPI Bus Programmable Gain Amplifier; Family Enables Digital Control of the Analog Domain
ViASIC Announces New 0.13 TSMC Library; High-Performance Configurability with a Single Photomask
Texas Instruments Announces Volume Production of World's First 90nm DSPs Running at 1 GHz
Agere Systems Unveils Industry's First Single-Chip 48-Port Gigabit Ethernet Switch and Lowest Power Octal PHY
Fairchild Semiconductor's New Integrated Power Switches Meet the 1 Watt Initiative
Matrix Semiconductor Demonstrates Innovative 3-D Memory at the Popular Science "Best of What's New" Showcase
National Semiconductor Announces New Family of Color LED Drivers for Cell Phones, Smart Phones, Personal Digital Assistants and Other Portable Handheld Devices
Atmel and NTRU Announce an Unbeatable Hardware/Software Security Combination for Trusted Computing Products
Conexant Names Dwight Decker Chief Executive Officer; 15 Percent Operating Expense Reduction Planned for Current Fiscal Year
Atmel Announces Four New LCD AVR Microcontrollers
UMC Reports Sales for October 2004
Infineon Reports Increased Revenues and Earnings for Fourth Quarter and Fiscal Year 2004
TI Unveils New Class of Industrial Data Converters With Widest Bandwidth 24-Bit Industrial ADC
National Semiconductor Announces New High-Performance Converters That Slash Power Consumption in Cell Phones up to 80 Percent and Extend Talk Time as Much as 90 Minutes
National Semiconductor Introduces New WEBENCH Online Design Environment for Audio Amplifiers
Adam Traidman Joins Giga Scale IC as President, Company Confirms Commitment to Reduce Electronic Design Cost
NVIDIA Introduces GeForce Go 6800 - Unleashing the World's Fastest Notebooks
LSI Logic Announces Webcast of RapidChip(R) Platform ASIC Partner Program Presentations
ON Semiconductor Introduces Industry's First EMI Filters Arrays with ESD Protection in Thin DFN Packages
Toshiba Files Patent Infringement Action Against Hynix Semiconductor
Actel Offers High-Performance, Easy-to-Use FPGA-Based DSP Solution
AMI Semiconductor Introduces Two New Transceivers for Automotive Applications
Matrix Semiconductor Announces Second Generation 3-D Integrated Circuit Memory Technology
ANADIGICS Introduces Industry's Most Highly Integrated Dual-Band WLAN Power Amplifier
More EDA in the News and More IP & SoC News
Upcoming Events...
--Contributing Editors can be reached by clicking here.
|
|
|
|
|
|