An independent multi-trade contractor asked SearchLight to review their marketing data. They wanted to see what a data audit would uncover. Our team reviewed their SearchLight dashboard, recorded a 20-minute video walking through the findings, and flagged three opportunities. One of them saved money immediately.
SearchLight's dashboard surfaced three issues in the first pass:
1. $190K in revenue potential sitting in unsold estimates. Demand generation was strong. The biggest lift available wasn't more leads; it was closing more of what they were already earning.
2. Google Ads book rate was 24% vs. an industry norm of ~42%. Lead grading could help identify whether the gap was lead quality, call handling, or both.
3. Bing was underperforming compared to Google Ads. Bing was consuming 29% of total paid spend at a 0.6x ROAS. Google Ads was returning 1.1x. The recommendation: cut Bing spend significantly and reallocate.
"I really appreciated your video! You pointing out our Bing spend was a HUGE mistake on our end. We had Bing auto-synced with our Google Ads, which is why the ad spend was so high. Terrible ROAS on that."
They had enabled Bing's auto-import feature, which mirrors Google Ads campaigns into Bing automatically. Nobody was monitoring it. Bing was spending $10,389/month running the same campaigns as Google Ads with a fraction of the return. Once they saw the data, they cut Bing spend immediately.
The following month, after cutting Bing spend by 61%:
Bing ROAS went from 0.6x to 6.9x. Even with less spend, the channel produced more revenue: $27,863 vs. $6,441. The total account improved across the board: overall ROAS climbed from 7.7x to 11.6x, book rate improved from 33% to 39%, and closed revenue grew from $280K to $386K on $3K less spend.
Beyond the Bing fix, the contractor also took action on the first recommendation: focus on closing what's already in the pipeline. The results showed up immediately. Paying customer rate improved from 19.0% to 23.5%, meaning more booked jobs were converting to paying customers. Average ticket climbed from $1,867 to $2,093, a 12% increase. More of the existing demand was being captured at a higher value per job.
Here's how Bing performed before and after the fix:
| Bing Performance | November | December | Change |
|---|---|---|---|
| Spend | $10,389 | $4,058 | -61% |
| Leads | 17 | 11 | -35% |
| Booked Customers | 8 | 7 | -12% |
| Book Rate | 47.1% | 53.9% | +6.8pp |
| Paying Customers | 3 | 5 | +67% |
| Closed Revenue | $6,441 | $27,863 | +333% |
| Average Ticket | $2,147 | $5,573 | +159% |
| ROAS | 0.6x | 6.9x | +1,050% |
To check whether the improvement was simply seasonal, we compared unpaid channels across the same two months. If December had universally stronger demand, you'd expect all channels to improve at similar rates.
| Unpaid Channel | Nov Leads | Dec Leads | Nov Revenue | Dec Revenue |
|---|---|---|---|---|
| Google Business Profile | 250 | 205 | $91,287 | $103,416 |
| Direct | 111 | 115 | $89,462 | $109,978 |
| Google Organic | 23 | 24 | $7,006 | $5,854 |
GBP leads actually fell 18%. Google Organic revenue fell 16%. Direct was roughly flat on lead volume. December did not have uniformly stronger demand. The unpaid channels showed modest mixed performance while the paid channels, especially Bing, showed disproportionate improvement. Bing going from 0.6x to 6.9x was the auto-sync fix, not the calendar.
This contractor was spending $10K/month on auto-synced Bing campaigns that nobody was monitoring, and leaving revenue on the table from unsold estimates. The data to catch both was there. It just wasn't visible until SearchLight organized it by channel, by source, by revenue. One 20-minute video, three insights, one month to see results: 61% less Bing spend, 51% higher total ROAS, paying customer rate up 24%, and $105K more in closed revenue.
The Bing savings alone paid for a full year of SearchLight in one month.
Data Source
Independent multi-trade home services contractor · November vs. December 2025 · SearchLight revenue attribution with CRM integration · Account details anonymized
