John Cooley has published part 1 of his market survey of people on the ESNUG mailing list. The results are interesting, yet not surprising, unless you happen to be mostly stationed on another planet or singularly biased toward your particular EDA employer. There are 11 items (sections) in the report. Two of them have mostly editorial content where John attempts to validate the method and results of his survey. Four deal mostly with customer support issues, leaving five that cover comparisons among various EDA tools.
The first section (Mindshare vs. Marketshare) is another attempt by John to defend the value of the results. I have written about this before, although not on this web page. John insists in using the word "scientific" to characterize the method and results. Once again I disagree with the use of the term, since actuaries would find the sample not representative of the overall EDA market. But I must say that, given the sample, the results are accurate. Those who object clearly have not read, or not understood, Section 1 of the report. In it John states that the sample came from people whose contact information was in the ESNUG mailing list. This list is likely to be biased in favor of Synopsys since it should contain all of the members of the Synopsys users community in addition to a lesser number of onlookers, detractors, and disrupters.
There are approximately 25,000 names on the list. John received 218 responses. Readers have no way of knowing how many of the respondents are actually actively using or have used Synopsys tools. Clearly the respondents are people that felt it was important to make their opinions known, thus I agree with John's statement that the results show "Mindshare" and not "Marketshare". Whether the results should be important to marketing and sales organization in EDA companies, or whether prospective buyers should be influenced by these results, is not a decision based on scientific methods resulting in a binary choice. As with most things human, the results give a flavor of the market and should be used as indicators of trends. It is a fact, though, that the data cannot be gathered easily in any other form, so the survey is useful and informative.
Using the data from the sections dealing with customer support, I concluded that customer satisfaction is eroding a bit, but not enough to be an issue. As problems get more complicated, and schedule's pressures mount, it is understandable that customers would feel a more urgent need for solutions from EDA vendors. Miracles in EDA are rare, the need for them is increasing, thus a few more people feel support is slipping.
It is also interesting that as the four EDA companies, Cadence in particular, are increasing the marketing efforts through internally controlled channels, like customers events, the satisfaction of the participants is not increasing, in fact it is somewhat decreasing. Another interesting point is John's admission that his sample did not yield a significant overlap between Mentor's U2U attendees and attendees of the other three customers events. This is an indication that the sample is heavily populated by ASIC designers and that the Mentor acquisition of Sierra has yet to percolate into the ASIC design market.
I also find unacceptable that John is forced to defend the results of the distribution of use of specific tools. Anyone with the ability to reason and who pays attention to the EDA industry could have predicted the trend, although probably not with the accuracy shown by the survey. That market leading tools like Design Compiler, First Encounter, or Calibre are loosing some "Mindshare" is natural. Four factors work to make the trend a reality:
- if you are number one competition will take some market share away from you
- you do not employ all of the creative engineers available in the universe and thus your competitors are capable of creating useful tools
- given the complexity of problems to be solved, companies are prone to try alternative approaches to the same problem, hoping for a "breakthrough", and finally,
- for any given design, there is a probability that a given tool will work better than any other.
John's detractors may find his contributions to the industry folkloric, but they cannot find them irrelevant.