How we set up Scale Insights using Claude.
Scale Insights is a powerful Amazon PPC automation platform, but power without a plan just spends money faster. Our edge is the setup: a deliberate rule architecture we design with Claude, then let Scale Insights execute around the clock. Here's the approach we use on real accounts.
1 · Connect the account and define the targets
Before any automation, we agree the numbers that matter: target ACoS, the TACoS band we want to hold, and the real margin per SKU. We feed those into Claude alongside the business reports, and it helps us translate "profit goals" into concrete rule thresholds, the inputs Scale Insights actually needs.
2 · Build a bidding stack per SKU
We don't run one blanket rule across the account. Each SKU gets its own bidding stack, tuned to its margin and stage. High-barrier stacks go on SKAG campaigns that drive the most orders; broader stacks handle research. Claude helps us draft and sanity-check each stack, and because rules propagate, one change can lift the whole account at once.
3 · Launch autos the right way
Auto campaigns are underused by most brands. We launch bulk autos per product from day one, layered at different CPCs so variations surface across the funnel, plus catch-all low-bid autos to harvest cheap conversions. Claude helps structure the launch matrix so nothing overlaps or cannibalises.
4 · Automate negations
Wasted spend dies here. We run a custom negation stack with a simple rule of thumb: no search term should spend more than the sale price before producing a sale. Claude reviews the search-term reports, proposes the negation criteria per campaign type, and we load them into Scale Insights to run automatically.
5 · Daypart and protect margin
For the right products, time matters. An office product, for example, only needs ads in working hours, so we daypart to 8am–7pm and pull spend on weekends. Scale Insights handles the schedule; Claude helps us read the dayparting data to find the windows worth keeping.
6 · Let Claude read the data
This is where AI changes the game. Instead of an analyst eyeballing one report a week, Claude processes the full picture, every term, target and ASIN, and surfaces what changed and why. A senior operator owns every decision; Claude just makes sure nothing slips through.
The result
This setup is exactly what drove the outcomes in our case studies: 120% YoY growth on a multi-marketplace catalogue, a seasonal launch holding a 4.24x ROAS, and a seasonal product scaled past $1.1M. The tools matter, but the configuration is the craft.