• business
• model
• team

I have come across two types of people: the ones who enjoy building from the ground up, and the ones who have an urge for getting things done.

Myself, I know I belong to the first group. I derive a lot of pleasure from learning how to do new things, from electronics design to software development.

Startups run on limited resources, with a ticking clock that marks the moment money will run out. In that perspective, founders need to strike a balance between how much they can do in-house and the gain from outsourcing. And balancing how fast they can deliver their solution with how fast they can iterate on the offering.

When I started Dispertech, I was confident I could build a minimum viable product without relying too much on external help. It took a bit of time to learn some missing skills like electronics design and 3D modeling, but in the end, I managed to get something going with minimal costs.

Having full control in-house meant we could very quickly iterate after collecting user feedback.

The chances of hitting the mark on your first try are almost nonexistent. Iterating is crucial for success.

On the other hand, many startups don't have enough internal resources to develop a prototype by themselves. After all, complex instrumentation requires a broad range of skills. Hiring people to cover those gaps may simply be too expensive for the available budget. In those cases, outsourcing product development is a very reasonable approach.

However, outsourcing is not a magic bullet. Product development agencies may lack insights into what it means testing with users, and the level of uncertainty scientific equipment inherently has. They are used to working on fully specified products, with tolerances and risks.

When I built the first microscope, I wasn't sure about the appropriate magnification.

The optimal would depend on what the customer wanted to do, as well as on the camera we would use, and the quality of the optics.

To be honest, I didn't think about these degrees of freedom when designing the instrument. I didn't consider that perhaps lenses should have changed depending on the application. However, since the device was fully developed in-house, adapting it as we would learn from customers took hours instead of weeks.

Although the idea of developing in-house can easily be romanticized, it's important to keep in mind that manufacturing was never done by us. We outsource all the machining to a trusted supplier. There was no way to set up a fine mechanical shop by ourselves just to create some prototypes.

Other products may be impossible to develop in-house. Sometimes they require some very specific skills or tools that force the outsourcing.

Striking a balance between doing it yourself and outsourcing is a multi-dimensional problem that needs to be weighed carefully.