Dubstream
← All field notes

Field notes · Automation

We automated the consultant's Tuesday.

2026-06-12 · Dubstream · 4 min read

Watch a good Amazon PPC consultant work for a week and you'll notice something: most of the billable hours go to the same loop. Download the search-term report. Find the terms with spend and no orders. Negate them. Nudge bids on what's converting. Check that no campaign is pacing past its budget. Scan for low-stock SKUs so ads don't pay for clicks that can't convert. Compile it into a client report.

It's skilled work the first hundred times. After that it's a checklist with a login. And a checklist with a login is a job description for software.

The ritual, itemized

Here is the consultant's weekly loop, next to the rule that replaces it. Every one of these runs in production today on accounts we manage:

  • Search-term cleanup → a negation rule runs hourly: any term over a spend threshold with zero orders gets queued for negation, with the full list in the audit log.
  • Bid maintenance → bid rules keyed to your ACOS target raise winners and cut losers daily, instead of whenever the calendar says Tuesday.
  • Budget babysitting → a budget guard checks pacing every few minutes and caps runaway campaigns the hour they run away, not at the weekly review.
  • Overnight waste → dayparting drops bids in the dead hours and restores them in the morning, every single day, without anyone remembering to.
  • Stockout collisions → an inventory check flags SKUs under a cover threshold so spend pauses before the listing goes dark.
  • The client report → a daily audit compiles store, PPC and listing findings automatically — the report writes itself and the humans discuss what to do about it.

On the largest account we run this way, roughly 177,000 keywords sit under continuous watch. No human reviews 177,000 of anything weekly. The choice was never "consultant vs. software" — it was "software vs. not actually doing it."

What's left for the humans

This is not a story about consultants disappearing. The ritual disappears; the judgment stays. Which products to push in Q4, how to price a launch, when a category is worth entering, what the creative should say — no rule engine answers those. The consultants who win from here are the ones who stop selling the checklist and start selling the decisions, with software running the checklist underneath.

We see this firsthand: we run Amazon Ads operations together with a European consultancy, executing Sponsored Products campaigns for their brand clients through our command-center tooling. They bring the client relationship and the strategy; the machine brings relentless Tuesdays.

If you're paying for the ritual

A fair test for any retainer you pay: ask which parts of the monthly work are rules — could be written as "when X, do Y" — and which parts are decisions. You should be paying a person for decisions. The rules should cost you compute.

For sellers & agencies

We build this automation as agency work — and ship it in Commerce Agent.