60-Day ROAS Analysis
Feb 26 — Apr 27, 2026 · Every day we hit 3.0x+ TRUE ROAS
Days at 3.0x+
24
of 61 total days
Avg Spend on 3x Days
$5,658
vs $8,884 on sub-3x
Best Day
8.99x
Apr 23 on $5,409
01 / 06
The Pattern: Less Spend = Better ROAS
The data is clear. Days where the algorithm spends less are consistently the days it converts better. More budget does not mean more efficiency.
3.0x+ Days (24 days)
$5,658
Average daily spend
Sub-3x Days (37 days)
$8,884
Average daily spend
57% more spend on losing days — the extra budget goes to low-quality clicks that don't convert. The algorithm performs best when it's forced to be selective.
02 / 06
All 24 Days at 3.0x+
Every single day that cleared the target, with spend and revenue.
| # |
Date |
Spend |
Revenue |
ROAS |
| 1 | Mar 1 (Sun) | $8,544 | $27,393 | 3.21x |
| 2 | Mar 4 (Wed) | $5,537 | $25,840 | 4.67x |
| 3 | Mar 5 (Thu) | $5,203 | $21,552 | 4.14x |
| 4 | Mar 15 (Sun) | $8,249 | $25,366 | 3.07x |
| 5 | Mar 18 (Wed) | $4,314 | $14,238 | 3.30x |
| 6 | Mar 19 (Thu) | $3,613 | $19,190 | 5.31x |
| 7 | Mar 20 (Fri) | $4,144 | $14,533 | 3.51x |
| 8 | Mar 21 (Sat) | $4,924 | $15,895 | 3.23x |
| 9 | Mar 22 (Sun) | $4,895 | $17,260 | 3.53x |
| 10 | Mar 24 (Tue) | $5,222 | $20,390 | 3.91x |
| 11 | Mar 25 (Wed) | $3,798 | $15,788 | 4.16x |
| 12 | Mar 26 (Thu) | $3,179 | $25,256 | 7.95x |
| 13 | Mar 27 (Fri) | $3,618 | $19,469 | 5.38x |
| 14 | Mar 28 (Sat) | $8,021 | $27,405 | 3.42x |
| 15 | Mar 30 (Mon) | $5,440 | $22,219 | 4.08x |
| 16 | Mar 31 (Tue) | $6,779 | $29,135 | 4.30x |
| 17 | Apr 3 (Fri) | $5,018 | $21,568 | 4.30x |
| 18 | Apr 12 (Sun) | $6,505 | $20,616 | 3.17x |
| 19 | Apr 17 (Fri) | $4,163 | $16,702 | 4.01x |
| 20 | Apr 21 (Tue) | $6,723 | $25,284 | 3.76x |
| 21 | Apr 22 (Wed) | $5,635 | $26,704 | 4.74x |
| 22 | Apr 23 (Thu) | $5,409 | $48,636 | 8.99x |
| 23 | Apr 24 (Fri) | $6,742 | $39,070 | 5.80x |
| 24 | Apr 25 (Sat) | $9,840 | $48,435 | 4.92x |
03 / 06
Top 5 Best Days
Every single one had spend under $10K. The algorithm's best work happens when it's selective.
#1
Apr 23 (Thursday)
$5,409 spend → $48,636 revenue
8.99x
#2
Mar 26 (Thursday)
$3,179 spend → $25,256 revenue
7.95x
#3
Apr 24 (Friday)
$6,742 spend → $39,070 revenue
5.80x
#4
Mar 27 (Friday)
$3,618 spend → $19,469 revenue
5.38x
#5
Mar 19 (Thursday)
$3,613 spend → $19,190 revenue
5.31x
04 / 06
Before vs After tROAS Recalibration
On Apr 18 we recalibrated Branded Search from 16.79x to 3.5x tROAS and fixed NB campaign targets. The algorithm responded.
Before (Feb 26 — Apr 17)
Avg ROAS on 3x+ Days
4.15x
Avg Spend on 3x+ Days
$5,262
→
After (Apr 18 — Apr 27)
Avg ROAS on 3x+ Days
5.64x
Avg Spend on 3x+ Days
$6,870
Hit rate improved from 37% → 50% and avg ROAS on winning days jumped from 4.15x → 5.64x. The algorithm is spending more aggressively on winning days and pulling back on losers. Exactly what the tROAS fix was designed to do.
05 / 06
The Golden Streak: Mar 18–28
9 out of 10 days hit 3.0x+ during this run. Average spend was $4,815/day. The algorithm was in its zone.
Green = 3.0x+ · Gray = below target · Mar 23 was the only miss
06 / 06
Key Takeaways for the Team
1
$5,000–$7,000/day is the sweet spot. 69% of 3x+ days had spend in this range. Going above $9K consistently drops ROAS below target.
2
tROAS recalibration is working. Post-fix hit rate jumped from 37% to 50%, and winning days average 5.64x instead of 4.15x. The algorithm now bids on real purchase signals, not inflated engagement.
3
Branded Search is the anchor. On every 3x+ day, Branded Search delivered 60-75% of total revenue. When Branded Search performs, the whole account performs.
4
Subscription attribution takes 48-72 hours. Days that look like 1.5x at midnight often settle at 2.5-3.5x by Wednesday. Do not make knee-jerk decisions based on same-day numbers.
5
Weekend spend surges need monitoring. The algorithm ramps spend 2-3x on weekends (Apr 25: $9.8K, Apr 26: $15K). ROAS compresses but absolute revenue grows. Need to decide: cap weekend budgets or let the algorithm run?