impression-share-diagnostics

Impression Share Diagnostics Skill

Purpose: Provides expert analysis of impression share metrics to diagnose root causes of underspending and auction performance issues in Google Ads Search campaigns.

Type: Domain knowledge skill (auto-invoked)

---

What This Skill Deliberately Does NOT Do

• Calculate the budget change. Diagnosis ends at "budget is (or isn't) the constraint" — the sizing math (5-10% caps, adjustment factors, daily translation) lives in budget-recommendation-calculator.
• Diagnose PMax with the decision tree. Performance Max has no impression share metrics — the scenarios apply to Search campaigns (plus the bid-cap check for Display/Video remarketing). PMax gets budget-utilization and asset diagnostics instead (see the campaign-type section).
• Treat Rank Lost IS as a target. The skill reads it as evidence for the constraint diagnosis, never as a number to optimize down — chasing it is the misconception this skill exists to correct.
• Pull the metrics. The required API fields are listed under Data Sources Required; the query templates live in gaql-query-patterns.

---

Core Impression Share Metrics

Search Impression Share (Search IS)

What it measures: Percentage of impressions you received out of total eligible impressions

How to interpret:

• >80%: Excellent - capturing most available impressions
• 50-80%: Good - moderate market coverage
• 30-50%: Limited - missing significant opportunity
• <30%: Very limited - major opportunity loss

Note: Higher is better, but 100% is rarely achievable or necessary

Budget Lost Impression Share (Budget Lost IS)

What it measures: Percentage of impressions lost due to budget constraints

How to interpret:

• <10%: Budget is sufficient for current demand
• 10-30%: Moderate budget constraint (minor opportunity loss)
• 30-50%: Significant budget constraint (primary limiting factor)
• >50%: Severe budget constraint (major spend potential untapped)

Primary indicator: If high, increasing budget will likely increase spend

Rank Lost Impression Share (Rank Lost IS)

What it measures: Percentage of impressions lost due to ad rank (quality + bid)

How to interpret:

• <10%: Excellent competitive position
• 10-30%: Good competitive position (normal)
• 30-60%: Moderate competitive disadvantage
• >60%: Significant competitive disadvantage

CRITICAL CONTEXT: Rank Lost IS is informational, not a primary optimization target

---

Smart Bidding Context (CRITICAL)

Most Portfolio Campaigns Use Smart Bidding

Common bidding strategies:

• Max Conversions (no target) - Automated, no manual bid control
• Max Conversion Value (no target) - Automated, ROAS focus
• Max Clicks (with optional max CPC limit) - Automated, traffic focus

YOU CANNOT MANUALLY ADJUST BIDS - Google's algorithm controls them.

Primary Control Lever: Monthly budget (translates to daily budget)

Budget Management Philosophy:

1. #1 Priority: Hit monthly spend targets (pacing)
2. #2 Priority: Maintain acceptable CPA/ROAS goals
3. Budget increases are CONSERVATIVE: 5-10% monthly budget increases

---

Root Cause Diagnosis Decision Tree

Scenario 1: Budget Too Low (Most Common)

Pattern:

• Budget Lost IS: >10% (any level of budget constraint)
• Rank Lost IS: >50% (high competitive pressure)
• Search IS: <70% (missing impressions)

What's happening:

• Google's algorithm wants to bid higher to compete
• Budget cap prevents it from spending more
• Algorithm is forced to bid conservatively to stay within budget
• Result: Lower bids → Worse ad rank → Lost impressions

Diagnosis: Budget is too low for competitive auction environment

Recommendation: Increase monthly budget by 5-10%

Why this works:

• With smart bidding, budget is the primary lever
• Higher budget → algorithm can bid more aggressively
• More aggressive bids → better ad rank → more impressions → more spend

---

Scenario 2: Quality Score Issues (Secondary)

Pattern:

• Rank Lost IS: >60% (very high)
• Budget Lost IS: <10% (budget is sufficient)
• CPA significantly above goals
• Search IS: Low despite budget availability

What's happening:

• Low quality score = higher cost per click
• Google charges more for lower quality ads
• Algorithm can't compete efficiently
• Result: Fewer impressions even with budget available

Diagnosis: Ad relevance, landing page, or keyword quality issues

Recommendation:

1. Improve ad relevance (match ad copy to keywords)
2. Improve landing page experience (speed, mobile, relevance)
3. Refine keyword quality (remove low-performing keywords)
4. Secondary to budget fixes - quality improvements take time

---

Scenario 3: Low Demand (Normal, Not a Problem)

Pattern:

• Search IS: >80% (capturing most impressions)
• Budget Lost IS: <10% (budget is sufficient)
• Rank Lost IS: <10% (good competitive position)
• Still underspending vs. budget

What's happening:

• You're capturing most available impressions
• Low search volume for these keywords
• This is NORMAL - not every keyword has high volume

Diagnosis: Account is operating efficiently, just low organic demand

Recommendation:

• Option 1: Accept that this is normal (low search volume markets)
• Option 2: Consider reducing budget to match actual demand
• Option 3: Expand targeting (new keywords, looser match types, new geos)

DO NOT increase budget in this scenario - you're already capturing available demand

---

Scenario 3a: Brand Campaign Demand Ceiling (Budget Reallocation)

Pattern:

• Brand campaign Search IS: >90% (near-maxed)
• Budget Lost IS: <10%
• Brand campaign spending 40-60% of allocated budget
• Other campaigns (GEO, BDRM) have lower IS (more headroom)

What's happening:

• Brand search demand is capped - not enough searches for "[Property] apartments"
• At 90%+ IS, you're capturing nearly all available brand demand
• Budget allocated to Brand cannot be spent
• Meanwhile, GEO/BDRM campaigns have impression share headroom

Diagnosis: Budget misallocation - too much budget on demand-capped Brand, not enough on campaigns with headroom

Recommendation:

• Reallocate budget from Brand → GEO or BDRM campaigns
• Brand continues capturing all available demand (just with smaller budget)
• GEO/BDRM can compete more aggressively with additional budget
• Total account spend increases toward daily target

Example:

Before reallocation:
- Brand: $20/day budget, $10/day actual (50% util), 94% IS
- GEO: $45/day budget, $27/day actual (60% util), 18% IS
- BDRM: $45/day budget, $20/day actual (44% util), 44% IS

Action: Move $5-10/day from Brand to GEO/BDRM

After reallocation:
- Brand: $12/day budget (still captures 94% IS)
- GEO: $50/day budget (can now compete for more of the 82% lost IS)
- BDRM: $48/day budget (can compete for more of the 56% lost IS)

Key Insight: You can't force demand that doesn't exist - put budget where there's opportunity

---

Scenario 4: Recent Budget Increase (Ramp-Up Period)

Pattern:

• Budget was increased in last 3-7 days
• Currently underspending vs. new budget
• Impression share metrics not yet stabilized

What's happening:

• Smart bidding algorithms need time to adjust
• Algorithm gradually increases bids after budget increase
• Normal ramp-up period (3-7 days typical)

Diagnosis: Normal ramp-up period after budget increase

Recommendation:

• No action needed
• Monitor for 3-5 more days
• If still underspending after 7 days, re-investigate

DO NOT increase budget again during ramp-up period

---

Impression Share Context for Different Campaign Types

Search Campaigns (GEO, Bedroom, Brand)

• Full IS metrics available: Search IS, Budget Lost IS, Rank Lost IS
• Use decision tree above to diagnose root cause
• Primary focus: Budget Lost IS (indicates spend potential)

Performance Max Campaigns

• NO impression share metrics available (different auction dynamics)
• Cannot use IS for diagnosis
• Alternative diagnostics:
  • Budget utilization % (MTD spend ÷ MTD budget)
  • Asset performance scores
  • Auction insights (when available)

Demand Gen Campaigns

• Limited IS metrics (display network, different from search)
• Focus on: Reach, frequency, view rate
• Not applicable: Search IS decision tree

Display/Video Campaigns (Remarketing)

Key Diagnostic: Check for "Bid setting limited" status in Google Ads UI

Common Issue: Max CPC bid cap too restrictive on Maximize Clicks strategy

Pattern:

• Campaign status shows "Bid setting limited"
• Very low click volume despite budget availability
• Spending well below daily budget

Root Cause: Max CPC cap (e.g., $2-3) is lower than auction clearing prices for Display network. Campaign can't compete for impressions.

Standard Fix for Display Remarketing:

• Increase max CPC bid limit to $4.00 (standard floor)
• This provides headroom to compete while maintaining reasonable ceiling

Context: Accounts may use standardized content suitability settings which restrict available inventory. This means bid caps need to be higher ($4+) to compete in the smaller placement pool. If content suitability settings change, bid cap requirements may need adjustment.

Verification:

1. Check campaign settings → Bidding → "Maximum CPC bid limit"
2. If under $4.00, increase to $4.00
3. Monitor for 2-3 days - spend should increase toward budget

Expected CPC Range: $2-3 after auctions normalize (the $4 cap acts as ceiling)

Note: Display campaigns don't have Search IS metrics - diagnose via campaign status and bid strategy settings

---

How to Interpret Rank Lost IS in Context

Common Misconception:

"Rank Lost IS is high (60-80%), so we need to improve quality score or increase bids"

Reality:

• Rank Lost IS is informational, not a target
• High Rank Lost IS (60-80%) is common in competitive auctions
• Don't optimize TO a specific Rank Lost IS target
• Use it to understand WHY underspending (competitive pressure signals budget need)

Correct Interpretation:

• High Rank Lost IS + High Budget Lost IS → Budget is the constraint
• High Rank Lost IS + Low Budget Lost IS → Quality/efficiency is the constraint
• Low Rank Lost IS + Low Budget Lost IS → No constraint (low demand)

Primary focus: Pacing variance and CPA/ROAS performance, not Rank Lost IS itself

---

Budget vs. Quality Trade-offs

When Budget Increase is Appropriate:

✅ Underspending vs. monthly budget (pacing variance >±8%) ✅ Budget Lost IS >10% (spend potential exists) ✅ CPA is acceptable or below goal ✅ ROAS is acceptable or above goal ✅ No recent budget increase in last 7 days

When Budget Increase is NOT Appropriate:

❌ CPA is significantly above goal (>20% over target) ❌ ROAS is below minimum threshold ❌ Recent budget increase within last 3-7 days (ramp-up period) ❌ Search IS >80% + Budget Lost IS <10% (low demand, not budget constraint) ❌ Quality score issues evident (very high CPA despite budget availability)

The Conservative Approach:

• Start with 5% budget increase (not 10-20%)
• Monitor for 5-7 days before additional increases
• Never increase budget >10% in single change (algorithm shock risk)
• Quality improvements are secondary to budget optimization

---

Integration with Underspending Investigation

How This Skill Fits:

This skill provides the Step 3: Impression Share Analysis framework for underspending investigations.

Typical Investigation Flow:

1. Step 1: Check recent optimizations (ramp-up period check)
2. Step 2: Analyze campaign spend patterns (which campaigns, budget structure)
3. Step 3: USE THIS SKILL → Diagnose root cause via impression share
4. Output: Budget recommendation or quality improvement recommendation

Expected Output from This Skill:

After analyzing IS metrics, you should be able to answer:

• What is the primary constraint? (Budget | Quality | Low Demand | Ramp-Up)
• What evidence supports this? (specific IS metric values)
• What action is recommended? (budget increase | quality improvement | monitor | reduce budget)
• Why will this work? (mechanism explanation)

---

Quick Reference: IS Patterns Cheat Sheet

Search IS	Budget Lost IS	Rank Lost IS	Diagnosis	Recommendation
<70%	>30%	>50%	Budget too low	Increase budget 5-10%
<70%	<10%	>60%	Quality issues	Improve quality (secondary)
>80%	<10%	<10%	Low demand	Normal (or reduce budget)
Any	Any	Any	Recent budget ↑	Monitor 3-5 more days
50-70%	10-30%	30-60%	Mixed constraint	Increase budget 5% + monitor

---

When to Use This Skill

Auto-Invoked When:

• Investigating underspending issues
• Analyzing impression share metrics
• Diagnosing auction competitiveness
• Determining budget vs. quality constraints
• User mentions "impression share", "lost IS", "rank lost", or "budget lost"
• Creating budget increase recommendations
• Explaining why campaigns are underspending

Data Sources Required:

• Google Ads API: metrics.search_impression_share
• Google Ads API: metrics.search_budget_lost_impression_share
• Google Ads API: metrics.search_rank_lost_impression_share
• Campaign bidding strategy: campaign.bidding_strategy_type
• Campaign CPA/ROAS: metrics.cost_per_conversion, metrics.conversions_value_per_cost

CRITICAL: API Returns Decimals, Not Percentages

The Google Ads API returns impression share as decimals (0.0 to 1.0), NOT percentages.

API Value	Actual Percentage
0.2464	24.64%
0.1017	10.17%
0.0999	9.99%
0.6337	63.37%

When displaying to users or applying thresholds, always multiply by 100:

# Correct
search_is_pct = row.metrics.search_impression_share * 100
print(f"Search IS: {search_is_pct:.1f}%")  # Output: "Search IS: 24.6%"

# WRONG - will show misleading values like "0.25%"
print(f"Search IS: {row.metrics.search_impression_share:.2f}%")

Threshold comparisons must use percentage values (after multiplying by 100):

search_is = api_value * 100  # Convert to percentage first
if search_is > 80:  # 80% threshold
    print("High impression share")

---

Using Impression Share as Demand Ceiling (Budget Viability Analysis)

The Pattern

When evaluating whether an ad group or campaign can absorb a budget increase (or sustain spend after pausing other segments), use impression share as a demand signal to calculate the spending ceiling.

Key Insight:

• Current spend tells you what you are spending
• Impression Share tells you what you could spend (the demand ceiling)

Formula

Potential Spend = Current Spend / (Search IS / 100)

Example

Scenario: Client asks to pause all ad groups except one single-geo ad group. Can it absorb the budget?

Metric	Single Ad Group
Current spend	$91/month
Search IS	15.8%
Potential ceiling	$91 / 0.158 = $577/month
Required budget	$1,800/month
Verdict	Will severely underpace (68% under)

When to Use This

• Client asks to pause ad groups and consolidate budget into fewer segments
• Evaluating if a single geo/segment can absorb increased spend
• Predicting underspend risk before making changes
• Sizing addressable demand for a keyword set or geography
• Answering "can X absorb Y budget?"

Ad Group Level IS

Important: Search Impression Share is available at the ad group level, not just campaign level.

Query pattern:

SELECT
    ad_group.name,
    metrics.impressions,
    metrics.cost_micros,
    metrics.search_impression_share
FROM ad_group
WHERE campaign.name LIKE "%GEO%"
AND segments.date DURING LAST_30_DAYS

Nuances & Limitations

• This is a rough estimate, not exact
• Assumes similar CPCs at higher volume (CPCs may increase with more aggressive bidding)
• Smart Bidding may adjust behavior when segments change
• Works best for Search campaigns (PMAX has no IS metrics)
• The ceiling assumes 100% IS capture, which is rarely achievable in practice

Probability Assessment

Use the gap between potential ceiling and required budget to estimate underpacing risk:

Gap	Probability of Underpacing
Ceiling < 50% of required	>90% (near certain)
Ceiling 50-75% of required	70-90% (very likely)
Ceiling 75-100% of required	40-70% (likely)
Ceiling > 100% of required	<40% (may work)

---

When to Load Other Skills

Skill	When
budget-recommendation-calculator	AFTER diagnosis, when the verdict is "budget constraint" — it turns the diagnosis into a conservative, capped recommendation
underspending-investigation	The producing investigation workflow that routes here — this skill is its impression-share analysis step
investigation-methodology	When IS analysis is one layer of a broader performance investigation rather than a standalone underspend question
gaql-query-patterns	When pulling the IS metrics this skill interprets

---

Built by Kurt Henninger. More free skills: github.com/fourteenwm/ppc-ai-skills