Design From Day One
Startups are created with a unique domain perspective and with the idea that a particular market is not meeting an overlapping need. During the early days, when you haven't defined the solution deeply, if at all, the risk of missing the eventual target is very high. Enthusiasm usually leads to building straight away and figuring things out as you go. A considerable part of the initial sunk product costs arises from not having spent the minimal amount of time defining what the pilot product should be.
Discovery and prototyping are much more accessible, more efficient, and less costly to do before development has started. It is crucial to have somebody on the team that can transform domain expertise into a coherent user-facing experience that can purposefully put together elements to ensure that the product enables a particular user outcome coherently. It does not matter if they are a designer or not, internal or external. It all boils down to having the skillset available during those critical times. If you are in this situation now, you should solve this problem yesterday. Waiting will only make it more costly to fix later, diminishing your chances of success. Remember designing something and getting feedback on those artifacts is much cheaper than the build-rinse-repeat cycle.
Strike A Balance Between The Short And Long Term
If the design is not ahead of development by at least a couple of weeks, you run the risk of running out of the critical product definition runway required to align the team towards a common goal. On the other hand, if your design is projected too much in the future, you might lose the focus on the now, potentially making that future impossible, or even losing sight of vital signals that could change your direction.
The design needs to be in a consistent state of tension, solidly anchored in the now while eyeing the future and extrapolating its materialization every time new relevant information becomes available.
Stop Thinking You Are A Good Representation Of All Your Users
The pitfall of being a domain expert is that false consensus can set in, and along with confirmation bias, it might make you think that you are a good representation of your entire user base. In the early days, this might be true, you need to start from somewhere, and your own painful experiences with the problem you are trying to solve will generate a lot of valuable insight. As you advance, though, being able to bring in customers is crucial, whether it's through guerrilla research, paper prototyping, alpha programs, or customer advisory boards, nothing beats canvassing the market for directional signal. Remember, you are but one slice of your potential customer base.