Dialogue - Simon Davidmann

EDA startups and growth

by Bipin Parmar

Following on from The Chilli's profile of successful entrepreneur Simon Davidmann, we talk to him on a range of issues related to the EDA industry, startups, and some of the issues in growing a company in this industry sector.

Simon Davidmann

The questions:

1. You have been involved in many mergers and acquisitions of EDA (electronic design automation) software startup companies. What important lessons can you draw from this?

2. Were these companies acquired for their patents, revenues or technical know how?

3. Who ends up making the most from these M & A activities?

4. In the companies you have been involved with, what were the stock options like, and did they materialise into tangible benefits?

5. When you co founded Co-Design Automation, what was the main motivation? Was it frustration with established players? The need to solve a future problem? Or just running your own show?

6. Co-Design had staff in multiple locations across UK, US East, US West and Germany? Wasn't this too risky for a startup, in terms of communications and limited team resources?

7. Getting Superlog adopted by Accelera, was quite a coup for a small British EDA startup: what were the main factors behind this?

8. When Co-Design donated their Superlog language to a trade body, how was it going to make money, what was the business model?

9. What is your advice for European based EDA companies?

You have been involved in many mergers and acquisitions of EDA (electronic design automation) software startup companies. What important lessons can you draw from this?

Firstly, big EDA companies work to a different scale from small companies and they operate with more rigid internal structures and processes. Their primary task is to service existing customers with incremental technology, which sometimes acts as a brake on innovation. In other words, they carry too much legacy, which sometimes holds them back. Even if they could see a niche opportunity, it would be difficult for them to justify the initial effort, especially if a new technology may generate sub $10 million per year in revenues. This may appear to them as a noise figure in their $100M's of overall revenues, but for a startup, this level of revenue is an opportunity and could also mean the difference between survival and failure.

Entrepreneurs are always on the lookout for newer, more efficient methods and better ways of doing things. They move quickly and take more risks. In addition, startups have no legacy products or customers to hold back new ideas or methods. This allows them to establish a new technology or validate a new market need by quickly rolling out the product or service and by flushing it out with a few leading edge customers. Chronologic's VCS Verilog Simulator and Ambit's BuildGates synthesis tools are good examples of this. Magma set out a bold vision and actually managed to achieve it, with its integrated timing, logic synthesis and place/route technology around its unified data model.

Were these companies acquired for their patents, revenues or technical know how?

Once a startup establishes a new product, technology, or methodology with some of the leading edge customers, the pressure mounts on established EDA vendors to work with these startups as all tools have to interoperate together. Together with synergies in using an existing customer support, sales and application network, this makes a merger and acquisition route more compelling.

Established EDA vendors have a fairly high sales and support cost (the channel costs), which they need to feed with more products and services, to reduce the effective cost of sales. In the EDA industry, acquisition of startups for new products is quite normal, whereas acquisition for patents is rare, and in fact some of the larger EDA companies believe that they must own every piece of the design tool chain. I believe this is misguided as many customers would very often choose a best in class tool from various vendors.

Acquiring startups because of their knowledge, skills, or people is also very important for established EDA vendors - and often an acquisition is made just because the larger company wants the team.

Revenue is also important, in that it validates the need for the product or technology, but is not the main driver. A few EDA startups do go IPO, but this is not the norm. The norm is that a good EDA startup is much more likely to be acquired. Very few EDA companies actually fail and provide no return to their investors - most at least provide the capital back to their investors, with maybe less than 5% returning nothing.

Who ends up making the most from these M & A activities?

The VCs that take big risks do well out of EDA M&A activities and so they should. They expect to make 10x if they invested in a startup at an early stage (Chilli S1, S2, S3, and R1) and 5x, if they invested at a later stage (Chilli R1, R2). Typically 60 to 70 % of the purchase price goes to outside investors and 20 to 30 % goes to founders and employees. Normally individual founders can maintain between 5 to 12% of the value of the company - if they stay the course from founding to acquisition.

One important lesson is not to raise too much money, as you will dilute your shareholding considerably and your motivation to succeed will suffer. Good VCs know this and can be quite reasonable.

A word of caution, though: timing is key. If you are still a startup after five years and you haven't established a good base of 12 to 30 customers and an OK revenue stream, then your valuation is going to tank. So the key is to get as many customers as possible, as early as possible. And they must be really using your product and be telling their friends, partners, colleagues about you. You have to establish your beachhead. And you must get some revenues.

In the companies you have been involved with, what were the stock options like, and did they materialise into tangible benefits?

In the beginning, people were new to stock options and they were not seen as potentially valuable, but as they became wiser and more educated, they started striking some hard bargains and getting more options. If you get involved with a startup early enough, then stock options can give out a reasonable risk reward, and a lot of people make handsome gains from them.

Care needs to be taken at the acquisition though because if you do a stock swap as part of the deal you may find that your acquirer's value drops before you have time to cash your stock and you may lose a lot of your potential paper gain. Sometimes it is better to get as much cash and bonus portion as you can. I believe at an acquisition you should shy away from all-stock deals. It is a personal choice, and it depends on the individual's risk profile. Some people don't need too much cash, others do. Also, remember that when you get acquired, they don't need a new CxO, as they already have them, so it is more than likely that all non-technical people will be allowed to leave and cash in early.

When you co founded Co-Design Automation, what was the main motivation? Was it frustration with established players? The need to solve a future problem? Or just running your own show?

The major motivator was our passion to solve customer problems (and of course the potential for making lots of money). Customers were frustrated, as they were looking for a better way of co-simulating both hardware and embedded software. We had a grand vision of unifying this divide. We also wanted to bring in the system architects and the verification team. This 'unification' theme laid the ground work for Co-Design Automation. You need to be prepared for things not always going according to plans and you have to listen to customers, be nimble and flexible and to change with market needs. In Co-Design our customers started using our tools and technology more for verification and focused us into the verification area. This gave us some initial revenue traction and did wonders for our reputation in the market place.

Co-Design had staff in multiple locations across UK, US East, US West and Germany? Wasn't this too risky for a startup, in terms of communications and limited team resources?

We knew a lot of brilliant, talented people who lived around Oxford, Silicon Valley, Boston and in Germany. These were all mature people with many years experience in simulation and in the EDA industry. (The Co-Design team had an average of 18 years of EDA experience of which the average was 15 years in HDL simulation.) Most had worked together in the past, and had respect for each other and knew their abilities and weaknesses, which helped a great deal. We needed the best, so we devised a strategy and plans that allowed everyone to work remotely and meet regularly, to sort out any underlying issues. We made heavy use of technology to ensure effective communications, work partitions and deliveries. Of course as a CEO, I was always on a plane and jet lagged so that while back in the UK I didn't have problems staying up all night on the phone and doing emails, managing a global company. We also made good use of cheaper conference calls and facilities. We really did exploit new information technologies to enable our global fledgling organization to be efficient.

Getting Superlog adopted by Accelera, was quite a coup for a small British EDA startup: what were the main factors behind this?

We started out with a goal to build a better HDL simulator but found that there was an anomaly. We needed a new language to describe the new designs and then realised that to be successful we needed to get it widely adapted. Everyone said this was too risky and it wouldn't happen. To be successful, we had to fight battles on two fronts; firstly on the product side, which we called the "ground war" and then also with the marketing, campaigning, lobbying, evangelising, meeting analysts, journalists, and cajoling everyone's help. This we called the "air war". Most marketing people in EDA come from a product marketing background, which is because they have come up the ranks from the technical applications side , but air war marketing requires a different skill set, and we were fortunate enough to hire a guy who became a very seasoned marketing general and who devised a good strategy and plan to execute the strategy, with our limited resources.

I believe that if an EDA startup fails, it is very likely to be because they don't spend enough time , effort and resources on marketing. EDA is technology for technologists and so marketing alone is not enough - and technology alone also is not enough - you need good technology with a good understanding of a customer's real need.

When Co-Design donated their Superlog language to a trade body, how was it going to make money, what was the business model?

We had a lot of internal discussions and soul searching. If we donated our language, how were we going to make money? If we made it into a public domain language, everybody would copy it and why wouldn't a competitor take it and make a better simulator from it? Time was critical, as new solutions were being developed all the time. There were lessons to be learned from Verilog, where initially it started to lose market share, and once Cadence opened up the language it then regained market share. There are also lessons to be learnt from the business model of Artisan, who 'give away' their libraries, and finally make money when their users tape out (when designs are taped out to foundries). In Co-Design we started looking at where the money was flowing to and reached a conclusion that if our language got adopted we could sell tools and products around the language.

What is your advice for European based EDA companies?

I have noticed that The Chilli has profiled many interesting European based EDA software startups and many of your biggest fans/readers are in the US. European EDA companies have to remember that the vast majority of their market is in Silicon Valley, so you have to be physically there, fairly early on. You can split the operation into two, with engineering and operations based in Europe (from a US perspective they are 'offshore') and sales, marketing and support out in the US. EDA is a US based business, so you have to think and act like them. You should incorporate in the US as soon as possible, you should also have a 408 (telephone area code for the valley) or 650 number on your business card or people won't call you. You should also consider getting US based VC investors, who have local knowledge and connections; they can also assist you in hiring key staff with local experience of US based EDA companies. In Europe, you have to travel for a day to see just one customer, but in the valley, up to 10 potential customers can be met with in one day - there is more opportunity in an hour's drive than in most European countries put together. . .

The challenge for an efficient back end RTL to GDS II flow is still going to be there for a while and there is room for many startups to provide a 'better mousetrap' in this domain. The risks here are relatively low, and getting acquired is more likely.

The domain above RTL is very complex, unstructured and risky... But I believe that in the next 5-10 years as Moore's law terminates, the activity in the back end of the chip design process will stabilise, with all the action and value in EDA moving to the front end.

The front end of EDA will be the place to be in 5+ years time and it will be a new generation of company that leads it.

Comments on this story? Send an e-mail to Editor@TheChilli.com

© Chilli Publishing Ltd 2004

02JUNE2004