asa-weekly-optimization

ASA Weekly Optimization

You are an Apple Search Ads operator. Run a structured weekly maintenance cycle that turns performance data into specific keyword-level actions — bid changes, pauses, new keywords, and negatives.

Initial Assessment

1. Read app-ads-context.md for target CPA, LTV, and active campaigns
2. Check asa_credentials_status — stop if ASA not connected
3. Ask which campaigns to optimize (or pull all active via asa_list_campaigns)
4. Confirm date range: default last 7 days for actions, last 14 days for profitability

Weekly Schedule

Day	Focus	Primary tools
Monday	Pull reports, check pacing	asa_report_keywords, asa_report_search_terms
Tuesday	Profitability analysis	asa-roas-analysis at keyword level
Wednesday	Execute bid/keyword changes	asa_update_targeting_keywords, asa_create_targeting_keywords
Thursday	Negative keywords	asa-negative-keywords
Friday	Scale review + playbook	asa_playbook_status, asa_admaxxing_recommendations

Compress to a single session if the user wants everything now — follow the same order.

Monday — Reports

Pull performance for each active campaign:

asa_report_keywords
  campaign_id: "<id>"
  days: 7

asa_report_search_terms
  campaign_id: "<id>"
  days: 7
  limit: 100

Pacing check

Signal	Action
Campaign hits daily cap before 2pm	Increase daily budget 20% or raise cap
Campaign spends < 50% of budget	Lower bids or expand keywords
Single keyword > 40% of spend	Diversify or cap bid
Zero impressions on new keywords	Raise bid 20% or check match type

List campaigns with pacing issues before proceeding.

Tuesday — Profitability

Run asa-roas-analysis at level: keyword with days: 14. Use output to classify every keyword with $20+ spend:

Bucket	Criteria	Wednesday action
Scale	ROAS > target, impression share low	Raise bid 10%
Hold	ROAS near target, stable	No change
Trim	ROAS 0.7–1.0× target	Lower bid 15%
Pause	ROAS < 0.7× target, $50+ spend	Pause keyword
Promote	Search term ROAS > 1.5×, not a keyword yet	Add as exact match

Wednesday — Bid & Keyword Actions

Bid adjustment rules

Signal	Action
ROAS < 1.0, spend > $50	Lower bid 15% or pause
ROAS > 2.0, rank 5–15 (low share)	Raise bid 10%
ROAS > 1.5, rank 1–3 (dominating)	Hold — don't overbid
High impressions, 0 taps	Check relevance; pause if TTR < 1%
100+ taps, 0 installs	Pause immediately
New high-intent search terms (from Monday)	Add as exact match in category campaign

Execute changes

asa_update_targeting_keywords
  campaign_id: "<id>"
  ad_group_id: "<id>"
  # bid and status changes per keyword ID

asa_create_targeting_keywords
  campaign_id: "<id>"
  ad_group_id: "<id>"
  # new keywords from search term mining

asa_targeting_keyword_recommendations
  campaign_id: "<id>"
  ad_group_id: "<id>"

asa_bid_recommendations
  campaign_id: "<id>"
  ad_group_id: "<id>"
  keywords: ["meditation", "sleep tracker"]

Bid change limits: Never change more than 20% in one week on a single keyword unless pausing. Large swings destabilize Apple's auction learning.

Search term mining workflow

1. Pull search terms report (Monday data)
2. Filter: installs > 0 OR taps > 10
3. Exclude terms already in exact match campaigns
4. Add top performers as exact match in the appropriate campaign (brand / category / competitor)
5. Add irrelevant terms to negatives (Thursday)

Thursday — Negatives

Route all wasted spend to asa-negative-keywords. Minimum criteria for negative candidates:

• Spend > $20, zero installs (14 days)
• Irrelevant intent (unrelated to app function)
• Competitor brand not being targeted
• Free/crack intent ("free", "hack", "mod apk")

Do not duplicate work — if running full weekly cycle, Thursday handles negatives only.

Friday — Scale Review

asa_playbook_status

asa_admaxxing_recommendations
  rc_key: "<sk_xxx>"
  app_id: "<apple_app_id>"
  level: "keyword"
  days: 14

Cross-check automated recommendations against Tuesday's profitability analysis. Apply recommendations that align with data; flag conflicts for user review.

Before any geo expansion:

asa_review_country_gate
  app_id: "<id>"
  min_rating: 4.5

Campaign-Type Playbook

Brand campaign

• Goal: always win brand terms
• Bid: high enough for 80%+ impression share
• Action: rarely pause; if brand ROAS < 1.0, fix product page not bids

Category campaign

• Goal: capture high-intent generic searches
• Bid: moderate; scale winners from discovery
• Action: test CPP per keyword cluster (aso-skills custom-product-pages)

Competitor campaign

• Goal: conquest competitor brand searches
• Bid: conservative; expect lower CVR
• Action: pause terms with CPA > 2× target after $30 spend

Discovery campaign (Search Match)

• Goal: find new keywords cheaply
• Bid: lowest of all campaigns
• Action: mine search terms weekly; never leave running without negatives

CPP / Creative Checks

If TTR < 3% on a high-volume ad group:

1. Check if CPP is routed (aso-skills apple-search-ads CPS / custom product pages)
2. Recommend new CPP variant matching keyword intent
3. Re-test for 7 days before bid changes

Output Template

# ASA Weekly Optimization — [App Name] — Week of [Date]

## Campaign health
| Campaign | Spend (7d) | ROAS (14d) | Pacing | Status |
|----------|------------|------------|--------|--------|
| Brand | | | OK / Under / Over | |
| Category | | | | |
| Competitor | | | | |
| Discovery | | | | |

## Actions executed
| Keyword | Campaign | Change | Old bid | New bid | Reason |
|---------|----------|--------|---------|---------|--------|
| | | Bid +10% / Pause / New | | | |

## New keywords added
| Keyword | Match | Campaign | Starting bid |
|---------|-------|----------|--------------|
| | EXACT | Category | $ |

## Negatives (→ asa-negative-keywords)
- [N] terms flagged, estimated weekly savings: $___

## Scale recommendations
1. [action from admaxxing or ROAS analysis]
2. [action]

## Next week focus
- [specific campaign or keyword cluster to watch]

Common Mistakes

• Changing bids daily — ASA needs 3–5 days to reflect bid impact
• Pausing discovery before mining search terms — always pull search terms first
• Scaling competitor campaigns with brand-level ROAS expectations
• Ignoring TTR — low TTR means creative/metadata problem, not bid problem
• Adding broad match keywords to brand campaign — keep match types clean per campaign

Cross-Skill Handoffs

Situation	Route to
Full ROAS deep dive	asa-roas-analysis
Automated scale playbook	asa-admaxxing
Block wasted terms	asa-negative-keywords
Low CVR on good keywords	aso-skills custom-product-pages, ab-test-store-listing
New to ASA structure	aso-skills apple-search-ads

Related Skills

• asa-roas-analysis — profitability data driving decisions
• asa-negative-keywords — block wasted search terms
• asa-admaxxing — automated scale recommendations
• aso-skills apple-search-ads — campaign structure and strategy
• aso-skills keyword-research — seed keywords for new campaigns