/ Systems
Automation as a compounding asset
Automation is not about doing less work. It is about paying for a piece of work once and having it keep returning value.
6 min read
Automate the second time, not the first
The first time a task is done, the process is still being discovered. Automating it locks in a guess. The second and third time reveal what is actually stable and what only looked stable.
We keep a plain list of things done more than once by hand. That list, not enthusiasm, decides what gets built.
Pick by frequency and blast radius
The best candidates are frequent, well defined, and low consequence when they fail. Release checklists, error reporting, transactional email, backups, uptime checks.
The worst candidates are rare and consequential. Automating a quarterly judgement call takes longer than doing it and fails in ways nobody notices until it matters.
- Frequent enough that the time saved is real
- Stable enough that the rules will not change next month
- Observable, so a silent failure becomes a loud one
- Reversible, so a bad run can be undone
Silent automation is a liability
The dangerous failure is not the one that throws an error. It is the one that quietly stops running and takes weeks to notice.
Anything automated needs a heartbeat: a log you can read, an alert when a run does not happen, and an obvious way to check that today's run did what it was supposed to.
Where the compounding actually comes from
The saved minutes are the small part. The real return is the removed decision — the thing nobody has to remember, prioritise, or feel guilty about.
Across a portfolio, that is the difference between operating several products and being operated by them.
Building something this applies to?
We take on a small number of outside builds each year — websites, apps, automations, and internal systems. Or look at what we build and operate.
Start a project