Inside the Brand Leakage Engine
Every metric in your Brand Leakage audit traces back to a specific query and classification rule. This page documents the brand detection heuristic, cannibalization calculation, incrementality weighting, and every threshold used to generate your verdicts.
Jump to metric
Data Sources
read-only OAuth · no campaign changes everAll data is read via Google Ads API and GA4. Derived intelligence tables are pre-computed in our pipeline before your audit runs.
Google Ads API — Campaign Performance
google_ads_campaign_performance
Campaign names (for brand detection), spend, conversions, revenue, impression share metrics per day
Google Ads API — Search Terms
google_ads_search_terms_historical
Search term text, cost, conversions for cannibalization overlap detection and incrementality calculation
Google Ads API — Keywords Performance
google_ads_keywords_performance
Keyword text, match type, Quality Score for brand keyword QS analysis
Brand Detection Heuristic
How we classify campaigns as brand vs non-brand
We classify campaigns by scanning campaign_name for known brand naming conventions. This covers the most common patterns used by agencies and in-house teams. The audit also reports a brand_detection_config showing how many campaigns matched — so you can verify the classification worked for your naming convention.
Why this matters: Every downstream metric — CPA comparison, ROAS split, cannibalization, incrementality — depends on this classification being accurate. If your campaigns use non-standard naming, the split may under-count brand spend, making the account look healthier than it is.
In plain English: We look at each campaign's name for words like 'brand', 'branded', 'bb', 'tm', or 'trademark'. If the name contains any of these, it's classified as a brand campaign. Everything else is non-brand. Simple, but it works for 90%+ of accounts.
Brand Dependency Status
How reliant is the account on branded search traffic?
Total brand campaign spend divided by total account spend, expressed as a percentage. The status verdict has three tiers. Above 50% means more than half your budget captures people already searching your brand name — users who would likely convert organically.
Why this matters: Brand dependency is invisible in blended metrics. An account with 4x ROAS and 60% brand spend might only have 1.2x non-brand ROAS — meaning growth through paid search is actually unprofitable. The dependency verdict forces this conversation.
In plain English: We add up all spending on brand campaigns, divide by total spending, and tell you the percentage. Above 50% = you're overly dependent on brand traffic. Above 30% = moderate risk. Below 30% = healthy balance.
Brand vs Non-Brand CPA Comparison
The CPA illusion exposed at campaign level
CPA calculated separately for brand and non-brand campaigns by joining campaign_performance with the brand classification CTE. Each campaign shows its segment label, spend, CPA, and ROAS. The CPA gap between segments quantifies how much brand traffic inflates account-level metrics.
Why this matters: Brand traffic has inherently lower CPA because the user already intends to buy. Blending brand and non-brand CPA produces a meaningless average. Reporting them separately reveals the true cost of acquiring NEW customers.
In plain English: We calculate cost-per-conversion for brand campaigns and non-brand campaigns separately. Brand CPA is almost always much lower because those users already know you. The gap between the two numbers is the size of the CPA illusion in your blended reporting.
Cannibalization Overlap
Brand terms appearing in non-brand campaigns — the double payment
A three-CTE query: (1) identify distinct search terms from brand campaigns, (2) sum spend on those same terms appearing in non-brand campaigns, (3) express the non-brand cost as a percentage of total account spend. Every term in the output has been confirmed to appear in BOTH campaign types — not guessed from keyword similarity.
Why this matters: Every term in this table means you're paying for the same user twice — once in the brand campaign and once in the non-brand campaign. The non-brand payment is pure waste because the brand campaign (or organic listing) would have captured it.
In plain English: We find search terms that appear in both your brand AND non-brand campaigns. For each overlapping term, we show how much the non-brand campaign spent on it. That spend is waste — you're paying twice for the same search.
Incrementality Cost (70% Weighting)
The true cost of paying for demand you already own
Total non-brand spend on overlapping brand terms, multiplied by 0.7. The 70% weight is a deliberate conservatism: not all overlap is pure waste — some brand terms in non-brand campaigns may serve a competitive defense purpose (preventing competitors from winning the auction). The weighting acknowledges this nuance while still quantifying the cost.
Why this matters: This is the hero stat of the audit. It tells a founder or CMO: 'You're paying $X per month for demand that already exists organically.' The 70% weight keeps the number credible rather than alarmist, which is critical for stakeholder trust.
In plain English: We take the total non-brand spend on overlapping brand terms and multiply by 70% (because some of that overlap might be defensively useful). The result is how much you're paying for demand you already own — money that could fund new customer acquisition instead.
Brand Impression Share Trend
Are competitors stealing your brand traffic?
Daily brand impression share, lost-to-rank, and lost-to-budget from campaign_performance filtered to brand campaigns with non-null IS data. A declining IS line with rising lost-to-rank proves competitors are actively bidding on your brand name and winning auctions.
Why this matters: Losing brand impression share to rank means competitors are outbidding you on YOUR brand name. This forces your brand CPCs up and diverts your most valuable traffic to competitors. Lost-to-budget means your own campaign ran out of money — the cheapest fix available.
In plain English: We track three lines over time: what share of brand searches you win, how much you lose because competitors bid more aggressively, and how much you lose because your budget runs out. Declining share + rising competitor pressure = someone is actively attacking your brand.
Brand vs Non-Brand ROAS Trend
Is the gap widening or narrowing over time?
Daily ROAS calculated separately for brand and non-brand segments. Converging lines mean non-brand efficiency is improving (positive trend). Diverging lines mean brand dependency is structurally worsening — non-brand ROAS is falling relative to brand.
Why this matters: Direction matters more than the absolute ROAS numbers. A 6x brand ROAS and 2x non-brand ROAS might be fine if the lines are converging. But 4x brand and 1.5x non-brand with diverging lines is a crisis — the growth engine is weakening.
In plain English: We plot brand ROAS and non-brand ROAS on the same chart over time. If the lines move together, non-brand is catching up. If they move apart, the dependency problem is getting worse — your growth engine is becoming less efficient while brand coasts.
Budget Reallocation Plan
Which campaigns to cut brand spend from and where to redirect
Each campaign's ROAS compared to account average. Brand campaigns with above-average ROAS are candidates for REDUCE (they'll maintain performance at lower spend). Non-brand campaigns with above-average ROAS are candidates for SCALE. Campaigns at or below average get HOLD.
Why this matters: Most managers know they should rebalance, but don't know which specific campaigns to touch. This table removes the guesswork — each campaign has a REDUCE/SCALE/HOLD verdict backed by its ROAS relative to the account benchmark.
In plain English: We compare every campaign's ROAS against the account average. Brand campaigns that outperform should be trimmed (they'll still perform with less budget). Non-brand campaigns that outperform should be scaled (they're proven winners). Everything else holds steady.
Proprietary notice: The methodology, scoring models, waste classification logic, and recovery projections are proprietary to ClickCatalyst Digital and provided for informational purposes only. Results are subject to standard market volatility. Recovery projections are estimates, not guarantees. Contact us within 14 days if any metric appears inaccurate.
Also worth reading
How to Read Your Brand Leakage Report →
Every section explained with exact action steps
See these metrics applied to your account
Run My Free Brand Leakage Audit