Why EDA differs from ERP in more than two letters

Posted on January 16, 2009 by paulmcl

Oracle HQWhat is it about EDA that makes it different from other software businesses? When the CFO of Texas Instruments buys Oracle or SAP he or she doesn’t study what algorithms they use in their relational database. EDA purchasers are the only people who “take the cylinder head off and look at the valves” before buying.

I think it is the speed of change. EDA, as we are all tired of hearing, is driven by Moore’s law. But the effect is that every few years a complete technology re-investment needs to be made and the incumbent does not have much of an advantage in discontinuous change. It’s not like that in other software industries. For Oracle, the last major change in databases was the relational database superseding hierarchical databases, and that was starting in the 1970s (first at IBM and then when Larry Ellison founded Oracle under its original name, Software Development Laboratories). It looks like there may be another change starting, driven by the internet, to schema-free databases which scale better to thousands of servers. So one turn of the handle of Moore’s law in 30 years, 15 to 20 times slower.

You could give me half a billion in VC money and I’m not going to be able to put together a startup to displace Oracle, no matter how many of the best database programmers I hire. Because it is not mainly about technology. And even if, by some miracle, I succeeded it would take 20 years. By contrast, when EDA companies miss a transition they tend to vanish quickly. Calma was not a major force in gate-level design. Daisy missed synthesis despite Synopsys being staffed with many of the same people, and, to add insult to injury, in the same buildings. Cadence lost its Dracula physical verification franchise to Mentor’s Calibre as quickly as it took 0.35um to come online, a few years. Yes, mistakes were made. Daisy made an ill executed acquisition of Cadentix. Cadence decided to make its own hierarchical DRC, Vampire, incompatible with Dracula so it didn’t harm its cash-cow. But mistakes will always be made (“new” Coke, Ford Edsel..) but in other industries one mistake is rarely fatal.

The speed of change also means that startups can get traction in a way that they don’t in other industries. The barriers to entry are really low, just a few people who really know the technology and it is possible to build a world-class product that is better than anything else out there. Most importantly, customers will buy it since the risk of using a product from a startup was lower than the risk of not doing so. Other businesses don’t move so fast. Waiting for the big guys to have it is usually the safest approach.

This may be changing as the importance of integration increases and so the important of point technology diminishes. I spent years at Compass Design Automation with a fully-integrated toolset that was very productive. But customers would only buy “best-in-class point tools” and do all the integration themselves. We were selling an engine when people wanted to buy their own ignition system, their own fuel-injectors and make the wiring harness themselves. Despite the limited commercial success (Compass only reached $55M before it was acquired) I still believe that the integrated approach really was superior for everything except unusual chips like memories and microprocessors. The evidence was that every chip VLSI Technology produced until the late 1990s was produced entirely on Compass tools on very short time-frames. In that era, for example, a huge percentage of mobile phone chips, then and now a market with short product cycles, were done that way.

It is no longer clear that most semiconductor companies have the inclination or manpower to look at tools from startups. Their focus is more on reducing cost and reducing the number of vendors they use. It is completely unclear how EDA will evolve in the current downturn. In Mike Santorini’s memorable image, semiconductor is the speedboat that pulls the EDA water-skier. When it is at full speed we ski well. Right now I think the water is around our waist and we are still sinking.

Posted in eda industry, marketing | Comments Off

We are fleas on a sick dog

Posted on January 15, 2009 by paulmcl

Well it was the annual EDAC CEO panel last night. The quality of the food was much better than normal, but it turned out that this was not because the CEOs were predicting a rich future in 2009 but that the San Jose Convention center had provided it (thanks).

For the first time Rajeev Madhavan of Magma was on the panel, and ghost of Cadence didn’t show although I assume they were invited (or maybe not, are they still even in EDAC?). Predicting their own future for a few months is hard enough, for the entire industry over a year is a distraction too far. Aart and Wally were there as always. Chris Rowen of Tensilica rounded out the panel carrying the IP torch (although when asked whether IP was truly a part of EDA he gave the one-word answer “yes”).

Ron Wilson’s coverage of the panel is here.

Aart went first and had us all text what type of recession we expected, using his Recession Compiler to add up the answer. We weren’t optimistic, expecting a long recession of a couple of years in total with unemployment going over 10% which it has only done once in the last 50 years. Apparently 38% of the CEOs Aart talks to, along with a few Wall Street types, think the same. Aart seems to think that the Obama government will wisely spend our future taxes today and that this will be good for tech as we unwind the discontinuity between the way we have been living and our levels of debt. Personally I am skeptical that the right solution to not being able to make our mortgage payments is to take on a bigger mortgage. Isn’t that at least partly what got us into this problem in the first place? Aart didn’t predict a number for the whole industry but was bullish for Synopsys themselves, having been giving guidance of 3.2-5.5% to the street.

Chris Rowen insisted that EDAs troubles were nothing to do with EDA or even semiconductor. “We are fleas on a sick dog.” But there is lots of growth in IP in the dataplane so Tensilica should be OK.

Wally was next. “In God we trust, everyone else bring data”. And as usual Wally had lots of it. It turns out that semiconductor is fairly disconnected from the real economy. If the economy is bad then things might not be too bad for semiconductor because we might be short on capacity. If the economy is good, semiconductor can shoot itself in the foot by overbuilding fabs. What EDA does tie well into is semiconductor R&D, which has previously only ever had one down year. Wally’s model for last year predicted 2% growth but turned out to be -12% although a lot of the difference is due to Cadence’s implosion when its business model ran out of steam. His model predicts 12% for this year but tempered down by 10% (the downturn in semiconductor R&D spending came late in the year just like 2002) and another 8% down for Cadence’s business model change giving -6% overall.

Rajeev also thinks we can farm the subsidies of the new administration as all that money flows into green technology. Plus, hard times might make new technologies get adopted faster, although it is not clear to me how Blu-Ray not surviving would drive new EDA business, although I suppose EDA gains most from new technologies in development and there is no further upside from chips that have already been designed.

I think that the whole economy is too uncertain to have much hope of making a solid prediction and that, in particular, nobody can even make sensible decisions until the government stops doing things and everyone knows what the rules are. We were asked to predict the stock prices of Cadence, Mentor, Magma, Synopsys and MIPS at the start of 2010. My guesses were 6 for Cadence (it’s 4 today), 5 for Mentor (same as today), Magma would have gone private, MIPS would have been acquired and Synopsys would have declined to 15 (from 18 today). Let’s see how that holds up.

Posted in eda industry | Comments Off

Super-models

Posted on January 15, 2009 by paulmcl

I wrote earlier in the week about open source software in EDA, or rather about the lack of it. One area where there is some free and open source software, as well as closed source software, is on that boundary between EDA for chips and software tools for embedded systems, namely what seem to be called virtual platforms or virtual prototypes (I hate the name “virtual prototype” since it is a chip-centric view of the world implying that the platform is useless once the chip shows up).

Virtual platforms, while they have some utility for chip development, are largely sold to software developers to allow them to do software development more productively and earlier than would be the case if they had to use the real hardware, which comes along too late and is too opaque. The performance of the virtual platforms is almost unbelievably high, running ARM or PowerPC code binaries at hundreds of MIPS on an off-the-shelf PC, often similar to the performance on the actual hardware.

I have worked for both VaST Systems Technology and Virtutech who both supply tools into this market. They charge per seat in the region of $5-25K/seat/year. In the IC design world these price point are low; in the software development world they are very high. Synopsys with its Virtio acquisition is also in this market. Imperas is a startup founded by Simon Davidmann in the UK to enter this market specifically to address the difficulty of programming multicore chips. Driven by a mixture of lack of funding but also a deliberate change of strategic direction, about a year ago they made their environment free and created Open Virtual Platforms (OVP).

Blue Ocean StrategyI met Simon at DATE last year where this was announced, and asked him why he did it. Firstly, he said that he is a big fan of Kim and Mauborgne’s Blue Ocean Strategy, changing the rules and competing where the competition isn’t. But the thing that really brought it home was discovering a fact about QEMU. QEMU is a similar type of simulator developed largely by one person, Fabrice Bellard, and distributed free (and open source). The fact Simon discovered is that QEMU has more Google hits than Synopsys.

qemuThink about that for a moment: a single free product that most of you have never heard of in a neighboring space to IC design has more web references than the EDA market leader has for all their products put together (they are almost identical at around 1.7M apiece when I looked just now).

Simon also realized that companies made more money from verification tools around simulators than selling the simulators themselves. So for Imperas the key would be to get people using the simulator so that there was a base into which to sell higher value tools. It is too soon to tell whether the strategy is working fully, but MIPS and Tensilica are both distributing models on the OVP foundation.

When I was at VaST and Virtutech it was clear to me that the market would be limited so long as models were not being supplied by the component vendors, either at the same time as or in advance of silicon. I always used the analogy of Synopsys in the early days. At first Synopsys themselves developed the ASIC vendor libraries necessary for synthesis. Bob Dahlberg, who ran the group, told me that at one point he had well over 100 people doing this. Then the ASIC vendors realized that it was their job if they wanted the job done how they wanted it done when they wanted it done. A year later Synopsys disbanded the group completely since ASIC vendors had completely taken over the task.

This is starting to happen in the automotive industry around VaST’s technology, For some time the main suppliers into the automotive industry (NEC, Renasas, Infineon, Freescale and others) have supplied processor models for VaST’s environment. NEC America is announcing today that they will be distributing complete virtual platforms on VaST’s foundation technology into the automotive industry, going beyond simply providing processor models. Software engineers in tier-1 suppliers (automotive-speak for people like Delphi, Visteon and Denso) and OEMs (automotive-speak for car companies like GM, BMW and Toyota) will be able to develop their software without having to wait for silicon to be available and in a much more productive environment than the real electronic control unit that will eventually ship in the cars.

However, I think that component suppliers will continue to remain reluctant to develop models for the virtual platform ecosystem while there are limited standards for interoperability or, as an alternative, a de facto winner in the same way as Synopsys was clearly the early winner in synthesis. Even a company like Freescale, which distributes VaST models into the automotive industry also distributes Virtutech models into the communication (think router and base-station) industry, which is clearly not optimally productive.

iPhone simulatorThe situation where OVP, VaST, Virtutech, Virtio, QEMU, Bochs and others all have incompatible virtual platform environments is not really sustainable. SystemC provides some standardization around modeling of peripheral devices where performance is not critical, but processor models depend heavily on the underlying simulation technology to get their blazing performance.

The other alternative is native cross-compilation environments. If you develop software for the iPhone Apple supplies a Mac-based iPhone simulator. It is fast but people complain about its accuracy especially for graphics. But presumably Apple decided it was not worth using true virtual platform to get the accuracy at some loss of performance, or maybe they didn’t even know just how fast simulation technology can be. Also, it is not so much iPhone application software but the call processing and low level software that absolutely requires a high accuracy platform.

It will be interesting to see how this all plays out.

Posted in engineering | Comments Off

Intel only needs one copy

Posted on January 14, 2009 by paulmcl

It is obvious that companies make money in EDA only if they sell enough software. One rule of thumb is that EDA companies thrive if each salesperson brings in $2M per year, and they don’t if they only bring in less.

But enough software really means enough hours of use of the software. For a large EDA company, most of the money comes from a relatively small number of large customers, and they optimize their use of licenses in server farms, sharing licenses world-wide and so on.

But enough hours of use of software in turn really means that either the software must run for a long time (like place and route or RET decoration) or else that customer engineers must sit in front of it for a long time (like a layout editor).

Other tools suffer from what I call the “Intel only needs one copy” problem. They have a hard time building license demand naturally. This is less of a problem in a startup, who are quite happy in the early days to sell one copy to everyone, but to get a good growth trajectory it is necessary to build on the beach-head of those first licenses and proliferate widely into at least some of the accounts.

If license demand isn’t built naturally then it becomes necessary to attempt to do unnatural things like try and charge per tapeout, or try and license on a per-named-user basis, or try and charge a royalty. These are all possible but at the very least the sales cycle will stretch out for a startup, and it will run out of cash, or for a large company it becomes too complex to include a weirdly licensed tool into a large contract (which, incidentally, is also one reason that OEM deals never work in EDA).

This is one of the big challenges of the ESL market. The tools are only needed occasionally, don’t run for a very long time and don’t require users to run them interactively for long periods.

Bottom line: it is really hard to sell a tool with an unexpected business model, which for EDA means some sort of floating license for a period of time. A nice analogy is the restaurant business. When you go to a restaurant you expect to pay depending on what dishes you order. That’s how restaurants work. But in fact most of a restaurant’s costs are fixed: the rent, the employees’ salaries, utilities, advertising. So rationally a restaurant might charge by the minute no matter what you eat. That changes things a bit (caviar is cheap, that espresso after dinner is really expensive) but even so I suspect you’d have a hard time running a business that way. It’s just not what the customers expect.

Posted in eda industry, marketing, sales | Comments Off

Where is all the open-source EDA software?

Posted on January 13, 2009 by paulmcl

Open source logoThere is a big cultural difference between tools for IC design and tools for software design. A difference in the way they are developed, the way they are sold, the way they are deployed. I think there are two reasons.

Firstly, IC designers are not software designers (duh!) and so are not generally capable of writing or extensively modifying their own tools, so it doesn’t cross their mind that it might be a good use of their time to do so. After all, it wouldn’t be.

Secondly, the culture has arisen that the studliest IC designer (or whatever the equivalent is for a woman) has the most money spent on their tools. Companies capitalize IC designers with large amounts of software to increase their productivity. But it is not really productivity, it is the ability to get the job done at all. Only in the most technical sense is improving the time taken to design a chip from a millennium to six months merely a productivity increase. The EDA industry has improved engineering productivity by a factor of maybe 100,000 in the last thirty years.

Neither of these are true in the world of tools for software engineers. Firstly, they are software engineers and so can develop software fast by either just going and developing it on whatever is available (without requiring centuries) or they can write themselves some productivity tools and then use those tools to produce their product with higher efficiency. They are their own customers in that sense.

Open source development has also been shown pretty conclusively to be more efficient the closed source. Eric Raymond’s1 book “The Cathedral and the Bazaar” is a little dated but still the best survey of this. But there is a problem. Open source software is also known as “free software”. Originally this meant “free as in freedom not free as in beer” but has come to mean “free as in beer” too. Or at least very cheap.

Open source means that if you buy the product, and perhaps even if you don’t, you also get the source code and can do what you want with it. Of course, since you can do what you want, it becomes really hard to sell a second copy since you can always build that yourself from the source, so the second copy had better be really cheap or that’s what you’ll do. For example, Wind River used to have a proprietary royalty-bearing operating system called VxWorks but the world is going to Linux and so Wind River supplies their own version of Linux. But it is hard to charge much and impossible to charge a royalty on this since a customer can get Linux from many places and even hire a few engineers and build their own version. Or a competitive company can take the same source and customize their own product. It is hard to see what Sun has got from Java to justify its investment, and even harder to see why it recently bought MySQL for about a billion dollars when the price for a copy of MySQL is exactly zero.

In that way, open source is a bit like Craigslist. Craigslist didn’t steal all the money from newspaper classified advertising. It took a billion dollar business and made it into a million dollar business, making it impossible for anybody, even Craigslist themselves, to make real money in classified advertising.

So the result of all of this in the world of tools for software development is that all the best tools are open source, but nobody can make any money selling them. This works fine as long as enough people like Sun and IBM pay their developers to do open source development on the basis they make money on the hardware, or enough programmers do this in their spare time for fun (and because if you want a job at somewhere like Google, one of the things they’ll take a look at is what open source projects you work on after hours). But there are no Microsofts of open source, no Oracles nor Adobes. Not even Intuits or Mathworks.

IC design tools are all closed source, apart from a few bits of infrastructure like openAccess. Synopsys isn’t about to give you the source code for Design Compiler just because you bought a license, and they certainly aren’t going to put it up on the web so Cadence can grab themselves a copy too. It is arguable whether the quality would even improve that much if they did so since most of the users are not itching to get into the millions of lines of source code and add a few enhancements.

Now that electronic systems contain large amounts of software content, these two worlds are starting to collide a bit. The investment for a system design team in tools for the IC part is maybe in the millions of dollars, and for the software part it is essentially zero. And not because they are forgoing the best tools because they are too cheap; those free tools are the best tools.

The number of chip designers is fairly small, perhaps 20,000 engineers. There are ten million software engineers. The price of EDA tools has to be high to recover the development cost over a small base; software tools have a much larger base. After all, Bill Gates got rich selling copies of MS-DOS and Windows for less than $100 each. But he sold a lot.

1 Interestingly, since open source sounds like some sort of socialist commune, everyone tends to assume that Eric Raymond and other open source luminaries are pinko lefties. Actually he is hard-core libertarian with interesting views on other issues than programming. His blog is Armed and Dangerous.

Posted in eda industry, engineering | Comments Off

City Slickers marketing

Posted on January 12, 2009 by paulmcl

I have done a fair number of consulting projects for EDA startups and a lot of them start out with what I like to call “City Slickers marketing”, named after the movie City Slickers. For those of you who have not seen it, there is an old cowboy, Curly (played by Jack Palance) and a young advertising account manager Mitch (played by Billy Crystal). The marketing is named for the following conversation:

Curly: Do you know what the secret of life is? [holds up one finger] This.
Mitch: Your finger?
Curly: One thing. Just one thing. You stick to that and the rest don’t mean shit.
Mitch: But, what is the “one thing?”
Curly: [smiles] That’s what you have to find out.

When I arrive at startups where the CEO is the key technologist, or even just an engineer by background, I tend to have a conversation like this:

“What’s the one reason people should buy your product?” I ask.

“One reason, I can give you twenty,” the CEO replies.

Technical people (and I am one, so this is a lesson I had to learn the hard way too) tend to overvalue technical features in a product and assume that if the technology is good then the product will sell itself. And if one feature is a good reason to buy, then lots of features are even more of a reason.

But the world doesn’t work that way.

A colleague recently reminded me about a store in San Jose that sold “Fine art and bicycles”. Presumably a fine art dealer who also was a keen cyclist. This is an extreme example of how multiple features are not necessarily additive and how you have to take account of the way the customers look at the world. There simply isn’t a “fine art and bicycle” market that you can be the leader of. In fact, even if you are the best art shop in the area, selling bicycles too isn’t a plus, it detracts from your message.

A lot of early marketing in a startup is working out what the single compelling reason is for a customer to engage with you. And it has to be just one (or maybe a couple if you can segment the market a bit). The early stage of engaging with customers is sometimes referred to as throwing mud against the wall and seeing what sticks. You can’t find the single compelling reason as an intellectual exercise, you have to get out and engage with customers and work out where your technology solves problems the customer cares about.

When I arrived at Ambit, I learned that our value was that we produced faster circuits than Synopsys. And we had better time-budgeting. And we could run top-down. And we ran faster. And our pricing was bundled. And physical synthesis was in development. And…and…and.

It was once we realized that we could handle large million gate designs in one gulp that we really started to get traction. The other things were all true, but none of them was compelling enough to get a company to engage with a startup. But if you had a million gate design and you couldn’t get it through Design Compiler, then you were calling us to take your money.

This is different from the elevator pitch for investment purposes. Ambit was “Design Compiler only better” but that isn’t focused enough for marketing a product or driving a detailed development roadmap. You have to find out the one thing.

Posted in marketing | Comments Off

Molten lava

Posted on January 9, 2009 by paulmcl

I was going to write an entry about how it didn’t look like Magma (NASDAQ: LAVA) was going to survive, what with losing cash, a lot of debt due, and not much in the bank. Employees don’t like it when you don’t make payroll. I heard that private equity people were looking at Magma but were not really interested. It wouldn’t have cost much to buy Magma but it wouldn’t be quick to fix so it would take a fair bit of investment before it was dressed up ready for a new party.

The latest rumor is that it is indeed going private but that Rajeev himself, the CEO, along with Andy Bechtolsheim, one of the original investors, are going to buy it off the public. They already own 10% or so each, and much of that Magma debt has been trading at 30 cents on the dollar making it cheap to get rid of. If this happens the hard work then begins. Rebuilding the technology base takes time but Magma still has some outstanding engineers and they did it all before in the face of a lot of skepticism when they first started, but don’t forget it took nearly $100M in venture capital to get there.

Posted in eda industry | Comments Off

Guest blogger: Jim Hogan on IP

Posted on January 9, 2009 by paulmcl

Jim HoganThis is the first of what I hope will be a regular feature of this blog: guest entries from other people with insight into EDA and the ecosystem around it (venture capital, semiconductor, embedded software…)

Jim is one person in EDA who really doesn’t need any introduction. He was one of the early guys at Cadence and was a mainstay there for many years, in particular driving a lot of their (then!) strategy doing advanced product R&D by acquiring successful startups. Since then, after a stint at Artisan, he has been a VC, a private investor and a board member of many EDA startups. Not many people would try to start an EDA company without running it by Jim.

The rest of this entry is Jim not me (Paul).

It is clear that EDA in its core flow is increasingly under pricing pressures from customers and is being commoditized.  Using the “building the house” analogy, EDA needs to move from building hammers to extracting more value by being at least a significant subcontractor and if possible the general contractor.

I believe that EDA will need to move to the system level and software signoff specifically.  A necessary component of this strategy is IP. 

Interestingly enough, it appears customers are also viewing their IP assets in a new light.  An event unimaginable a few years ago has occurred. Nokia sold off its IC implementation team to ST Micro.  It appears they view this asset as something they don’t have to build or hold as a core competency, but can buy.

Let me try to put this in a context familiar to everyone.  When was the last time you bought a car and looked under the hood?  You don’t need to know what the engine is to use the automotive appliance.  You still care about the car cost, performance, features and power, but if those needs are met then who cares what the engine is.

Synopsys clearly has integrated this thinking into their long-term strategy and are executing.  Here are some recent quotes related to IP from their earnings call:

“Synopsys is uniquely well-positioned to be the partner of choice:

One, we are financially stable, with a strong cash position, no debt, good cash flow, strong backlog and a conservative business model–making us a stable partner as customers actively de-risk their future.

Two, we have strong, well-integrated technology, in a comprehensive solution, that is complemented by crucial IP and very experienced global support ready to help customers streamline their overall cost of design.

And three, we have the strategic vision, and resources, to continue to invest in the key technologies of the future.

…

Our IP business also deserves mention. We not only added a significant collection of new cores to our offering, but also grew the business as outsourcing semiconductor IP for cost reasons is becoming more mainstream. "

Synopsys, because of their history DC and DesignWare, comes by this strategy naturally enough.  However, they clearly in recent acquisitions have demonstrated they understand that FPGAs are becoming more than just prototyping solutions but viable production vehicles.

EDA, by focusing on differentiated IP, or System-Level IP, offers several benefits to the customer: Predictability, Price, Performance and Power.  It begs the question is EDA the right term at all anymore. We all need to re-define the space we are and will serve.

One example of System-Level IP is a processor.  I believe the processor decisions have arguably already been made and customers have voted to use ARM especially in wireless applications.  

In my opinion, there is a substantial untapped System-Level IP opportunities in block level interconnect and memory control–the architecture through which IP is integrated and IP interoperation is achieved.

The new platform opportunity for what EDA becomes lies in the successful integration of third party blocks in a System on Chip (SOC) or an FPGA.

Posted in eda industry, guest blog | Comments Off

Three envelopes

Posted on January 9, 2009 by paulmcl

Three envelopesCan there be any subject more boring than revenue recognition for software? If you listen to the conference calls of the public EDA companies, you’ll either hear them discuss or get asked about how much of their business is ratable versus term. What does this mean? Should you care? Also, what does it matter how long the term is, isn’t longer more money and so better?

When Jack Harding was CEO of Cadence, he lost his job because of these details. Cadence had been selling permanent licenses (for historical reasons I’ll maybe go into at some point, EDA had a hardware business model). The sales organization had come up with the concept of a FAM, which stood for flexible access model. The basic idea was great. Instead of selling a permanent license valid forever, sell a license valid for only 3 years for not much less. Then, three years later sell the same license again. The lifetime of a permanent license had proved to be about 6 years in practice, so this was almost a doubling of the amount of money extracted per license. This was then scaled up into “buy all the licenses you will for the next three years today”, with some flexibility built in by throwing extra licenses into the mix. This was done in a way that meant all the revenue, or most of it, could be recognized up-front.

There turned out to be two problems with this once it was scaled up. Firstly, the customers didn’t really know what licenses they needed in year 3 although they had a pretty good idea about years 1 and 2. So to get them to go for this, the third year discount had to be huge. The second and bigger problem was that in two years the Cadence sales force closed three-year deals with every large account they had. The combination of these two things mean that every customer acquired all the licenses they needed for the next three years, but it was all booked in two years (and for not much more than two years’ worth of money). Numbers looked great for two years but in year three there were no customers left. The wheels came off and the numbers cratered. Jack Harding was CEO at the time and a couple of days after the quarterly conference call he was gone.

There’s an old joke about a new CEO starting his first day and being left three envelopes by the outgoing CEO. He is told to open them when things get really bad. Things go OK for the new CEO for the first few months and then there is a downturn in business so he opens the first envelope. “Blame your predecessor” is on the card inside. So he makes speeches about how he inherited a company on the brink of ruin from the old CEO and the analysts and press give him a break. The second time things look bad, he opens the second envelope. “Reorganize.” So the newish CEO takes all the business units and carves them up into functional divisions. That seems to fix business for a time. But eventually the future is not looking so bright any more and the now-not-so-new CEO opens the third envelope and read the card: “Prepare three envelopes”.

So after Jack Harding left, Ray Bingham came in, opened the first envelope, and said Jack Harding and FAMs were bad, Cadence would henceforth book ratable business. Depending on details of the wording in the license, FASB (Financial Accounting Standards Board, a bunch of academic accountants from the east coast who’ve never run anything, but that’s another story) forces the revenue to be recognized up front (like a FAM) or quarterly (“ratably”) over the three years of the contract. With ratable business almost all of a quarters revenue comes out of backlog so it is very predictable, and there is a lot less pressure to close business at the end of a quarter (because only 1/12 will drop to the bottom line) which should lead to better discounting behavior.

However, there was pressure for Wall Street for growth and one way to provide that was to start to mix some term FAM-like business in with all the ratable stuff, since it dropped to the bottom line immediately. That was why smart analysts were so focused on what percentage of business was ratable. If you don’t know that, you have no idea if the numbers are good or bad, or if they are sustainable.

[Full disclosure: Cadence acquired Ambit, where I was working, towards the end of Jack Harding’s tenure. I then worked for Cadence for three years including working directly for Ray Bingham for a period. I left before Mike Fister came on board.]

Eventually the board brought in Mike Fister as CEO and Ray became Chairman. Mike Fister hadn’t heard the joke, obviously, since he omitted to open the first envelope. He had a perfect opportunity to take a big loss, switch to ratable business and generally blame anything he wanted on Ray. Instead, he kept going on the same treadmill. To all of us observing from outside it was clear that Mike Fister wasn’t going to make some new and interesting mistakes, he was going to make the same old mistake all over again. So it was no surprise when it turned out that the rate of growth was not sustainable, that they were booking ridiculously long-term deals of 5 or more years. The reason that this is bad is that a 5 year deal is not a green-field deal with a virgin customer, it is a 5 year deal with a customer who already has a 3 year deal, and customers don’t pay much for a deal to buy software for years 4 and 5, let alone 6, 7, 8; they are not under any pressure. So eventually the wheels came off again. There was even some restatement of revenue associated with, surprise, whether some deals were correctly recognized as term or ratable.

So Mike Fister got to prepare his three envelopes and now we know it is Lip-Bu Tan who will open them. Watch for a big reset, blaming Mike for all the terrible deals he left behind, and lots of talk about starting with a clean sheet.

Posted in eda industry, management | Comments Off

“Put your money where your mouth is”

Posted on January 8, 2009 by paulmcl

Talking of being CEO of Cadence, not many people have been parachuted into running a troubled EDA company.

But I have.

At the end of 1996 Compass Design Automation was coming up to a fifth successive quarter in which the revenue declined, not just from the same quarter the previous year but simply from the quarter before. However, revenue had not cratered and was still running at well above $50M/year. Personnel turnover in the US had reached over 100%. With around 200 employees in the US, over 50 had resigned in the previous quarter. Overseas turnover was still much lower. So things were pretty bad, but not so bad that recovery was out of the question.

I had already left Compass a year or more earlier, due to major differences about how the company was being run, and I was working at VLSI, the parent company.

Al Stein, the CEO of VLSI came into my office. “You know you are always giving me advice on what I should do with Compass,” he said. “Well, put your money where your mouth is and go and run it.” Just like that I was CEO of Compass.

Al wanted me to turn around Compass and find an acquirer. I told him that was impossible. I could try and turn it around, which I guessed would take a minimum of two years, or I could try and stabilize it and find a buyer. Trying to fix it and find a buyer is like dismantling your car and advertising it on Craigslist at the same time. “Trust me, once you reassemble it you’ll have a great vehicle”. So we went for the second option, to stabilize the company and sell it.

So that’s your challenge. What would you have done? And for a bonus where we don’t know the answer1, what would you do if you were CEO of Cadence?

1 We managed to get our turnover down, produced three quarters of successively increasing revenue and sold the company to Avant! for over 1X revenue. When I first took the job, I wasn’t certain we’d last another quarter so I count that as success.

Posted in eda industry, management | Comments Off