Skip to content
Lachlan Loveberry
Approach

Boring deploys are a moral position.

Confidence is a byproduct of small reversible steps, real-world tests, and observability you trust. Here's how I build it in.

The most expensive moment in any engagement is the one where the team stops trusting the deploy button. Once that trust goes, everything slows down: people batch changes, releases turn into events, “let’s just wait until Monday” becomes a strategy.

I aim for the opposite: a team that ships on a Friday afternoon and goes home.

What confidence is made of

Small, reversible steps. If a change can be rolled back in 30 seconds, you can ship it. If it can’t, we either reshape it until it can, or we plan the rollback before we plan the rollout.

Real-world tests. Tests that catch what the team is actually afraid of. Not 100% line coverage. Not test theatre. The handful of tests that wake people up if they break.

Observability that’s load-bearing. Logs you’d actually read, metrics you’d actually act on, alerts that fire only when someone should care. Anything else is noise that you’ll quietly start ignoring.

Migrations as a discipline. The Lasting SEO CallRail → Twilio migration shipped zero dropped calls because we treated cut-over as a craft, not a step in a Gantt chart. (Case study →)

What I do, specifically

  • Add a rollback section to every deploy plan. If it can’t be filled in, the deploy isn’t ready.
  • Ship behind feature flags by default for anything user-facing.
  • Run new code against real (sampled) production data before it serves real users.
  • Wire alerting before launch, not after the first incident.
  • Document the runbook as I build, so the team that inherits the system can keep it alive.

What confidence isn’t

It isn’t:

  • A massive QA gate that delays every release.
  • A 12-page rollout doc nobody reads.
  • “Run it in staging for a week and pray.”
  • A culture of blame when something breaks.

It’s the opposite of all of that — calm, deliberate engineering that treats reliability as a feature.

Ready when you are

Let’s see if there’s a fit.

A 30-minute call. I'll ask honest questions, and you'll get honest answers about whether I can help — and whether you should hire me at all.

Book a discovery call →