PPC Bid Optimizer (with apply)

Recomputes Sponsored Products keyword bids toward a target ACoS and applies them through DataDoe Actions - always dryRun first, real change only on your approval. The write follow-through to the Wasted-Spend Watchdog. MCP-native end to end.

When to use this

• After the Wasted-Spend Watchdog flags bleeders, to actually fix the bids.
• Weekly bid tuning toward a target ACoS.
• Trigger phrases: "optimize bids", "lower my bids", "fix ACoS", "adjust keyword bids", "bid optimization", "reduce ad spend on losing keywords".

The framework. Target-ACoS bidding

For each keyword: new bid = current bid x (target ACoS / actual ACoS), clamped.

• ACoS too high (bleeder) -> cut bid toward target (floor at a min bid).
• ACoS well below target + converting -> raise bid modestly to win volume (cap the step, e.g. +/-30% per run, and a hard max bid).
• Not enough clicks -> leave alone (insufficient data). Never move a bid more than one step per run - PPC needs time to settle.

Configuration

• MCP base: https://mcp.datadoe.com/mcp/v1
• Read: Keyword Targeting Performance (amazon_ads_targeting_by_campaign_by_date) - spend/sales/clicks/orders per keyword.
• Current bid + target id: AMAZON_ADS_TARGETS_FIND action (returns targetId + live bid). The performance export does NOT carry a reliable bid/targetId, so FIND is the source of truth for both.
• Write: AMAZON_ADS_TARGETS_UPDATE action (targets[] with targetId, bid, adProduct: SPONSORED_PRODUCTS, targetType: KEYWORD, marketplaceScope, marketplaces).
• Inputs from user: target ACoS (e.g. 30%), min/max bid, max step (default 30%).
• Enablement: AMAZON_ADS_TARGETS_FIND and the real UPDATE require the org to enable these Actions (Settings > Actions) with the connection set to read+write. dryRun works even when disabled - so the analysis + preview always run.

Step-by-step workflow (MCP-native)

1. sellers_and_vendors_list -> pick the seller.
2. Performance: exports_create on amazon_ads_targeting_by_campaign_by_date, last 30-60d, groupBy [ad_keyword, ad_match_type, ad_campaign_id, ad_group_id], sum ad_spend/ad_sales/ad_clicks/ad_orders, filter ad_campaign_type = SPONSORED_PRODUCTS. Compute ACoS per keyword.
3. Current bids: actions_start AMAZON_ADS_TARGETS_FIND (targetQuery, adProductFilter SPONSORED_PRODUCTS, targetTypeFilter KEYWORD, stateFilter ENABLED), poll actions_get -> map targetId + current bid by keyword.
4. Compute new bid per the framework (clamp to min/max + max step). Build the change list, biggest ACoS offenders first.
5. dryRun: actions_start AMAZON_ADS_TARGETS_UPDATE with dryRun: true. Show the validated result (status VALIDATED) and the full before/after bid table.
6. On explicit approval only: actions_start again with dryRun: false, poll actions_get, report per-target acceptance. Otherwise stop - nothing changed.

Output format

1Bid Optimizer (dry run) - {marketplace} - target ACoS {t}%
2
3keyword              match   ACoS   cur bid   -> new bid   why
4{kw}                 PHRASE  156%   {cur}0.55    {cur}0.35  cut to target
5{kw}                 EXACT   14%    {cur}0.40    {cur}0.48  raise (headroom)
6
7Dry run: {n} bids validated, 0 issues.
8Reply "apply" to push these bids, or edit targets/target-ACoS first.
Copy code

Worked example (illustrative)

A phrase keyword might show ~€182 spend / ~€117 sales / 176 clicks / 14 orders -> ACoS ~156%. At a 30% target, new bid = current x (30/156) ~= a large cut, but clamped to the max single-step + min bid, so likely one ~30% cut this run, revisit next week. Contrast a keyword at ~18% ACoS - leave it or nudge up. The AMAZON_ADS_TARGETS_UPDATE dryRun validates cleanly (status VALIDATED, 0 issues) even when Actions are disabled - so the preview is always safe to show.

Quality self-check

• Did I get current bid + targetId from FIND (not guess)?
• Did I clamp to min/max and cap the per-run step?
• Did I dryRun and show before/after before any real write?
• Explicit approval before dryRun: false?
• Right marketplace + currency for the account?

Common mistakes

• Slashing a bid to zero on one bad week (over-correction).
• Raising bids on high-ACoS "converting" keywords (they're still unprofitable).
• Using the export's bid/targetId (unreliable) instead of FIND.
• Running the real update without enabling the Action / read+write connection.

Notes

• Write skill. Real writes need the Action enabled + read+write connection; dryRun always works for demo/testing.
• Pairs with ppc-wasted-spend-watchdog (find) and ppc-negative-keyword-applier (negate the truly dead terms).
• A DataDoe skill, built on DataDoe targeting data + AMAZON_ADS_TARGETS_UPDATE.