Method How I work

Routine runs on rails. Moments get a human.

Two modes. The system owns the routine: data pulls, signal checks, segment refreshes, guardrails. I own the moments: the launch, the breakage, the budget call, the creative bet. You pay senior rates for the second list, never the first.

01 The operating rhythm

23:05

Pipeline night shift

Shopify, Meta, Google and Klaviyo data lands for every brand. Catalogue jobs, segment refreshes and scheduled builds run while nobody waits on them.

06:00

Signals and verdicts

Eight signals scored per brand. Each account gets a colour and a reason. Anything amber or red carries a recommendation specific enough to act on before the first coffee.

06:20

Operator review

I read every verdict and act on the ones that need a human: budget moves, creative rotations, escalations. Routine mode ends here; judgment starts.

DAY

Moment mode

Launches, fixes, experiments, builds. The reading finished at six, so execution starts before nine. That's the compressed decision cycle in practice.

02 Working rules

The rules I won't bend.

R1

Evidence before opinion

A recommendation needs at least two comparable situations behind it. If I haven't seen it work twice, I'll say I'm guessing.

R2

Numbers over adjectives

Where a real number exists, it does the talking. If a report ever says “strong performance” with no figure attached, bin it.

R3

Staging before live

Nothing ships to a live store, account or flow without a staged version and a check. The speed comes from the system; the checks stay.

R4

Your account, your data

Everything is built in accounts you own. If we part ways, you keep the infrastructure, the data and the documentation.

03 Engagement shapes

Three ways in.

Sprint: one problem, fixed scopefrom $2,500
AI build: a system, deployed$2,500–$7,500
Retainer: the whole enginefrom $4,500/mo

Most clients start with a sprint or a build, then move to a retainer once the working style proves itself.