Skip to findings
UPLIFY16 checks
Demonstration report

A Search audit that localizes loss before changing bids

We trace query, keyword, ad, landing, sale and tracking so higher bids do not compensate for irrelevance.

Evidence status
Demonstration report
Source
Synthetic, internally consistent data
Period
Synthetic 30 day window

Metrics and names are synthetic. They demonstrate depth and arithmetic, not a client result.

Control snapshot

  • 90,000 spend synthetic snapshot
  • 180 Ads conversions CPA 500
  • 126 confirmed in CRM qualified CPA 714.29
  • 2,400 search queries classified by intent

The 90 second view

  1. Stop leakage Separate noncommercial terms without importing generic negative lists.
  2. Clarify routes Separate brand and nonbrand, then route each intent to the closest page.
  3. Protect Do not rewrite a brand campaign with confirmed commercial outcomes for Quality Score alone.

What was checked and where confidence ends

Sources

  • Campaigns, keywords and search terms
  • RSA, assets and landing pages
  • Bids, budgets and change history
  • Conversions and synthetic CRM

Coverage

All active Search campaigns, 2,400 terms and all primary conversions.

Limitations

Budget growth remains conditional without real margin.

Measurement confidence

Platform conversions overstate the commercial result

90,000 spend / 180 conversions = CPA 500. CRM confirms 126 outcomes, so qualified CPA is 90,000 / 126 = 714.29.

16 checks

Full audit map

The three decisions above give an executive summary. The complete review register below shows the checks behind those decisions.

This page shows 16 of 16 checks in the demonstration register. A client report also links each check to its source, accountable role, and working artifact.

CheckPriorityDecisionWhat we established
01 Measurement and the CRM gateThe platform counts an action while the business confirms only part of it as a qualified result.
GS-01 P0 Reconcile Connect the Ads conversion to the CRM stage. 180 platform conversions and 126 confirmed outcomes produce different CPAs.
GS-02 P1 Import Return qualified without personal data. Bidding needs lead quality rather than form submission alone.
GS-03 P1 Protect Do not change bids before signal verification. A new target on an incomplete conversion automates the wrong objective.
GS-04 P2 Explain Document the attribution-window difference. Ads and CRM can correctly show different values.
02 Search Terms and negativesThe main lever is intent classification instead of bulk negative-keyword additions.
GS-05 P1 Classify Label 2,400 queries by intent. 18,000 of 90,000 spend went to queries without a commercial outcome.
GS-06 P1 Exclude with evidence Add negatives only for confirmed mismatches. Some informational queries may assist a sale.
GS-07 P2 Map Route each intent to the correct landing page. A relevant query loses alignment on a general page.
GS-08 P2 Repeat Recheck waste after 14 days. Read the effect with budget and bids unchanged.
03 Architecture, brand, and RSABrand is separated for clarity, and every ad promise must be supported on the landing page.
GS-09 P1 Separate Give brand and non-brand separate budgets. A blended CPA hides the cost of new demand.
GS-10 P1 Align Tie each RSA promise to landing-page evidence. Five of eight synthetic pages lack the main proof used in the ad.
GS-11 P2 Protect Keep three aligned routes as the control. They provide comparison after weaker pages change.
GS-12 P3 Ignore score as outcome Do not treat Ad strength as a business result. An interface score does not prove lead quality or profit.
04 Bidding, Impression Share, and migrationAutomation changes after the objective is clean and within a controlled budget.
GS-13 P1 Record Capture baseline CPA, qualified CPA, and volume. All three are needed to avoid buying cheaper but weaker leads.
GS-14 P1 Migrate in stages Change one strategy in one group. Moving every campaign at once removes the control.
GS-15 P2 Verify Read lost IS beside economics. A lost impression is not lost profit without a marginal outcome.
GS-16 P2 Roll back Restore the strategy when qualified CPA degrades. Rollback follows business quality rather than the platform conversion.

Priority decisions with evidence

Fix

Different intents compete in one group

Source: terms and CRM model. Period: 30 days. Status: synthetic. Confidence: high. Coverage: 2,400 terms and 90,000 spend.

Observation
18,000 of 90,000 spend, or 20%, went to informational and service terms without a commercial outcome.
Decision logic
The observed loss is localized to query and landing. Some terms may assist sales, so exclusions require journey review.
Action and verification
Search lead classifies actual terms. Acceptance: every negative has query evidence and qualified CPA is rechecked after 14 days.
Separate

Brand hides weak nonbrand economics

Source: campaign report. Period: 30 days. Status: synthetic. Confidence: high. Coverage: all Search campaigns.

Observation
Brand and nonbrand share budget and target, making the average cheaper.
Decision logic
Average CPA does not show the cost of new demand. Separation alone does not guarantee growth.
Action and verification
Google lead separates budgets and reporting without changing bids. Acceptance: independent goals, negatives and landing map with unchanged total budget.
Rewrite

RSA promises more than the landing proves

Source: RSA and rendered landings. Period: snapshot. Status: synthetic. Confidence: medium. Coverage: 12 ads and 8 pages.

Observation
Five of eight pages lack the key proof used in the ad. Ad strength is not business evidence.
Decision logic
Message mismatch may reduce conversion. Three aligned pages remain the control.
Action and verification
The copywriter and CRO specialist create intent-specific proof blocks. Acceptance: 8 / 8 routes align promise, proof and CTA while bidding stays unchanged.

What works and should not change

Preserve the brand campaign with confirmed commercial outcomes

Brand routes produce commercial outcomes without a current tracking defect.

Separate its budget and keep it as the control.

Target state and change rules

01

Target state

Queries are split by intent, brand no longer masks non-brand, and bidding uses confirmed quality.

02

How we test

CRM signal and negatives come first, followed by one staged bid change at a stable budget.

03

When we roll back

The strategy returns if qualified CPA breaches the agreed threshold or confirmed volume falls.

The 7, 30 and 90 day plan

  1. 7 days
    Classify terms and split brand reporting.
    Owner
    Search lead
    Acceptance criteria
    All 2,400 terms are classified, and brand and non-brand segments have separate baseline metrics.
  2. 30 days
    Fix intent to landing congruence.
    Owner
    Copywriter and CRO specialist
    Acceptance criteria
    8 / 8 routes pass QA.
  3. 90 days
    Test broad or value bidding after maturity gate.
    Owner
    Google Ads lead
    Acceptance criteria
    Tracking is stable, the negative list is documented and data volume meets the stated test threshold.

What the client receives

  • Query to landing map
  • Conversion and CRM reconciliation
  • Brand and nonbrand decision
  • RSA and asset review
  • Tests with stop conditions

Method and appendices

  • Classifier for 2,400 search queries
  • Brand / non-brand and negative-keyword map
  • RSA → promise → proof → CTA
  • Bid migration and rollback protocol

Need the cause of expensive Search, not just lower bids?

We identify the funnel stage, evidence, counter evidence and a safe test.

Discuss an audit for your business