Building a company inevitably involves creating processes to handle routine and non-routine tasks. And most processes suck. They suck because they are often designed with an idealised view of what the work is like.
Just like products, processes need to solve real problems. They are either adopted, or they die.
Trying to overhaul any process all at once poses two major threats:
- It takes more time to learn if it worked or not.
- It’s more difficult for your users (e.g. employees) to adopt it.
You can avoid some of these pitfalls by iterating your processes over time, just like you would with your product.
Document as-is, not as you want it to be.
If you’re already doing something but there is no formal process for it, don’t try to create a process around it and change it at the same time.
For one, you’ll be surprised how often how you think things are done and how they are actually done are different. It’s something you can only learn by monitoring your current process first before redesigning it.
Start as small as possible.
If you’re introducing an entirely new process, start with something simple. You’ll spend less time monitoring it and it will be adopted quicker. Simple processes are also easier to learn from, because they are easier to analyse.
Figure out where you want to go.
Define parameters of your new process—why you need it, what your desired outcome looks like and what your success criteria are.
Do your research and document your findings, but don’t spend too much time on implementation details. Always favour action over excessive planning when the cost of failure and course correction is low.
Apply small changes until you get there.
Now that you know what your current process looks like, and where you want to go with it, start by applying small incremental changes and measure what impact they have.
Keep your momentum.
Small iterative changes add up very quickly. Think of it like you do of compound interest. If you keep making small changes every week, after a year you’ll be farther away than you think.