Ambrosia Treatment Center · Baseline benchmark
Where we are, where we’re going
Prior 3 months vs June MTD · Prepared June 6, 2026
Current census: 70 of 119 (59%). June is the first month of the engagement. The plan: beat May. No big swings promised this month — incremental improvements measured against a real baseline.
Current census (as of Jun 6)
Cost per qualified opportunity & cost per admit
| Month | GAds Spend | GAds Qualified | Cost / Qualified | GAds Admits | Cost / Admit |
|---|---|---|---|---|---|
| March 2026 | $100,063 | 37 | $2,704 | 13 | $7,697 |
| April 2026 | $103,837 | 46 | $2,257 | 22 | $4,720 |
| May 2026 | $106,978 | 36 | $2,972 | 15 | $7,132 |
| June MTD (6 days) | $21,361 | 7 | $3,052 | 2 | $10,681 |
Headline cost includes all spend and all GCLID-attributed conversions — both brand and non-brand. The brand-stripped picture below is the more honest read.
Brand-stripped picture — what we actually drove
The April "spike" was brand. Once you separate branded admits from non-branded, non-branded admits were flat at ~11/month across March, April, and May. The variance was entirely in branded conversions, which would have converted regardless of paid spend. This is the metric our work moves; the headline isn't.
Admits classified by keyword (Salesforce CTM_Keyword field)
| Month | Total Admits | Unattributed (no GCLID) |
GAds — Branded | GAds — Non-Brand | GAds — No KW captured |
|---|---|---|---|---|---|
| March 2026 | 90 | 77 | 2 | 11 | 0 |
| April 2026 | 85 | 63 | 8 | 11 | 3 |
| May 2026 | 86 | 71 | 2 | 11 | 2 |
| June MTD | 13 | 11 | 1 | 1 | 0 |
Branded = CTM_Keyword contains "ambrosia", "neuroscience institute", or "academy at ambrosia". The 11/month non-brand floor across three months is a much steadier signal than the headline 13/22/15 swing.
Qualified leads classified by keyword
| Month | GAds Qualified | Branded | Non-Brand | No KW captured |
|---|---|---|---|---|
| March 2026 | 37 | 7 | 30 | 0 |
| April 2026 | 46 | 9 | 29 | 8 |
| May 2026 | 36 | 14 | 19 | 3 |
| June MTD | 7 | 0 | 7 | 0 |
Non-brand qualified leads dropped from 30 (March) → 19 (May) even as brand grew. This is the real volume problem — and where our work shows up first.
Estimated cost per non-brand admit/qualified (the real numbers)
| Month | Non-Brand Admits | Cost / Non-Brand Admit | Non-Brand Qualified | Cost / Non-Brand Qualified |
|---|---|---|---|---|
| March 2026 | 11 | ~$9,100 | 30 | ~$3,300 |
| April 2026 | 11 | ~$9,400 | 37* | ~$2,800 |
| May 2026 | 11 | ~$9,700 | 22* | ~$4,900 |
*Includes "No KW captured" treated as non-brand. Spend is treated as ~99% non-brand because visible branded spend is <1% of total in our data (audit estimated ~$5-6k/mo brand, still a small fraction of $100k spend). Non-brand cost per admit is flat at ~$9k; non-brand cost per qualified varies more.
Admits per month — total and Google Ads-attributed
| Month | Total Admits | GAds Admits (GCLID present) |
GAds Share | GAds Est. Revenue |
|---|---|---|---|---|
| March 2026 | 90 | 13 | 14% | $139,327 |
| April 2026 | 85 | 22 | 26% | $272,765 |
| May 2026 | 86 | 15 | 17% | $188,406 |
| June MTD (6 days) | 13 | 2 | 15% | $14,200 |
Total admits has stayed flat (85-90/mo); the variance is in Google Ads attribution. April is the high-water mark.
Qualified leads per month — what we optimize toward
| Month | Total Qualified (Stage past Missing Info) |
GAds Qualified | GAds Share |
|---|---|---|---|
| March 2026 | 259 | 37 | 14% |
| April 2026 | 237 | 46 | 19% |
| May 2026 | 198 | 36 | 18% |
| June MTD (6 days) | 42 | 7 | 17% |
Qualified leads is the metric we optimize the account against — admit variance is too high to be a stable bidding signal. April is the peak; May regressed, June MTD on pace to match May.
Funnel breakdown — Google Ads opps by current stage
| Created in | Missing Info | Active follow-ups (VOB, Priority, FUs) |
In transit (Travel/Detox) |
Admitted | Alumni | Refer Outs |
|---|---|---|---|---|---|---|
| March | 21 | 14 | 7 | 1 | 12 | 3 |
| April | 25 | 22 | 4 | 3 | 14 | 3 |
| May | 19 | 18 | 5 | 9 | 3 | 1 |
| June MTD | 1 | 3 | 1 | 1 | 0 | 2 |
Two patterns: Missing Info bucket is the biggest leak (~30% of created opps stuck there); refer outs are real qualified leads we lose to other facilities.
June targets — beat May, don’t over-promise
The June bar — non-brand focus
| Metric | May actual | June target | June MTD (day 6) |
|---|---|---|---|
| Non-Brand Qualified Leads | 19 | ≥ 30 (back to March) | 7 |
| Cost / Non-Brand Qualified | ~$4,900 | ≤ $3,500 | ~$3,000 |
| Non-Brand Admits | 11 | ≥ 13 | 1 |
| Cost / Non-Brand Admit | ~$9,700 | ≤ $9,000 | — |
| GCLID attribution % of total admits | 17% | ≥ 20% | 15% |
Targets are framed around non-brand because that's what our work moves. Branded admits hold steady regardless of paid spend and shouldn't be the bar we set for ourselves.
The honest framing: the prior agency delivered an April spike that wasn’t repeated. May regressed. The first June ask is to beat May, not chase the April peak. Pace right now is tracking flat-to-slightly-below May — that’s expected; our changes go live this month, not immediately.
What changes in June that wasn’t in March-May
- Conversion tracking rebuild. Salesforce qualified-lead and admit signals now feeding Google Ads bidding (offline conversion imports went live this week). Replaces the 60-second-call signal the algorithm was optimizing against before. Pure bidding-quality improvement.
- GCLID retroactive bootstrap. Last 90 days of qualified opps and admits uploaded as historical conversions to seed the algorithm with real signal from day one.
- Search term cleanup. Bulk negative load from the audit: competitor brands, medication research, sober living, junk. Cuts wasted spend that was diluting the optimization signal.
- Brand separation. "Ambrosia" branded terms moved to a separate manual-CPC campaign. Stops paying inflated auto-bid CPCs on captured demand.
- New landing pages. Replace the broken site as destination pages for paid traffic.
None of these promise an instant jump. They tilt the math toward better unit economics over the next 4-6 weeks. The next monthly review on July 1 is when we’ll see the first real read.