Boldport

View Original

So you want to start an EDA business?

It's New Year's day. We look ahead. Some of you may be thinking today of starting an EDA business. I've been an EDA user all of my professional life and have tried and failed in starting an EDA "startup", and am now running another business in this space*.  Over this time I've had many realisations, ideas, conversations, highs and lows, all surrounding "EDA business". Since I find myself repeating some of what I've learned, I wanted to share my thoughts and opinions here.

Three points to start off with:

  1. EDA tools generally suck in most ways possible. Pretty much everyone agrees on this. There is a huge need for innovation on several fronts: optimisation, usability, team-based design, to name just a few. Lack of innovation on any of these will eventually grind the industry to a halt.
  2. The EDA business is dominated by a handful of well established, huge, resourceful corporations, with users that are generally conservative and bound by old-school ways and corporate constraints.
  3. As an industry we're very poor at exchanging information in an effective way.

All of what I'm about to say is a result of the points above.

Engineers are conservative

It's a gross generalisation, but when it comes to trying something new, we engineers are not as enthusiastic as our software dev counterparts. We tend to stick with how we're used to doing things, particularly as we progress with our career. It's quite sensible to do that because mistakes in hardware tend to be very expensive, and you're expected to get things right on the first go. This doesn't help innovation, though. (In a somewhat perverse way our affinity to tinkering may explain why we tolerate broken tools -- we like fixing them with clever elaborate scripts!)

Most engineers in big companies are possibly not even allowed to try anything new. That may be a sensible policy from a corporate point of view, but it too does not help with innovation.

Finally, it's rare that the engineer has any say in the tool choices. Those important decisions are the domain of "purchasing"! Doesn't help.

The lifestyle chasm

Over the years I've spoken to many EDA small business / startup owners. The picture that emerged is that unless you have a protectable innovation that is tackling a burning problem, get huge amounts of funding to develop it, and sell to one of a handful of companies, you will be sliding towards the "lifestyle chasm" and never climb out of it. Some of the people I spoke with were at it for more than five years, some got millions in funding and still shut down after that time. Some regretted sticking with it for that long. They had what I considered to be a great product.

10X is 1,000X hard

Say that you developed a product that everyone universally agrees that is 10X better in one important area compared to the nearest competition. Problem is that the EDA business, traditionally, is based on lock-in and reliance on high switching costs to stifle competition. That mandates a huge financial and educational commitment and reliance to and on a particular vendor, and their way of thinking and doing things. So now, despite having a recognised superior product (in one area), you have to overcome their investment in the "other" product. But, more crucially, the "other" product needs to only slightly improve in order to kill your 10X and prospects.

Prehistoric

While it's getting progressively better, we are still stuck somewhere in the 1980s in how we communicate data. Revision control is an emerging technology, designs are sent by email, Gerber and Excellon formats are still ubiquitous, exchange formats are proprietary and designed " by committee", and every component has a different format of a datasheet with no machine readable information. It's very difficult to innovate where the basics aren't even there.

"Big EDA" can be ruthless

The large players in this industry guard their entrenched positions well. They even discourage performance benchmarking in their EULAs. It's actually very hard to publicly prove that your product is better! As a small business, you can't compete with the lawyer power of those mythical beasts. Here it’s worth mentioning that purchasing cycles in EDA are loooooooong. If you're to survive, you will need a large reserve to cope with the one to two year it typically takes for a company to actually buy your product.

So you want to start an EDA business?

"Will this guy stop whining; it's hard to succeed in any industry" some of you are thinking. That's true. I'm not writing this to stop anyone from trying, though hearing about some of the more common pitfalls from an industry-specific view can help. If you're thinking of starting a business in the EDA space, consider the following:

  1. There's a lot of room for innovation. The temptation is huge. But be prepared for the best case for your idea to become a lifestyle business. If what you're doing doesn't seem like something you'd be happy doing for a "salary", then it might be worth re-thinking the concept.
  2. Don't plan on converting all engineers. Some engineers will never try something new. Some engineers will not even think that there is a problem with their broken tools since the tool has abused them for so long they are blind due to the Digital Stockholm Syndrome. Identify early who will and is able to use your tools.
  3. Unless you have a protectable and significant innovation, don't plan to be bought out by Big EDA. Try charting a different path that will free you from the constraints that EDA is infamous for.
  4. Understand engineering culture. Understand the EDA business culture. Do not require those to change significantly as a requirement for your company's success. If anything, bypass them!

If you're an hardware developer craving for better tools, please support new projects by trying them out, providing feedback, or simply saying a good word. It means a lot.

* My first product was a web-based automated FPGA build management tool called 'boldport flow'. Good concept; poor execution. I learned a lot about the EDA business in the process. After a year of doing that I didn't enjoy it enough to continue doing it as the inevitable lifestyle business, so I pulled the plug and moved on.

I now run Boldport and created an open source PCB design tool called PCBmodE. While Boldport can be considered an "EDA business" by some, I certainly do not. I try my best to not live up to their reputation in some areas. If Boldport succeeds, we'll have to find a new acronym for that kind of business. I love what I do now, and if the worst that can happen is that it becomes a lifestyle business, I'm fine with that even though I'm aiming much higher.