Bad Design (Thank you for the use of this image).

Lack of a design authority is one of the biggest maladies in (technology) engineering today

I’m in the process of (getting a) garage-workshop built. Before submitting a planning application, I drew up a design, and got a good idea of what I wanted. Compromise, limited budget, small garden, but there is a clear concise idea of what is required, the steps required to build the garage, and deliver a great workshop.

Deliver is the key.

There is no danger of an abstract heli-pad replacing it at the last moment.

I cannot believe that this situation is present in so many tech companies. People fail to define what they actually want. They will use terms like “fail fast”, “iterative prototyping”, and “abstract data model.”

Not being able to find the right people, lack of budget, lead to questionable decisions being made.

Spreading a team geographically all over the place, sub-contracting, or ….. This does not resolve the hot potato of what do you actually want? Time ticks by, the budget gets burned.

I have been faced with this situation three times professionally. I have been able to fix things twice.

Fixing involved me designing, pitching and implementing components, which were tweaked, and became the production norm. Because of the (manageable) scale, I was able to crack-on and get things done. I wonder if people realise the amount of blood, sweat, and teeth-gnashing which was involved. Fix is a highly subjective term.

The third one was huge, and involved ambiguity around the design, features, the data-model, and how, who, and what was going to be delivered. Intended cross customer, software as a service with real time, high volume, concurrent, events increased the complexity no end. This was at the exploration phase.

The lack of a rock-star CTO, team of engineers in the same location as the product team, caused a real engineering disconnect. Enough said.

Going back to my garage-workshop, I’m confident it will deliver, and be spot-on :)

You cannot boil the ocean.