Auto Max Val Decision — Deep Dive

The single biggest decision in this audit. ~$1.6M/yr is on the table.

What "Auto Max Val" is

Why it exists (the media buyer's strategy)

The BAU / ALC / Max Val trio is intentional, not accidental:

• BAU (Business As Usual) — optimizes for OFFSITE_CONVERSIONS, gets the lead at lowest cost. The volume engine.
• ALC (Advantage+ Leads Campaign or similar) — Meta's automated optimization, broader reach.
• Max Val (Maximum Value) — optimizes for RETURN_ON_AD_SPEND against the Purchase event. The premise: pay more per lead, but get higher-VALUE leads (better accept rate, higher payout) that more than make up for the CPL premium.

The media buyer's bet: at Max Val, the algorithm should learn to find leads that QuoteWizard will accept and pay premium for. Even at $36.44 CPL, if those leads have a 50% higher accept-rate × payout vs BAU leads at $21.84, Max Val wins.

That's a coherent strategy IF the value signal works.

Why it's broken

Meta returns purchase_roas: "Not available" at every level of the account — campaign, ad set, ad. This is because:

1. The Purchase event is firing (good)
2. The Purchase event has high EMQ (9.1 — good)
3. But the value parameter on the Purchase event is either malformed or flat (Mo's audit: "98% of Purchase price data has formatting issues" — needs Events Manager confirmation)

The result: Meta's RETURN_ON_AD_SPEND optimization has no signal to optimize against. It's optimizing against noise. The campaign is structurally paying for premium optimization that isn't happening.

Visualizing the bleed

Auto BAU                Auto Max Val
$338,882 spend          $341,592 spend
15,587 leads            9,375 leads
$21.84 CPL              $36.44 CPL
                        ↑ $14.60 PREMIUM per lead

9,375 leads × $14.60 = $136,912 / 30 days
                     = $1.64M / year

                     PAID for an optimization
                     that isn't happening

The three options, with math

Option A — Pause Max Val, reallocate budget to BAU

Action: Pause Auto Max Val campaign. Move the $7,000/day budget to Auto BAU (currently underspent — $11.3K/day actual vs $19.8K/day cap).

Math: - Auto BAU pacing gap is $8.5K/day undeployed - Pausing Max Val saves $7K/day = $210K/30d - Reallocating $7K/day to BAU at $21.84 CPL = ~320 additional leads/day - Current Max Val produces 9,375 leads/30d = 312 leads/day at $36.44 CPL - Same lead volume, ~40% lower CPL.

Pros: - Immediate savings: ~$110K/mo at same volume (BAU CPL × current Max Val lead count) - Eliminates the "paying for broken optimization" bleed - Frees Auto BAU pacing — actually USES the budget that's been sitting

Cons: - Loses the campaign structure — relighting Max Val later means re-warming the learning phase - Some optionality loss if the value-fix lands quickly and Max Val WOULD have worked

Best if: dev says the value-fix timeline is "6+ weeks" or "unsure"

Option B — Cap Max Val at current $7K/day, fix value pipe in parallel

Action: Don't pause. Don't scale. Lock the daily budget. Start value-pipe dev work in parallel. Revisit in 2 weeks.

Math: - Continues $341K/30d at $36.44 CPL - Continues paying $137K/mo premium during the fix window - If fix lands in Week 2: Week 1 + half of Week 2 of premium spend ≈ $50-70K wasted - If fix lands in Week 3: $100K wasted - If fix lands in Week 4: $137K wasted

Pros: - Preserves campaign learning (no re-warm cost) - Maximum optionality — if value-fix lands fast, Max Val starts working - Lowest risk politically (no "Mo paused our biggest campaign" optics)

Cons: - Bleeds $4.5K/day of premium spend during the fix window - Doesn't actually solve anything — just delays the decision

Best if: dev says the value-fix timeline is "2 weeks" with confidence

Option C — Keep running, fix value pipe in parallel, decide later

Action: Don't change anything. Start dev work. Revisit at the value-fix landing.

Math: - Continues $341K/30d unchanged - Total wasted spend during typical fix window (3-4 weeks): $300-500K

Pros: - Zero operational change required - No risk of disrupting current performance

Cons: - Costs ~$300-500K in premium spend during the fix - "We knew it was bleeding and we did nothing" if Hassan asks why

Best if: never. This option exists only to be rejected unless there's a specific reason to preserve current state.

My recommendation

Cap to $7K/day for Week 1 (Option B), then re-decide Week 2 based on dev confidence:

• If dev says "value-fix lands in 2 weeks with confidence" → stay capped (Option B → C)
• If dev says "value-fix is 6+ weeks or uncertain" → pause + reallocate (Option B → A)
• If dev says "3-4 weeks" → pause + reallocate. Reallocation savings in 3 weeks (~$110K) exceed the optionality value of preserving the campaign structure.

The framing for Hassan:

"I want to cap Auto Max Val at its current spend for Week 1 so we stop the bleed but don't lose the campaign yet. By next Tuesday, dev tells us how long the value-pipe fix takes. If it's two weeks, we hold the cap. If it's a month or more, we pause Max Val and shift that budget to BAU — same lead volume, 40% lower cost. Either way, we're not paying for broken optimization in Q3."

That's an operator framing Hassan can sign off on without a 30-minute conversation.

What to say if Hassan or Anthony push back

"We've been running Max Val for months. It must be working."

Response: "It's running, but Meta returns purchase_roas: Not available everywhere. That's the algorithm telling us the value optimization isn't happening. The CPL premium versus BAU on the same audience is structural — same audience targeting, 67% higher CPL. That's the cost of running ROAS optimization against a broken signal. The campaign isn't delivering on its premise; we just haven't measured it."

"Pausing breaks our delivery."

Response: "Auto BAU has $8.5K/day of unspent budget right now. The reallocation absorbs Max Val's volume at lower CPL. We don't lose delivery — we lose CPL premium."

"Max Val gets us higher-value leads even if Meta can't see it."

Response: "Possible, but unverifiable. The downstream lead quality data (accept rate, payout per lead) isn't visible to us in the Meta-only view. If you have that data, let's look at it — if Max Val leads have a 50%+ higher accept rate that justifies the premium, we keep it. Without that data, we're paying a premium on faith."

"What's the downside of just leaving it?"

Response: "$137K a month of premium spend for at least the 3-4 weeks of the value-fix dev cycle = $400-500K. That's the lower bound. If value-fix takes longer, more. Compared to a campaign pause that costs ~$0 to reverse, the do-nothing option is the most expensive option on the table."

What this audit can't tell us (and what Mo should ask)

The Meta-only view doesn't see downstream lead quality. To make this call with full confidence Mo needs from QW/Delty:

• Accept rate on Max Val leads vs BAU leads (same time period)
• Payout per accepted lead on Max Val vs BAU
• Return rate on Max Val vs BAU

If Max Val accept-rate is materially higher (say 30%+ over BAU), the campaign may be working despite Meta not being able to measure ROAS. In that case Option B (cap) becomes more attractive — preserve the campaign for when the value-fix gives the algorithm what it needs.

If Max Val accept-rate is roughly equal to BAU, the campaign is straight-up overpaying for nothing. Option A (pause) is the clear move.

This is a Tuesday ask of Anthony / Lauren: get me the accept-rate breakdown by campaign for the last 30 days.

Decision tree

Tuesday Tuesday Tuesday:
                          │
              [Dev value-fix timeline?]
                          │
              ┌───────────┼───────────┐
              │           │           │
        2 weeks     3-4 weeks     6+ weeks
              │           │           │
              ▼           ▼           ▼
         Cap (B)    Pause (A)    Pause (A)
         Stay B     Reallocate    Reallocate
         until      to BAU        to BAU
         fix lands  immediately   immediately
              │
              ▼
        Fix lands?
              │
        ┌─────┼─────┐
        │     │     │
       Yes   No    Late
        │     │     │
        ▼     ▼     ▼
     Scale  Pause Pause
     Max    Max    Max
     Val    Val    Val

If I get one thing from Tuesday on this topic

An honest dev timeline. Best case + worst case in weeks.

With that number, the Max Val decision makes itself. Without it, we're guessing.