The $73K Drain: How Automatic Placements and Dayparting Gaps Bled a Sportsbook Dry

The Situation

A multi-market sportsbook was running $10,500/week across Meta Ads↗ to promote live betting products in the Philippines and Thailand. The campaigns used Campaign Budget Optimization (CBO) across 8 ad sets with automatic placements — Meta's recommended default.

On the surface, the numbers looked acceptable: ROAS hovered around 1.1, which the team considered "break-even with LTV upside." But a deeper analysis revealed that nearly half the budget was generating effectively zero return.

What Went Wrong

Three budget leaks were operating simultaneously:

Leak 1: Placement Bleed (41% of spend wasted)

Automatic placements directed budget wherever Meta could deliver cheapest impressions. The breakdown:

Placement	% of Spend	Conversion Rate	ROAS
Facebook Feed	31%	2.8%	3.4
Instagram Feed	18%	2.1%	2.9
Instagram Stories	10%	1.4%	1.8
Audience Network	28%	0.03%	0.05
Messenger	13%	0.07%	0.09

Audience Network and Messenger combined consumed 41% of total spend ($29,930) while generating $1,800 in revenue. The algorithm chose these placements because CPMs were low — but the traffic was almost entirely accidental clicks and bot traffic.

Leak 2: Dayparting Blindspot (estimated $15K wasted)

The sportsbook's product was live betting — users bet during live matches. Peak conversion hours were 6 PM - 11 PM local time, coinciding with Premier League, NBA, and local league schedules. But the campaign ran 24/7 with even budget distribution. Analysis showed:

• 6 PM - 11 PM: 68% of all deposits, but only 22% of daily spend
• 2 AM - 10 AM: 3% of deposits, but 31% of daily spend

Leak 3: CBO Misallocation

Campaign Budget Optimization was sending 35% of budget to an ad set targeting "general sports" interest — a broad, low-intent audience with a $134 CPA. Meanwhile, the highest-performing ad set (targeting "football betting apps" interest) was budget-capped at 12% of total spend.

Diagnosis

RedClaw's budget efficiency audit mapped every dollar to a conversion outcome. The findings were stark:

• Effective budget: Only $31,100 of the $73,000 (42.6%) reached placements, times, and audiences that could realistically convert
• Blended ROAS of 1.1 masked a bifurcation: efficient spend generated 3.4 ROAS while leaked spend generated 0.06 ROAS
• True CPA on converting placements was $34, not the blended $78 the team was reporting

The Fix

We implemented a three-part budget containment strategy:

1. Placement restriction: Removed Audience Network and Messenger entirely. Restricted to Facebook Feed, Instagram Feed, and Instagram Stories only. Added Instagram Reels as a test placement with a 10% budget cap.

2. Dayparting schedule: Built custom ad schedules aligned with live sports:

   • Weekdays: 4 PM - 12 AM (pre-match through post-match)
   • Weekends: 12 PM - 12 AM (full matchday coverage)
   • Off-hours: paused entirely

3. Manual budget allocation: Replaced CBO with ad-set-level budgets, weighted by historical ROAS. Top performers received 60% of budget, test audiences received 15%.

Results

Within 12 days of implementing the fixes:

• ROAS jumped from 1.1 to 3.7
• CPA dropped from $78 to $29
• Total weekly spend actually decreased 34% (from $10,500 to $6,900) — but FTD volume increased 41%
• Wasted spend dropped from 57% to under 8%

The most striking metric: the client acquired more first-time depositors spending $6,900/week than they had spending $10,500/week. Budget leaks do not just waste money — they actively prevent the algorithm from learning what works.