How a Tech & SaaS E-commerce Account Reduced CPA by 83% in a Single Quarter

A professional-tier Technology & SaaS e-commerce account operating at a $7,000 monthly Google Ads budget reduced cost-per-acquisition by 83.04% (from $175.85 to $29.82) and grew conversion volume by 807.82% over a 90-day optimization window through structured search-terms hygiene, Performance Max asset-group refinement, and bidding strategy recalibration against a clear target ROAS of 7.0x.

Key Takeaways

• CPA dropped from $175.85 to $29.82 (-83.0%) in a single 90-day window against the prior quarter baseline.
• ROAS expanded from 1.87x to 10.99x (+487.7%), crossing the stated 7.0x target and operating at a meaningful margin above it.
• Conversion volume grew from 54 to 491 conversions (+807.8%) on a 54% spend increase, evidencing efficiency gains rather than pure budget-driven volume.
• The account executed more than 75 documented optimization actions across the quarter, concentrated on search-term exclusions, Performance Max keyword themes, and campaign-level negative lists.
• Clickthrough rate improved from 0.40% to 1.14%, a 2.85x lift, indicating that traffic quality improved in parallel with conversion efficiency.

The Account

A professional-tier e-commerce retailer in the Technology & SaaS vertical, selling a national product catalogue through a direct-to-consumer online storefront. The account was running a mixed Search and Performance Max structure on a monthly budget in the $6,000-$8,000 range, with a documented target ROAS of 7.0x and no explicit target CPA ceiling. Traffic volume was moderate (~4-6M impressions per quarter), and the account had been operating for more than a year before the optimization window opened.

Before the intervention, the account was underperforming its own target by a wide margin: ROAS was sitting at 1.87x against a 7.0x goal, and CPA had drifted to nearly triple the median CPA for the Technology & SaaS vertical across the wider platform benchmark. The underlying issue was not budget, creative quality, or product fit — it was structural inefficiency in how the account was matching traffic to intent.

The Challenge

In the 90 days leading up to the optimization window (June 25 to September 25, 2025), the account produced the following performance:

Metric	Baseline (90 days)
Spend	$9,516.97
Conversions	54.12
Conversion Value	$17,816.69
CPA	$175.85
ROAS	1.87x
Clicks	18,095
CTR	0.40%
Impressions	4,545,806

Three structural problems were masking each other:

1. Performance Max was consuming budget against low-intent traffic. Impression volume was high (4.5M) but CTR was at 0.40%, indicating that the algorithm was finding impressions cheaply but not matching them to qualified intent. With no keyword exclusions feeding back into the P Max learning loop, the campaign had no signal to stop spending on irrelevant themes.

2. Search-terms hygiene was effectively absent. The account had a negative-keyword list on paper, but it was static and had not been updated to reflect the long-tail terms that the platform was actually matching against. Every week, the same wasted-spend patterns were recurring.

3. Bidding strategy was not aligned to the stated ROAS target. The account was using a conversion-volume optimization posture without a Target ROAS constraint, so the algorithm had no incentive to discriminate between a $50 order and a $500 order.

The result was an account that was spending nearly $10,000 per quarter to generate $17,816 in conversion value — a 1.87x return on a 7.0x target.

The Approach

The optimization program was executed as a structured 90-day engagement, with changes deployed in overlapping waves rather than as a single migration. The sequencing was deliberate: structural changes first, then bidding strategy, then ongoing hygiene.

Step 1: Full search-terms audit and negative-keyword restructuring. The team pulled the complete search-terms report for the previous 90 days and identified clusters of irrelevant matches that were consuming budget. These were consolidated into a structured negative-keyword list and applied at both campaign and account levels. Recurring low-value query patterns were excluded at the theme level to stop the Performance Max algorithm from re-expanding into them.

Step 2: Performance Max keyword theme expansion. With the negative-keyword infrastructure in place, the team then expanded the positive-signal side of Performance Max. In one documented action, 97 keywords were added in a single campaign (“MB - P Max - Stock Medio e Push 2026”) to give the algorithm clearer intent signals on which product stock levels and query patterns to prioritize. This was paired with asset-group refinement to align creative themes with the new keyword targeting.

Step 3: Bidding strategy recalibration against the 7.0x ROAS target. The campaign migrated fully into a value-based bidding posture (Maximize Conversion Value with Target ROAS), with the target set at 7.0x. This gave the algorithm a clear efficiency constraint and allowed it to down-bid low-value orders while scaling into high-value ones.

Step 4: Weekly search-terms review cadence. The team established a weekly review cycle in which every campaign’s new search-terms report was reviewed, flagged terms were excluded, and the cumulative savings were tracked. Over the 90-day window, this produced more than 20 documented search-term review completions, each one feeding new exclusions back into the account.

Step 5: Automated change-monitoring and alerts. Structural changes (budget adjustments, campaign status changes, keyword additions) were monitored through an automated change-history alert system, so that any drift or unintended modification could be caught within 24 hours rather than discovered weeks later in a performance report.

The Results

Over the 90-day optimization window (September 26 to December 25, 2025), the account produced the following performance:

Metric	Before (90 days)	After (90 days)	Change
Spend	$9,516.97	$14,650.70	+54.0%
Conversions	54.12	491.31	+807.8%
Conversion Value	$17,816.69	$160,995.12	+803.7%
CPA	$175.85	$29.82	-83.0%
ROAS	1.87x	10.99x	+487.7%
Clicks	18,095	63,327	+250.0%
CTR	0.40%	1.14%	+185.0%
Impressions	4,545,806	5,543,067	+21.9%

The most important signal in the before/after comparison is not the headline CPA reduction — it is the ratio between spend growth and conversion growth. Spend rose by 54%, but conversions rose by 808%. That delta is the unmistakable signature of an account that was leaving volume on the table because of structural inefficiency, not because of budget constraints.

ROAS reaching 10.99x against a 7.0x target also confirms that the optimization was not about squeezing more volume at any cost — the account was operating at a 57% margin above its stated efficiency target, which is the correct operating posture for a sustained scale-up phase.

Lessons Learned

1. Negative-keyword lists must be treated as a living asset, not a one-time setup. The single highest-leverage change in this account was establishing a weekly search-terms review cadence. Static negative lists decay within weeks as query patterns shift.

2. Performance Max requires explicit signal on both sides of the ledger. Adding 97 keywords at once looks aggressive, but it was only effective because the negative-keyword infrastructure was in place first. P Max campaigns that are given positive signals without negative constraints will re-expand into low-quality traffic.

3. A bidding strategy without a constraint is not a strategy. Migrating from unconstrained conversion maximization to Target ROAS gave the algorithm the efficiency frame it needed. The target should match the actual business target, not a conservative fallback.

4. CTR is a leading indicator of traffic quality. The account’s CTR almost tripled (0.40% to 1.14%) as the structural changes took effect. CTR movement of this magnitude is almost always a sign that impression mix has shifted toward higher-intent traffic, and it typically precedes CPA improvement by 2-4 weeks.

5. Spend growth and efficiency growth are not mutually exclusive. The conventional framing treats cost reduction and volume growth as a trade-off. In this case, both improved simultaneously because the starting point was so far below efficient operation that removing waste automatically unlocked volume.

Methodology Note

Data sourced from a managed Google Ads account in the Technology & SaaS vertical operating at the professional budget tier. All identifying information has been removed. Performance metrics reflect the best 90-day window (September 26, 2025 to December 25, 2025) compared against the prior 90-day baseline (June 25, 2025 to September 25, 2025). The account executed 77 documented optimization actions during the measurement period. Metrics are reported in USD.