If you google “7 deadly sins of startups” you get lots of hits on mistakes for social networking ventures, only a few of which are relevant to EDA startups. In EDA you have to demonstrate real growth quickly with a very tech-savvy audience in a handful of bluechip accounts. So throw away the research you did on the web because it isn’t going to help. Success is still possible though if you avoid some of the following blunders. I’ve committed a few myself and have seen up close the anguish of the others.
1.The “me-too” product
It’s amazing how often would-be entrepreneurs look at a successful product in the industry and think “hey, I could do that too”. This is a huge warning sign that you don’t have an original idea and that your differentiation will at most be in quality of result. The first problem is obvious. The second is a delusion unless you can prove 10X superiority over market leaders. That requires a fundamentally different approach, (aka an original idea). Why isn’t 3X good enough? Because everyone knows at least some of that advantage will evaporate in a production environment and 1.5X – 2X isn’t enough to justify a switch from a leader who’ll almost certainly catch up and who already provides bundled pricing and local support. (And don’t even think you’re going to compete on cost, unless you’re a Walton.)
If you don’t have an original idea, think harder or find a partner who does have an original idea.
2.The “one-customer” product
This is deceptively easy to fall into but equally dangerous. You have a good contact at a major account, you build a custom product and they’re happy with the result. They swear everyone in the industry is going to need this. But any single customer is a terrible judge of the general market because they’ve seen very few flows. What you built is unlikely to translate to another design team, much less another account. Above all, do not get dragged into “integration products” (build me a flow). These never translate, even within a company.
Service companies often fall into this trap. Some products developed in services make the transition to general application, but not many.
Check with multiple prospects in multiple companies and dig deep before you decide you’ve found a trend. Don’t cling to the first positive thing you hear as validation – make sure you get the whole response, warts and all.
3.The “nice-to-have” product
You found general agreement that your product would be useful but is it essential? Is solving this problem their organization’s priority 1 or 2? Unless you hear a resounding YES, they’re grading this a nice-to-have product, which they’ll never buy. Avoid building stuff to improve productivity. Vitamins are always nice-to-have. And avoid solving short-term problems. If the big guys are promising a solution in a year, forget it (sometimes they don’t come through, but the risk is still too high).
Get multiple independent viewpoints; the majority should grade this as must-have. Don’t be afraid to ask about solutions from competitors. Disqualifying bad ideas quickly is a good thing and you may uncover new ideas.
4.The “one license per account” product
OK, so you found a product that has broad appeal and is an absolute must-have. But only one engineer in the whole company needs it and then maybe only for a couple of months each year. Your total available market just shrank to maybe 40 licenses. Design companies are prepared to pay up to a few tens of $K for an EDA product, even if it cures cancer. That may give you an interesting lifestyle company, but an exit is out of reach given lack of room to grow.
A product has to be able to grow to at least multiple 10’s of seats if not 100’s of seats per account. Check that your customers agree your product could do this. (Engineers are often scared to ask customers this kind of question. Don’t be. Reasonable customers expect well-run startups to size the opportunity.)
5.The “revolutionary as long as you radically change your design flow” product
This was one of my sins – I built a product around IP-XACT design creation. Looked good because IP-XACT is an industry standard and several other companies were already on-board. But for others adoption required ditching RTL-based design at the SoC level and switching to IP-XACT-based design. That’s a huge change. Design teams look for incremental flow changes; tight schedules make major changes too risky. There are exceptions (when a team can start from scratch) but you can’t build a viable business on exceptions.
Run, don’t walk away from anything that requires a radical shift in methodology. Even SystemVerilog allowed for incremental adoption.
6. Uncritical love for your product
Passion for your creation is essential but it can also blind you to problems, especially deep problems in translating the concept to user needs. You can’t rely on your team or prospects to tell you your baby is ugly; your team wouldn’t dare and prospects would rather avoid conflict (at least before they have bought). Absent critical feedback you fail to see the need for major course corrections and your business mysteriously evaporates.
You need to decide early on if you want to be right or you want to be successful. You probably can’t have both. Recruit an independent advisor whose job is to tell you what you don’t want to hear and can contribute to discussions on corrective action.
7. Confusing cost with opportunity cost
It’s important to be frugal when running a startup, but there are limits. Reinventing commodity components (netlist/RTL readers and GUIs are two obvious examples) will tie up resources for much longer than you think and will ultimately cost more in slower progress on real differentiators and lost goodwill than you would have spent on off-the-shelf solutions
Never forget you are in a race to demonstrate differentiated product traction with key customers within 5 years. Wasting any of this effort on reinventing commodity functionality you could have bought (or partnered to get) is crazy.
It’s possible you could defy one or more of these guidelines and still have a winner. But then you’re betting against the house. Maybe you’ll win, but I doubt you’ll find investors.