One-stop Shopping

who wins and who loses ...

by Peggy Aycinena

This week’s article is about a subject near and dear to the hearts of many in EDA. Is it good or bad for customers and/or vendors to have one CAD tool vendor "owning the flow" – to have one vendor supplying basically all, or almost all, the tools needed to execute a design? The idea is described in many ways; however frequently the terms "One-stop Shopping" and "All You Can Eat" are used to capture the concept.

This is a lengthy article and I would suggest you go get that extra tall, extra hot latte and read the thing to completion. If you do, you’ll access the nuances of various points of view related to this very important topic.

One-stop shopping has always been a sub-context for everything that goes on in this industry, but I would argue that today the subject even more critical. It’s risen from sub-context all the way up to the level of context, and beyond.

The following people have contributed comment here:

* Daya Nadamuni, Principal Analyst for Dataquest/Gartner Group

* Graham Bell, Senior Marketing Director at Nassda Corp.

* Mike Schuh, General Partner at Foundation Capital

* Wally Rhines, CEO & Chairman of the Board at Mentor Graphics Corp.

* Pravin Madhani, Founder, President & CEO at Sierra Design Automation

* Alberto Sangionvanni-Vincentelli, EE Professor at U.C. Berkeley and Chief Technology Adviser at Cadence Design System.

I spoke by with Daya, Graham, and Mike by phone. Those conversations helped to generate the questions I then posed to Wally, also by phone. I used those questions as well when I communicated via e-mail with Pravin and Alberto.

I want to thank everyone on the list for the time and effort they gave to this discussion. And I also want to say a special thank you to Ed Lee of Lee PR for proposing this project and facilitating a number of the conversations included herein.

This is an important topic. It’s not going to go away.

***************************************

Daya said that CAD tool users are interested in the interoperability and price of the design software that they buy. They’re also interested in the amount of time needed, or available, for interfacing with the vendors and their sales people, as well as the learning curve on the various tools for the engineers who use them, and the amount of integrated support available for the tools.

There are some advantages to having one vendor supplying the entire flow, according to Daya. The tools are usually second-sourced anyway, which means that there’s less compelling need for multiple vendors. Also, many times with large user companies, they’re standardizing on a flow – which on the face of it means one vendor could supply that flow.

However in reality, no vendor can supply all the tools needed for the complete design flow because there are still specific point tools that are only come out of a few specialty vendors. It’s true however, there are often issues with these small specialty vendors. Large tool customers frequently make their purchase decisions by looking at the financial stability of the vendor. That can be a problem for small start-ups who are not yet profitable.

In any case, Daya said that back in 2000, some companies were talking as if they could actually make a play as an all-you-can-eat vendor, but that market position didn’t last. It just never really worked.

Daya enumerated the advantages of having multiple vendors in the flow. She said that leading-edge customers make purchase decisions based on the strength of the tool they’re purchasing, not necessarily based on the financial attractiveness of a bundled set of tools from one vendor. Most big customers, she said, do not go with just one vendor, and again – there are so many small companies that come out with best-in-class point tools, the large customers are not going to go with one vendor if it means eliminating the ability to use those point tools where appropriate.

Per Daya, the innovative thinking in EDA happens in the small start-ups. She said it’s true that the large EDA companies do spend a lot of money on R&D, but the leading-edge breakthroughs still tend to come out of the start-ups. She added, however, that every company has the potential to be an R&D leader – large or small – it all depends on their in-house skill set.

Daya said that consolidation is a natural evolution in the business cycle of every industry. She said that in her opinion, consolidation has characterized the EDA industry no more or less than many other industries – particularly in software industries.

Daya believes the optimal number of leaders in a software industry is three, possibly four, and that the dynamics in many different sectors could be used as evidence to support that opinion. She acknowledged that it’s difficult to know how it feels to be in the shoes of a small start-up that’s having trouble getting market traction because of having to compete against the large vendors. But, she said, large companies that try to establish and maintain control over an industry always face a long-term uncertainty in that position.

Finally, Daya said her numbers indicate that today in EDA, the top vendors control 60 to 75 percent of the market. The incumbents always have the advantage – but if you’re a small tool vendor and you have something innovative to offer to the market, you can and will overcome the barriers to entry thrown up by your larger competitors. There may be turf protection by the big companies, but new entries into the market need to look past that, and try to integrate themselves and their products into the overall design ecosystem to establish a market position.

***************************************

Graham Bell, Senior Director of Marketing at Nassda – an organization in the process of being acquired by Synopsys, Inc. – spoke to the issue by referencing a recent article in EE Times where Sequence Design, Inc. President and CEO Vic Kulkarni apparently spoke about the larger EDA firms bundling their tools.

"[Kulkarni] also lashed out at larger EDA firms, saying that discounting by them is hurting the industry as a whole. 'Some of them are cutting prices, bundling things and locking in customers with promises for the future, and these are not good for the EDA industry,' Kulkarni said. 'They have spoiled customers who in turn are asking for price cuts, especially around the time of the customers' quarterly results announcements.'"

Per Graham: "I think Vic is talking about some wall-to-wall tool suppliers who are really trying to consolidate their positions in the marketplace. They're asking their customers to reduce the number of suppliers they work with. They're saying that it's easier for you to manage your CAD budget if you only have one tool vendor."

"Sure, it’s easier for a customer to deal with one big EDA vendor than a number of small ones. From the customer's point of view, they're probably getting the software they need to do the design, but are they missing out on next-generation innovation? As with many if not most industries, small startups may have better technological ideas, but getting to market is tough."

"Traditionally, the small companies need years to ramp up to a point of profitability. With so much pricing pressure in the market, that process for small start-ups has become extremely difficult. In the end, the wall-to-wall EDA vendors’ sales strength gives them the advantage in closing the deal. The customer has to weigh the convenience of dealing with one big vendor against the advantages of blazing new technology from a handful of small vendors that may or may not be successful. Choosing between what’s economically beneficial versus what is technologically beneficial is the quandary."

"So who makes the purchase decision at the customer? The executive who is in charge of financials and is looking for the cheapest solution, or the end-user within the customer company? Clearly, the EDA vendor needs a champion at some level in the organization, and in large companies it's often the design team and their managers who are pulling for the ‘best’ solution. Whether the productivity and performance gains of using the ‘best’ solution justifies a difference in price and tool maturity is a judgment call."

"Certainly semiconductor companies try to influence their tool vendors as to what direction their R&D should go, and in that case – it's the large customer who has influence even with the large vendors. So the question here is: who determines and defines the inherent value of EDA technology? And, is that question different from which EDA technology will get me furthest along in my next design project with the minimum amount of cost and grief?"

"Perhaps, the end result of all of these pressures is that we're going to see a lot fewer EDA companies in the future. However, there are more startups bucking that trend and the number of exhibitors at DAC has been growing for the last 3 years running to over 230. That's because engineers love to do start-ups and to get their new solutions to market. And even though there's been a sustained downturn in the market, and even with customers having up to a 40 percent drop in sales in recent years, the EDA companies have, with a few exceptions, robustly survived"

"Users want next-generation technology but try to get the best deal possible, whether from large EDA vendors or startups. The large EDA vendors have created a model that offers pretty good value to large customers. But the question remains, as we get into 65nm designs, is "pretty good" good enough? And if not, how can the customer mindset be changed to encourage EDA innovation in all EDA vendors, big and small?"

***************************************

"It's hard to get excited about the EDA market leaders. The prevalent business model is one key reason: one-stop shopping – a single-vendor supplied tool flow. My point of view is parochial, as an investor in early-stage technology companies, but I think this business model is one of the things that ails the industry and one that is within the control of the market to fix. Vendors are to blame for locking in to a model with limited flexibility. Customers have managed to effectively negotiate licensing and pricing terms that result in public EDA companies being unattractive investments."

"From a start-up perspective, it isn’t simply the all-you-can-eat of the license model that's the challenge – it's allowing the customers to re-mix the tool suite that is often more problematic. Current licensing schemes are constructed such that there's an incremental price over the annual fee that you pay in order to be able to modify the mix of tools you use. To practice this business model in an industry with only 3 or 4 companies who provide everything they make to their customers, including all of the technology to be developed in the future at a set price, is unhealthy. I find that the large suppliers are using this business model, together with claims of future product development in a period of shrinking CAD budgets, to compete with start-ups. This is a problem for the industry."

"I don’t believe that the full line suppliers are developing the disruptive technologies that leading edge designs require. Publicly owned companies are often precluded from investing in a development team of 20 engineers for 24 months before being able to generate any revenue in order to meet profit goals. There is also a reluctance to disrupt (replace) what customers are buying. Over the past 10-20 years, new technology has come primarily from start-ups."

"Those fledgling companies need to be financed. Venture capital firms are designed to meet this need. While friends and family and some angels may continue to provide seed-level financing, the big expense in growing an EDA company is in the direct sales force. Seed-stage investors will be hard pressed to provide the cash required to build out a direct sales channel. This money can only come from institutional investors. We are attracted by interesting technology and smart entrepreneurs, the ingredients of a lot of EDA start-ups, but we also need to see real customer traction – revenue."

"Competitive tactics of the large vendors that focus on licensing schemes – offering to mix and match yet-to-be- developed technology for their existing customers – make it very difficult for the early-stage companies to sell effectively. If it’s not possible for early stage EDA companies to generate customer traction, then the venture capital will dry up and any disruptive technology that might be developed will die. Prospective customers of EDA start-ups might be wise to develop a sensitivity to this issue. The all you can eat terms are most frequently found in larger user organizations. These companies used to be the primary target for start-ups. The degree of difficulty in cracking one of these accounts is so high that often the young company turns to a smaller opportunity. I might note that the cost of selling is pretty much the same."

"Presently, the venture community looks at the EDA market opportunity as small. There are only a few venture capitalists who understand the market. Even the early stage companies, who should qualify for funding, have trouble raising it in part because of the difficulty of closing business. I am skeptical that the captive venture functions within the large vendors have the skills and experience to assist small companies to be successful, and many question the motives with regard to building a strong independent company. In order to ensure the availability of money from quality venture firms, the returns have to improve."

"In summary, if young companies disappear because they cannot achieve lift off, both EDA users and the large vendors will suffer. How will the problems of design verification, systems-level design and DFM, for example, be addressed if not by start-ups? Customers might benefit greatly by setting aside some budget dollars to invest incrementally in new technology from start-ups. They maybe early stage and therefore potentially short on mature processes, but they can be very responsive. It would be great to see somebody like Costello in the industry; someone who thinks exciting thoughts and acts energetically."

***************************************

Rhines – Is there such a thing? Not really, but for mature flows with the technology changing dramatically – there is sometimes an opportunity for one-stop shopping. So, for instance – between 1990 and 2000 in the PCB market, it was sort of that way. But that is no longer the case. Now, there are all sorts of value-added tools that go into the flow. Things like signal integrity analysis, the library infrastructure, the multi-user distributed computer tools, manufacturing-interface tools – they have all come together, so that now we have our sales [working into] their flows, and their flows into ours.

Another places where one-stop shopping has been successful is in low-risk flows like FPGAs. In that case, if the user is really a non-demanding customer – is not doing anything that challenges the technology – he can use the supplied tools from Altera, Xilinx, and other FPGA vendors. But if a customer wants a turnkey flow, Mentor is the only one providing that. For just synthesis, Mentor or Synplicity are both vendors.

In general, however, only the most mundane projects will go with a one-vendor flow. And certainly in the area of custom IC design, one-stop shopping has never worked in the past, and shows no sign of working in the future.

It's always been the desire of some customers to take their vendors' dogs with their stars – but it hasn't worked. The engineer rebels at that, and the risks are just too high. So typically, best-in-class tools emerge in the development of new technologies.

What happens is – if best-in-class tools come from start-ups – one of the major EDA companies will acquire them to try and package the sale of that product with other products. But the problem with [that kind of] one-stop shopping is that no one company can be best at everything. If you're involved in challenging technology or there's a high risk of failure on the project, you're just better off using best-in-class tools.

Rhines – Anyone who uses one-stop shopping is at a disadvantage relative to their competitors. Some companies have tried to, but their engineers are typically forced to incorporate inferior tools – or switch from what they're using. And so, they're at a disadvantage and may not be able to overcome that.

The part I find most interesting is that one-stop shopping means poor interoperability. Initially, you may say that doesn't make sense. But when I ask people whose tools are best integrated in the flow, they tell me of course, it's the start-ups. That's because the start-ups don’t sell anything unless they're integrated into the flow. The fact is that the start-ups do the best job of interoperability. They have no choice. And their best interoperability is the very day they are acquired by a large company – it only gets worse from that point on.

As long as the tool is based on its ability to integrate well into the flow, a lot of effort is made to make sure it works well in the flow. But one the company is acquired, the emphasis changes, and so does the ease of interoperability. So, if you're part of a big company and you have to make another division manager make changes to their products – it's a battle in the executive office as to which division has to change.

However, if you're an independent company, you have no choice. You have to change. So, usually the bigger companies have more difficulty in maintaining interoperability in their flow – and in other flows.

As an example – you'll find if you do a survey that Calibre has the best integration into the Milky Way flow, the best integration into the Cadence flow, and the best integration into all other flows as well. That's because we have to – we don't have a choice. The customers have a variety of different layout tools, and Calibre has to work to all of them.

Rhines – No, it's a true complaint. The negative is that, many times, superior tools are passed over because the existing ones are just good enough. The challenge for start-ups is to distinguish themselves with their capabilities.

Rhines – It's not the one-stop shopping that causes the problems. It's the sales and marketing reach of the start-ups. So in general, one-stop shopping is actually disliked by all but the laziest of customers. In one-stop shopping, you're using tools that are kludged together instead of those designed for interoperability.

Second of all – a single vendor negotiation is not attractive to a purchasing manager, because you have only one supplier and no alternatives. Engineers want well-integrated, best-in-class tools, and so they're not likely to be supportive of a bundled corporate strategy. Sometimes there are tensions with the management in charge of cost reduction, when they make an argument for taking the dogs along with the stars because they come at an attractive price.

But the engineer struggles with the tools if that argument wins.

Start-ups by their very nature can't have a large sales force that support adoption around the world. And really, it's during the adoption phase that the engineer makes the decision about what best-in-class tools will be used for emerging needs. They may not be aware or, or get support form, the emerging companies so they may go with whatever is available.

Rhines – In rapidly changing areas of technology, I believe the answer is no. No one can attach the wide diversity of specialty needs that exist within EDA in just one company. Hundreds of start-ups provide the opportunity to peruse specialty niches, and also non-obvious technical discontinuities. Now big companies often take a different approach. In the Mentor case, however, we identify specific emerging areas of new methodology or technical discontinuities, and choose the ones we think we can have the greatest impact on – and, we just don’t do things in those other areas.

There's another approach as well in EDA. That's to simply wait for the smaller companies to fight it out and then wait for a leader to emerge. Then you go and acquire that leader. Mentor hasn't benefited as much as others with that, but it certainly is a viable strategy. But the question is – can the big companies address all the specialty niches and all the discontinuities? I think the answer is definitely no.

Rhines – I'd say that start-ups need to attack the niches and the non-obvious discontinuities. And the customers/designers need to be creative and see EDA as a differentiator. I think that, in fact, continues to happen. So, we were told in late 2002 by some of our competitors that they'd be able to squeeze us through one-stop shopping – but exactly the opposite has happened. We've gained share almost every year, and have had compound growth metrics.

That's because we've gone for best in class, rather than one-stop shopping.

We don't have the strategy of providing money to small start-ups. We, ourselves, are in the business of hitting the discontinuities, and we've hit more than our share. We view start-ups as either complementary to our technology, or as a back-up strategy for failure to develop the technology in-house.

Rhines – There's probably some concern over the fact that the overall industry has not been growing much over the last few years. But you have to look within to find the growth areas. Anytime there's a new methodology, there's growth in the Bay Area. There are lots of start-ups today in DFM, system-level design, and IP – things like that. There's also a lot happening in platform-based design.

There's not that much growth in the traditional areas for start-ups – things like place and route, or traditional types of verification. Although there have always been attempts to stimulate a discontinuity there, or to react to one.

Rhines – I would say watch their feet and not their mouths. Watch where the VCs go with their money, and not what they say. The number of EDA start-ups increased last year, and the funding is growing. There was a weak period due to the excesses of 2000 and 2001, but I believe the number of start-ups in EDA is beginning to increase again.

One reason is that there's an exit strategy in EDA that does not include an IPO – one that's not always as obvious in other industries. In EDA, the economics of acquiring a company, and the ability to do it, works well. A start-up can start out with the strategy of being acquired, and provide good return on investment for the venture capitalist. You have to be compensated for the original risk, but it's not like 1999 or 2000 where people without a product or a business plan could sell themselves and get lots of money. The returns today can be good – historically EDA investments have been good.

***************************************

Pravin – There is no such thing as one-stop shopping for EDA customers because no single company can provide best in class technology for all the design challenges customers face. They will buy best in class tools irrespective which vendor provides it.

Pravin – One-stop shopping is definitely to the disadvantage of the customer. One-stop shopping appeals more to the finance guy rather than the designer. EDA vendors typically throw in their "sub-standard" tools for free during one-stop shopping. End of the day, the designer is stuck with a flow that is the "cheapest", but not necessarily the "best".

Pravin – One stop shopping is something big EDA vendors are trying to sell to the customers to lock them up and make them pay more over time. If you analyze any of the "volume purchase agreements" that the big EDA firms sign with their biggest customers, there will be products that have been "sold" cheap or "thrown free in the mix".

Pravin – One stop shopping does not discourage start-ups from entering the market. I can tell you from my personal experience at Sierra that customers are willing to adopt innovative solutions from startups and are willing to pay for it as long as you solve their design challenges. EDA customers know that major technology advances in EDA industry has mostly come from startups. And also remember the big EDA vendors were once a startup!

Pravin – History is proof to the fact that innovation happens in the small start-ups in EDA or any other industry. Innovation definitely happens in the big companies as well, and one thing they definitely have is ‘deep pockets’ but it is much slower, there are legacy issues that clouds their vision of what the new product needs to do, there is a worry about cannibalization – i.e. what will happen to my existing product line if I launch a new product, and than there is politics - between the new and the old product teams.

There is just too much baggage in the large companies to develop the cutting-edge solutions

Pravin – As always, leading edge customer need to work with startups to create cutting edge solution to solve their complex design challenges. Also, Venture capitalists need to continue to invest in innovative EDA startups.

Pravin – The Venture Capital industry’s age-old mantra is "Innovation Always Pays". They also know that every new generation of process node brings new complexities, which require innovative solutions in EDA. Startups will get traction and VC funding if they provide innovation.

***************************************

Q – Is there such a thing as one-stop shopping for customers of EDA tools?

Alberto – In principle there is. However, today there is no single vendor that offers the entire spectrum of tools needed to implement a complex flow even though two companies come to mind that come close to it: Cadence and Synopsys.

Alberto – From a support point of view, it is clear that a one-stop shopping situation is favorable for customers: one flow to train your designers on, one company to go to for bug fixing and additional features, one company to partner with at a deeper level than possible today. In some sense, one stop shopping enables a different level of relationship between customer and provider so that different business models can be followed that have more the flavor of a win-win relations than it is today where most of the times EDA vendors are kept at arm length and looked with somewhat of a suspicious attitude.

Of course, a one-stop shopping situation may end up in a dependency situation for customers. Second-sourcing is a common strategy in electronics companies to mitigate the risks associated with a single technology provider. Is second-sourcing a feasible avenue in EDA? What is a stable economic balance for the EDA sector? These are questions that have to be answered soon. I believe that the balance between advantages and disadvantages of a one-stop shopping is tilted towards the advantages if the EDA industry matures and can entertain a real partnership relationship with customers and if customers are ready to see EDA players as desirable partners to share risks and benefits.

Alberto – At this point, as I said earlier, the one-stop shopping solution is not fully possible but there is a definite possibility that this may take place in the near future.

Alberto – I believe that there are just not enough resources (people, skills, capital, management attention) for any company to blanket the EDA field with all possible innovations. Innovation will come from a combination of factors including young start-ups. The one stop shopping companies will make sure that whatever innovation is offered in the market can be easily integrated in their global offering so that it never becomes obsolete. For example, OA is a way of providing a command ground where young start-ups can develop their solutions without the need of spending precious technical resources building an infrastructure.

Alberto – I already pointed out that I do not believe that a one-stop shopping situation will discourage the existence of small companies or that the talent and bandwidth of established companies will be enough. It will create though a higher barrier to entry for some companies since the advantages over existing solutions will have to be non-marginal to be appealing. I actually believe this is a better situation than the one we have today where companies are born based on marginally interesting approaches.

Alberto – First, I would challenge the concept that cutting edge technology comes only from start-ups. In my experience with large companies like IBM, GE, BMW, HP and Bell Labs, the most striking and game changing technology came from either large companies who have enough resources to be able to invest in research with critical mass or from Universities. Start-ups are good transmission chains from revolutionary concepts and prototypes to commercially viable offerings. This commercialization aspect is not, in general, a strength of established companies for a number of structural reasons even though there are examples where this has been possible.

Thus, I see a number of scenarios:

1) Internal incubation from established companies to protect innovators from the day-to-day routine work that can absorb a very significant part of engineering time;

2) Spin-offs to favor quicker commercial introduction of new products;

3) Start-ups supported by established vendors in addition to external VCs and possibly customers.

4) Start-ups supported by external VCs and possibly customers.

Models 1, 2 and 3 have the advantage of potentially leveraging the sales channel of established companies to make the technology available. Model 4 has the advantage of full independence. (It was actually the model that led to founding of Cadence and Synopsys in the 1980s where customers teamed up with VCs to facilitate the path of important concepts into their design teams). In all cases, open domain tools and frameworks facilitate substantially the development of innovative technologies in EDA.

Alberto – I believe that the future environment, if it will take shape as I outlined above, will eliminate present difficulties and costs so that the VC investment could be much more leveraged and less risky. For example, having customers to participate in the capital structure provides substantial advantages in terms of feedback, guidance and lower adoption barriers.

Having established EDA companies as co-investors provides market experience, technical and managerial experience and potentially an already established sales channel. The new initiatives will also be of better quality since the technology they promote will have to be measurably and substantially better than what is available.

**********************

March 7, 2005

Peggy Aycinena owns and operates EDA Confidential. She can be reached at peggy@aycinena.com

Copyright (c) 2005, Peggy Aycinena. All rights reserved.