I talked earlier about how it seems to take $6M to build a channel in EDA once you get to the “just add water” stage where all you need to do is to ramp up a salesforce and distribution. However, typically you are not really ready for this when you first think you are. More EDA (and other) companies are killed by premature scaling than anything else. Ramping up a channel is very expensive and will burn a lot of money very fast for little return if the product is not ready, either killing the company completely or requiring an additional unexpected round of funding at an unimpressive valuation, diluting everyone’s stock significantly.
When to scale is the most difficult decision a startup CEO faces. Too early and the company dies due to lack of financial runway. Too late and the company risks either missing the market window or losing out to competition during the land-grab phase of a new market.
There are two ways the product can be “not ready,” and there will usually be a mixture of the two. The first, and most obvious, is that the first release won’t include every feature that every user requires and so isn’t ready to serve the entire market. This is probably even known and acknowledged; it’s not as if engineering doesn’t have a long list of stuff for version 2.
The more dangerous way the product is “not ready” is that you are not completely sure precisely which is the most important problem that it solves for the user, and which subsets of users will value this the most. Value it enough to consider engaging with you, an unproven startup with an immature buggy first release. For instance, you might have a product that you think serves the entire market, everyone will need one, you can’t do 45nm designs without it. In fact, if you are just starting to engage with real customers, you might never have run a real 45nm design through your product, just doubled up 65nm designs and switched the library or something. It is often easier to get great results on those older designs. For example, in the last company where I worked, Envis, we got great power reduction results on 130nm designs, and public domain cores that had been around for even longer. After all, nobody cared that much about power back then so they didn’t put much effort into designing to keep power under control. When we tried our tool on 90nm and 65nm designs, the results were initially less impressive. Designers had already done many obvious things by hand meaning that we had to work harder to produce compelling incremental savings.
The reality is that your initial product certainly doesn’t serve the whole market. If it does, you should have gone to market earlier with a less complete product. Worse, the precise feature set you have implemented might not serve any submarket completely either. But you need to have a product that serves at least some of the market 100%, even at the cost of being useless to a large part of the market, as compared to a solution that is 90% for everyone. At least in the first case there is at least one customer who might buy the tool; in the second case, nobody is going to buy the tool, everyone is going to wait for you to add the remaining 10%. A different 10%, perhaps, for every customer.
It is a fallacy to think that taking a product to market is a linear process. Do the engineering, prepare sales collateral, start selling. Taking a product to market is more of an iterative exploratory process. There is a phrase, “throwing mud against the wall to see what sticks” that sounds derogatory. But, in fact, the early stage of going to market should be like that. In the best of all worlds you’ll have had one or two customer partners since the early stages of development, helping you spec the product and helping guide the early parts of development. But it doesn’t always work out that way. Sometimes you don’t have early partners, or your champion leaves, or sometimes those early partners turn out to be atypical in some important way, so that you are forced to choose between satisfying their unique requirements and developing features with wider applicability.
That leaves you with a product that you have to explore how to take to market, and to explore which if the many aspects of the product you should emphasize in your positioning. This is the City Slickers marketing problem, discovering the one thing that your customers value enough to buy the product and focusing your marketing and engineering on making that value proposition strong before worrying about other aspects of the product that might broaden the appeal to a larger segment of the whole market.