Healthcare App Discovers $50K/Month Wasted Spend

Challenge

A successful digital health platform was spending €180,000 per month on Google Ads and Meta Ads.

Performance looked decent on paper.

The metrics:

• 3.8 ROAS (not bad)
• €85 CAC (acceptable for healthcare)
• Growing steadily
• Board was happy

But the CMO had a gut feeling.

"We're spending too much. Something doesn't add up."

The problem: Can't pinpoint what's wrong.

Dashboards showed:

• Campaigns trending up
• ROAS stable
• No obvious red flags

But they knew:

• Competitors acquiring cheaper
• Some campaigns felt inefficient
• Analysis was surface-level

The team was too busy to investigate:

• 4 marketers managing €180K/month
• All hands on creative, landing pages, tests
• Zero time for deep analysis
• Dashboard reviews = 30 minutes/week max

Hidden waste is invisible in dashboards.

That's the nature of waste. It hides in:

• Keywords that look OK but don't convert
• Audiences that overlap
• Campaigns that used to work
• Geos that underperform
• Times of day that drain budget

You need to dig through everything to find it.

At €180K/month across hundreds of campaigns?

Impossible manually.

Solution

They connected Cogny to see what AI would find.

Setup: 15 minutes

• Google Ads account (47 campaigns)
• Meta Ads account (23 ad sets)
• GA4 via BigQuery

48 Hours Later: First Analysis Complete

AI generated 67 tickets.

Top 10 tickets (sorted by monthly impact):

Ticket #1: Zero-Conversion Keywords

Issue: "128 keywords have spent €18,200 over last 90 days with zero attributed conversions"

Detail: Keywords that get clicks, sometimes even good CTR. But never convert.

Examples:

• "free telemedicine app" (€2,400 spent, 0 conversions)
• "healthcare app reviews" (€1,800 spent, 0 conversions)
• Dozens more...

AI recommendation: Pause these keywords → Save €6,100/month

Why team missed it:

• 780 total keywords
• These had decent CTR (looked healthy)
• Only conversion analysis exposed it
• Manual review would take days

Ticket #2: Audience Overlap

Issue: "4 Meta ad sets targeting 72% identical audience → Competing in same auctions"

Ad sets:

• "Healthcare Seekers"
• "Wellness Enthusiasts"
• "Telehealth Users"
• "Digital Health Adopters"

Impact: Driving up CPMs by bidding against themselves.

AI recommendation: Consolidate into 2 ad sets → Save €4,200/month

Why team missed it:

• Meta doesn't clearly show overlap
• Each ad set looked fine individually
• Only cross-analysis revealed it

Ticket #3: Geographic Waste

Issue: "12 regions converting at <1.0 ROAS → Burning €8,400/month"

Specific regions: Lower-income areas where healthcare app adoption was low.

Performance breakdown:

• Top 10 regions: 5.2 ROAS average
• Middle 20 regions: 3.8 ROAS
• Bottom 12 regions: 0.8 ROAS

Same bids everywhere.

AI recommendation: Pause bottom 12, reallocate to top 30 → Save €8,400/month + increase conversions

Why team missed it:

• 42 regions total
• Looked at aggregate, not regional breakdown
• Required granular analysis

Ticket #4: Search Terms Bleeding Budget

Issue: "247 search terms triggered ads but never converted → €9,200 spent last 90 days"

Examples:

• "free healthcare advice" (broad match picked it up)
• "healthcare jobs" (irrelevant)
• "healthcare news" (informational, not transactional)

AI recommendation: Add 247 negative keywords → Save €3,100/month

Why team missed it:

• Thousands of search terms
• Manual review is tedious
• Not obvious which are problematic

Ticket #5: Time-of-Day Inefficiency

Issue: "Ads running 2 AM - 6 AM converting at 1.2 ROAS vs 5.1 ROAS during daytime"

Budget spent overnight: €12,000/month for minimal return.

AI recommendation: Adjust ad schedule, shift budget to high-performing hours → Save €4,800/month + improve conversions

Why team missed it:

• Didn't analyze by hour
• Assumed 24/7 was fine
• Only aggregate daily metrics reviewed

Ticket #6: Creative Fatigue

Issue: "8 top ads showing 40% CTR decline over last 30 days → Performance degrading"

Still spending heavily: €18,000/month on fatigued creative.

AI recommendation: Rotate to backup creative → Recover CTR → Save ~€3,600/month

Why team missed it:

• Tracking 85 ad variations
• Fatigue is gradual
• Easy to miss in weekly reviews

Tickets #7-10: Various Issues

• Device performance variance (mobile worse in certain campaigns)
• Competitor brand terms not converting (€2,100/month waste)
• Duplicate campaigns running simultaneously (€1,800/month)
• Underperforming placements on Meta (€2,600/month)

Total from top 10 tickets: €44,700/month waste identified

Additional 57 tickets found more opportunities.

Grand total: ~€52,000/month wasted spend discovered

That's 29% of their €180K/month budget.

Implementation

Week 1: Quick Wins

Implemented top 5 tickets:

• Paused zero-conversion keywords
• Consolidated overlapping audiences
• Paused worst-performing regions
• Added negative keywords
• Adjusted ad schedule

Time to implement: 3 hours

Immediate impact:

• Spend dropped to €165K/month (€15K saved)
• ROAS increased to 4.4 (maintaining conversions)

Week 2-4: Deeper Optimization

Implemented tickets 6-20:

• Rotated fatigued creative
• Fixed device bidding
• Removed duplicate campaigns
• Optimized placements
• Refined targeting

Ongoing: AI continued generating tickets daily. Team reviewed and executed high-priority items.

Results

After 8 Weeks

Wasted Spend Eliminated:

• Before: €52,000/month waste (hidden)
• After: ~€8,000/month (always some inefficiency)
• Eliminated: €44,000/month waste

Ad Spend Optimization:

• Before: €180,000/month
• After: €136,000/month (same conversions)
• Savings: €44,000/month = €528K/year

Or: Maintain Spend, Improve Results They chose to maintain €180K spend but reallocate:

• Cut €44K from waste
• Added €44K to winners
• Net: Same budget, way better results

ROAS Improvement:

• Before: 3.8
• After: 5.6
• Improvement: 47% increase

CAC Reduction:

• Before: €85
• After: €58
• Improvement: 32% reduction

Customer Acquisition:

• Before: 2,118 customers/month
• After: 3,103 customers/month
• Increase: 985 more customers (47% growth)

What This Meant for the Business

Annual Impact:

Option A (Reduce Spend):

• Save €528K/year in ad spend
• Maintain customer acquisition
• Improve profitability

Option B (Maintain Spend, Grow):

• 47% more customers acquired
• Better unit economics
• Faster growth

They chose Option B.

Business Metrics After 6 Months:

Revenue Impact:

• 985 additional customers/month × €420 average LTV
• Value: €413,700/month
• Annual: ~€5M additional customer lifetime value

Efficiency Gain:

• Marketing efficiency ratio improved 47%
• Same budget driving way more growth
• Path to profitability accelerated

Time Savings:

• No longer spending 20+ hours/week looking for optimization opportunities
• AI finds them automatically
• Team executes in 4-5 hours/week

Key Insights from AI

1. Waste Hides in Plain Sight

Their dashboards looked fine.

3.8 ROAS isn't bad. €85 CAC was acceptable. Steady growth.

But underneath: 29% of budget was wasted.

Why they didn't see it:

• Aggregate numbers masked problems
• Would need to analyze 780 keywords individually
• Review 247 search terms manually
• Check 42 regions one by one
• Track 85 ad creatives over time

Impossible manually at scale.

2. "Good Enough" is Often "Not Good At All"

Many of the wasted elements had decent metrics:

• Keywords with 3-4% CTR (looks good!)
• But 0 conversions (terrible)

Surface-level analysis misses this.

You need conversion tracking across everything. AI does this automatically.

3. Opportunity Cost is Real

€44K/month waste = €44K they could spend on what works.

Before optimization:

• Best campaigns: Limited by budget
• Worst campaigns: Getting plenty of budget

After optimization:

• Best campaigns: Fully funded
• Worst campaigns: Paused

Same total spend, 47% better results.

4. Granular Analysis Reveals Patterns

Aggregate view: "Campaigns performing at 3.8 ROAS"

Granular view:

• Top 30%: 6.2 ROAS
• Middle 40%: 3.9 ROAS
• Bottom 30%: 0.9 ROAS

Action: Kill bottom 30%, fund top 30%.

Can't do this without granular analysis.

5. Continuous Optimization Compounds

Week 1: Found €44K/month waste Week 4: Found €8K more (campaigns changed) Week 8: Found €6K more (new issues emerged)

Marketing is dynamic.

What works today stops working tomorrow. New waste appears constantly.

Need continuous monitoring.

Manual review can't keep up. AI can.

What The Team Said

"We thought we were doing fine. 3.8 ROAS seemed decent. Then AI found €52K/month in waste. We were shocked. That's almost 30% of our budget down the drain."

— CMO

"The scariest part: We would never have found this manually. Too many campaigns, too much data, too little time. AI found it in 48 hours."

— Head of Performance Marketing

"We're now operating at 5.6 ROAS instead of 3.8. That's not incremental. That's transformational. And it cost us €15K/year for Cogny vs €500K+ we were wasting."

— CFO

"Best investment we made this year. Period. The ROI is insane."

— CMO

Lessons Learned

1. Audit Before You Scale

They were about to scale to €250K/month.

If they'd scaled without optimizing:

• Waste: €52K → €72K/month
• Lost: €240K/year additional waste

Instead:

• Optimized first
• Then scaled
• Scaled efficiently

Lesson: Fix efficiency before adding budget.

2. Dashboards Lie (By Omission)

Their dashboard showed:

• ROAS: 3.8 ✓
• CAC: €85 ✓
• Growth: Steady ✓

All looked fine.

But dashboard didn't show:

• Which keywords waste money
• Where audiences overlap
• What geos underperform
• When creative fatigues

Dashboards show aggregate. Problems hide in details.

3. Manual Analysis Doesn't Scale

To find what AI found would require:

• 80+ hours of analyst time
• Spreadsheet analysis
• Cross-referencing data
• Still might miss patterns

Cost: €8,000+ in labor

AI cost: €1,200/month

Plus: AI does it continuously, not once.

4. Every Company Has Hidden Waste

"Our campaigns are well-optimized" they said.

But they had 29% waste.

Reality: Every account has waste. Even "well-managed" ones.

Because:

• Marketing is complex
• Changes constantly
• Manual review can't catch everything

AI finds what humans miss.

5. The ROI is Immediate

Month 1:

• Cogny cost: €1,200
• Waste eliminated: €44,000
• ROI: 36x (first month!)

Most tools take months to show ROI. This paid for itself in 1 day.

Replicability

This result is replicable if you have:

✅ €50K+/month ad spend (enough data for AI to analyze) ✅ Multiple campaigns and keywords (more complexity = more hidden waste) ✅ No recent deep audit (waste accumulates over time) ✅ Feeling you're missing something (usually correct!)

Expected findings:

• 15-35% wasted spend in most accounts
• Higher in accounts that haven't been audited in 6+ months
• Lower in recently optimized accounts (but still 5-10%)

Typical timeline:

• Week 1: AI identifies waste
• Week 2-4: Implement quick wins
• Month 2-3: Full optimization, see results

Expected outcomes:

• 20-50% ROAS improvement
• 15-30% CAC reduction
• Or: Same budget, 20-40% more customers

Want to Know Your Hidden Waste?

Every marketing account has waste.

The question isn't "if" but "how much."

Most companies discover 15-35% of budget is wasted on:

• Keywords that don't convert
• Overlapping audiences
• Wrong geos or times
• Fatigued creative
• Inefficient placements

You can't see it in dashboards. You need AI to analyze everything.

Find your waste:

Schedule a demo

We'll run a preliminary analysis and show you:

• Estimated waste in your account
• Specific examples
• Expected impact of optimization
• 30-day ROI projection

Usually finds $10K-100K/month (depending on spend level).

Often pays for itself in 1 day.

---

About This Case Study

Written by the Cogny team—built by the founders who created AI optimization systems for Netflix, Zalando, and Momondo at Campanja, and scaled growth for Kry, Epidemic Sound, and Yubico through GrowthHackers.se.

Company details anonymized to protect client confidentiality. Waste percentages and results verified and representative of typical audit outcomes.

Last Updated: December 27, 2024