01
Set the scope
Choose publisher, marketplace, campaigns, ASINs, audiences, or a keyword set.
Algorithm Studio
One workflow for every optimization decision.
Define scope, set conditions, preview publisher-specific changes, and approve every action before it reaches an account.
Decision preview
Amazon - 4 campaigns - 126 keywords
Raise bids on profitable exact terms
Guardrails passed - Ready for approval
The operating model
Every decision follows the same visible path.
Algorithm Studio turns optimization logic into a reviewable workflow. Teams see what will change, why it qualifies, and where it will run before anything is published.
02
Define conditions
Combine margin, conversion, spend, rank, inventory, and time signals with thresholds.
03
Preview branches
See how one intent maps to Amazon, Google, Meta, or future publisher actions.
04
Approve and run
Apply limits and publish only the changes that passed review.
Publisher-aware execution
The logic stays consistent while the action fits each channel.
Publisher branches translate a common optimization goal into the controls each platform supports.
AMZ
Amazon
Bids, budgets, placements, search terms, match types, and DSP audiences.
GGL
Search bids, budgets, match types, placements, and conversion value.
META
Meta
Audience, creative, delivery, and budget branches ready for the next connection.
KFD
Kaufland
Marketplace-aware branches with the same review, limit, and ledger structure.
40+available signals
4review stages
1decision ledger
Guardrails and approvals
Automation stays inside the boundaries you set.
Every workflow checks limits before it proposes an action. Approval mode lets your team review suggestions before live changes.
01
Scope limits
Restrict publishers, accounts, campaigns, products, or audiences.
02
Spend limits
Set daily and monthly caps, bid floors, ceilings, and change-size limits.
03
Approval modes
Choose suggestion, approval, or autopilot mode per workflow and account.
04
Preview diffs
Review before and after values, expected outcome, and affected entities.
05
Rollback context
Keep the previous value, reason, and outcome attached to each action.
06
Team controls
Give operators shared visibility with a searchable decision history.
The decision ledger
A complete record of what changed and why.
Every proposed, approved, skipped, and executed action is written to one audit trail.
01
Signal
The inputs and time window that triggered the recommendation.
02
Reason
The rule, threshold, and expected outcome behind it.
03
Action
Publisher, entity, old value, new value, and approval state.
04
Outcome
Expected versus actual impact, recorded for the next cycle.