The Component Breakdown: Diagnosing a Below-Average Quality Score Export

Marketing Academy · Field Work●Paid Ads

All projects·From: Quality Score & Ad Rank

MiniAudit· 20 minutes

Zendesk

Objective: Given a real keyword-level Quality Score export (10 keywords with the three component ratings), identify which component is dragging each keyword's score down and prescribe the correct fix, not just react to the composite number.

You're the paid search analyst at Zendesk reviewing why the 'Help Desk Software' campaign's CPCs have crept up 40% quarter over quarter, even though bids haven't changed.

Read the three-component breakdown, not the composite 1-10 score, and match each 'Below Average' rating to the specific fix the lesson prescribes for that component.

Before you start

What you'll need

Free path (everything below is enough to finish)

Google Sheets

FreeImport and filter the Quality Score export, flag Below Average components

Free, no account access required beyond a CSV export

Google Ads

FreePull the keyword-level Quality Score columns to build the export

Free to view for any account owner or manager

The process

2 steps

Step 01 of 02

Reading the three Quality Score components instead of the composite number

The lesson breaks Quality Score into three weighted parts: Expected CTR (~39%), Landing Page Experience (~39%), and Ad Relevance (~22%). Google shows each as Below Average, Average, or Above Average in the Keywords tab.

This export has 10 keywords. Sorted by which component is rated 'Below Average,' which single component shows up most often across the account?

Google Sheets— Import qs-export.csv, freeze the header row, and add a flag column per component.

Procedure

Import qs-export.csv into Sheets and freeze row 1
Add three helper columns: CTR Flag, Relevance Flag, LP Flag, marking each 'Below Average' cell
Sum each flag column to see which component fails most often across the 10 keywords

Sample output

Keyword                      QS   Exp.CTR        Ad Relevance     LP Experience
help desk software            4   Below Average  Average          Below Average
ticketing system software     3   Below Average  Below Average    Below Average
customer support platform     6   Average         Average          Below Average
it service desk software      4   Below Average  Average          Below Average
zendesk alternative            7   Average         Above Average    Average

Flag totals (10 keywords): CTR Flag = 6, Relevance Flag = 2, LP Flag = 7

Healthy

The flag totals point to Landing Page Experience (7 of 10 keywords) as the dominant failure, so the fix effort goes there first.

Unhealthy

Treating all 10 keywords the same way because the composite scores look similarly low, when the actual failing component differs keyword to keyword.

What this means

The composite score tells you something is wrong; the component flags tell you what to fix and in what order.

So what do I do about it?

Symptom	Action	Effort
7 of 10 keywords flag Below Average on Landing Page Experience	Prioritize dedicated landing pages per keyword theme before touching ad copy	half day
Only 2 keywords flag Below Average on Ad Relevance	Leave ad group structure alone for now; it isn't the bottleneck	5 min

YouYou can do this yourself, no engineering access required.

Step 02 of 02

Matching a Below Average rating to its specific fix

The lesson gives a distinct fix list per component: CTR fixes are about headline-keyword matching and extensions, Ad Relevance fixes are about ad group tightness, and Landing Page fixes are about dedicated pages, speed, and message match.

For the two keywords flagged Below Average on Ad Relevance ('ticketing system software' and one other), what is the single most likely structural cause, based on the lesson's mistake list?

Google Sheets— Same qs-export.csv, cross-referenced against the account's ad group list.

Procedure

Filter the export to the 2 rows flagged Below Average on Ad Relevance
Check which ad group each keyword sits in against the account's ad group export
Confirm both keywords share one broad ad group with 40+ other terms

Sample output

Ad group: 'Support Software - Broad' (47 keywords)
  ticketing system software    QS 3   Ad Relevance: Below Average
  it help desk software         QS 4   Ad Relevance: Below Average

Ad group: 'Zendesk Brand'  (3 keywords)
  zendesk alternative           QS 7   Ad Relevance: Above Average

Healthy

Both Below-Average-Relevance keywords sit in the same 47-keyword catch-all ad group; the fix is splitting it into 3-5 keyword clusters with matched copy, exactly the lesson's mistake #1.

Unhealthy

Rewriting ad copy for the whole broad ad group once and expecting Ad Relevance to recover, without first splitting the cluster.

What this means

Ad Relevance problems are almost always a structure problem (ad group is too broad), not a copywriting problem alone.

So what do I do about it?

Symptom	Action	Effort
A 47-keyword ad group has 2+ keywords Below Average on Ad Relevance	Split the ad group into 3-5 keyword clusters with dedicated ad copy per cluster	half day

YouYou can do this yourself, no engineering access required.

Final deliverable

A component-flag table across all 10 keywords with the dominant failing component identified, plus a one-paragraph fix recommendation prioritized by impact.

See a reference example

Sample output

Freshworks CRM, QS component audit (excerpt)

Flag totals (12 keywords): CTR Flag = 3, Relevance Flag = 8, LP Flag = 2

Finding: Ad Relevance is the dominant failure, driven by one 30-keyword 'CRM Software - Broad' ad group. Recommendation: split into 5 clusters (free CRM, CRM for startups, CRM pricing, CRM integrations, CRM alternatives) before any copy rewrite.

Success criteria

You're done when you can:

Correctly tallies which component fails most often across the export
Correctly identifies ad group structure, not copy quality, as the Ad Relevance root cause