Second of three parts: Integrating and aligning companies, what works and why.
By Ed Sperling
Buying companies is the easy part. Integrating them is the hard part. It’s also the point where most acquisitions that go awry actually run into problems.
There are widely different strategies for how to accomplish integration. Sometimes they work, other times they don’t. And sometimes both companies are surprised by the outcome—for better or worse.
“Either you think they’re going to work out and they don’t, or they work out vastly better than you would have thought,” said Wally Rhines, chairman and CEO of Mentor Graphics. “The biggest thing that can go wrong is when we don’t add the value that we think we’re going to add. So why are we valuable to a company? Frequently it’s the increased sales bandwidth. If we’re unable to add value, that’s a frequent cost of failure. Occasionally we have a cultural incompatibility, but that’s pretty rare. And you can tell who’s out to make a bunch of money and who’s going to stick around. You usually have a good idea of who’s motivated.”
As with any aspect of business, timing is everything. Joe Costello, the former CEO of Cadence, was famous for waiting for startups to seed products in the market—but just before the revenue stream started flowing in. He then paid what to most observers looked like high premiums for companies, acquired some highly motivated people, and the money started flowing in soon afterward. Those kinds of opportunities are much more difficult to find these days because there are fewer startups and more savvy investors. But that hasn’t stopped the acquisition frenzy in EDA. And despite decades of experience at this, each acquisition brings new challenges.
“You do become more efficient at doing this,” said Aart de Geus, co-chairman and CEO of Synopsys. “The faster you integrate companies, the better. You want to meet the team within 30 days, and you want a proposed product direction within 90 days. You still make mistakes that way, but you avoid more mistakes by moving quickly.”
De Geus noted that processes are more defined in companies that do it well. “EDA is not a special case, but in EDA the rate of change is so high that if the acquiring company is not willing to change it’s difficult to integrate. There is skill in everything. The key is to recognize mistakes quickly.”
That’s one approach, and it has worked well for Synopsys. Cadence takes an almost opposite approach—slowly and systematically assessing and digesting companies. Other companies tend to fall somewhere in between—taking time to understand and build teams on consistent processes while quickly cutting out redundant operations such as accounting and human resources.
Where the acquisitions come from
In the past, most of the acquisitions focused on startups. While startups still exist—there are dozens of EDA startups scattered around the globe, many still in stealth mode. But there also has been a shift in how those startups are funded. A sizeable portion are bootstrapped, while the bulk of the rest are funded by angel investors. Most VCs have stepped back from EDA, because the exit strategy is limited to either organic growth or acquisition. The big money is in software and social media, where companies can file IPOs.
“I worry about irrational business expectations,” said Kathryn Kranen, president and CEO of Jasper Design Automation. “Angels are not as rigorous in their investment strategies, and there is going to be a lot of infant mortality.”
One sign of this is the size of companies that large EDA vendors have been acquiring lately. Most of them are larger and established, such Synopsys’ acquisitions of Eve and Springsoft and Cadence’s acquisition of Tensilica. That makes it much harder to assess what’s happening in the rest of the market, as well as how that might affect future acquisitions.
“If you acquire a company of critical mass you can integrate it, but it’s a lot of work. And how do you do that with five companies? You also have to look at what affects valuation of public and private companies. There are two main factors. One is how fast they’re growing. The other is negative issues. Profit is only a tiny influence in all of this, because if you have a choice between being more profitable or growing faster, you’re probably going to grow faster.”
Cadence has added a new wrinkle into the process with IP acquisitions—understanding what kinds of new liabilities it is taking on.
“When larger companies acquire smaller companies, they don’t always know if the companies they’re acquiring know about all the export regulations,” said Larry Diesenhof, group director of export compliance and government relations at Cadence. “The acquiring companies can be fined hundreds of thousands of dollars for violations that occurred in the past.”
For EDA acquisitions, these kinds of issues are non-existing. But IP, as with software, falls under restrictions about what can be shipped where. Diesenhof, whose expertise is international trade regulations, is now brought into all IP acquisitions by Cadence.
Where we are now
Still, EDA has as many acquisitions under its belt as any industry, and in many cases far more. And put in perspective, the industry’s track record is very, very good. At least part of this has to do with the highly focused mission of some of the startups.
“The number of individuals that build a company, get acquired, leave and start another company is significant,” said Jack Harding, president and CEO of eSilicon. “They’re now on their second and third companies with a pre-determined fit into another company. They go off with semi-perfect knowledge of what’s needed by a specific company.”
He said that for a company to be attractive to the general market, it has to offer a 10x performance breakthrough. But to be attractive to a specific company, it simply has to fill a void in the tools lineup.
That view is echoed by Jim Hogan, a long-time EDA venture capitalist. “Would you start a new place and route company? No. It takes a good five to six years to have good technology, and another two or three years to get enough channel pressure to build momentum. That’s also a minimum $10 million investment. But if you have a new algorithm for power formats, you can build that for $3 million to $4 million and be profitable in five to six years and then sell the company for $25 million to $30 million. If you invest $10 million, you have to sell that for $60 million.”
But there are exceptions, as well. Hogan said Atoptech anticipated what would happen in place and route at 20nm and below and delivered a good solution that is gaining traction.
Moving to IP
After years of hype, another area that seems to be gaining traction for investment is IP. Many of the startups acquired over the past couple years are on the IP side, largely in reaction to the growing complexity of SoCs and the increased amount of third-party IP being used inside those chips. The percentage of externally developed IP inside of is expected to steadily increase for advanced nodes and in stacked die, as well as for platform-based designs for the Internet of Things, making this market particularly attractive for startups and companies looking to expand their IP portfolios.
“We’re seeing a lot more startups in the IP business than in classic EDA,” said Andrew Yang, president of Apache Design Inc. “The cost of chips is high and the IP business is definitely growing, but there’s a difference. Before, you could do a startup and create a narrow solution. Now, you need a broader skill set and the problem you identify has to be very crisply defined. The opportunities are still there for startups, but the execution is a lot harder.”
Coming in part III: Measuring success.
To read part I, click here.