The 8-month arc · May 2025 engagement → CTV Aug 11 2025 → Apr 2026.
The feedback loop was disconnected. The way they measured success was wrong. Acquisition strategies were hitting diminishing returns. Spend climbed +64% MoM yet new-customer CAC blew +55% and one-time CAC blew +142% - Meta CAC held flat at $138 and that was the headline they were tracking. The composed view told the truth: the system needed a strategic pivot.
Operating model reframe delivered. “It was never a Meta problem. It was a mix problem.” Audit → align → bridge. Cash-view (CFO) and accrual-view (CMO) reconciled in one operating model. Vendor measurement became a guide, not a gospel. Triangulation replaced single-source decisioning.
Phase 1 hero · Aug 11 – Oct 5 2025. Halo-forward CAC compressed from a $42.69 pre-CTV baseline to $32.58 over 8 weeks (−24%). Phase 2 (Oct – Dec, 9 weeks of holiday brand-layer cuts) drifted to $41.70. Phase 3 (Dec – Mar, 16 weeks, brand-layer restored at 40% intensity) settled at $40.20. Every week of the 33-week arc held below the pre-CTV baseline. MER trajectory aligns with brand-converted demand landing in highest-retention cohorts.
Brand pulse. Directional brand-lift read built from free public signal - baked in, no panel, no survey.
Weekly read. Seven independent public signals composed into one Brand Lift Index 0-100. This is not Disqo. No panel, no recall survey, no aided/unaided point estimate. What it is: a directional read that costs zero, runs always-on, and gets stronger when multiple signals converge. The value is the bundle - every UA partner brand gets this out of the box.
Every number on this page is synthetic - plausible values illustrating what the architecture renders once a brand's signal pipes are wired. The seven signal sources below are real APIs Stable integrates as part of running the brand's media. The architecture is the deliverable; the numbers populate when a brand's data flows.
When five-plus public signals rise together, the brand is moving - not just the click. That convergence is the read.
Seven independent public signals, each z-scored against a 28-day pre-flight baseline, capped at ±3, rescaled 0-100, composed into one composite. No panel rented. No survey commissioned. No vendor licensed. Refreshes weekly. Bands: 0-39 no read · 40-54 building · 55-69 lifting · 70-84 strong · 85-100 category-shifting.
The seven signals · what each one says, where it comes from
Competitive overlay · pulse vs category vs closest competitor
This brand's pulse runs +9 above category benchmark and +6 above closest named competitor over the trailing 12 weeks. Panels can't do this in real time. Continuous competitive tracking is the read CMOs ask for and survey vendors charge $30K-80K extra to deliver - Brand Pulse runs it free, weekly, baked in.
This brand has been outpacing the category for 12 weeks running and pulled away from the closest competitor in the last 6 weeks. The gap to category is widening, not closing - signal TV is doing the work the rest of the category isn't yet. Pull TV down here and the lead compresses.
Each signal is weak alone. Together they're hard to fake. Branded search can be paid-search-driven, so we pair it with non-brand category demand. Direct traffic can be loyal-customer-driven, so we pair it with press and sentiment that surface new audiences. Paid engagement can be creative-driven, so we cross-check it against organic. No single signal anchors the read - the convergence is what gets it past a CFO audit. When five rise together, the brand is moving for real. When they diverge, you have a diagnosis.
ComparisonWhat this is - and what it isn't (vs Disqo / Kantar)
| Disqo / Kantar (panel-based) | Stable Brand Pulse | |
|---|---|---|
| Cost to brand | $80K-$200K+ per year | $0 to brand · subsidized by UA partnership economics |
| Signal source | Opted-in survey panel · same person, exposed vs control | Seven public signals · API ingest · no panel |
| Output | Aided / unaided recall point-deltas, intent, favorability | Directional 0-100 index · convergence read |
| Cadence | Quarterly study windows | Weekly read · daily ingest |
| When to still want Disqo | When CFO needs a survey-grade aided/unaided recall number for the board, or when running competitive brand-tracking across categories. Stable Brand Pulse complements - doesn't replace. | |
MethodologyHow the BLI is calculated · signal weights · what gets greyed out
Composition. Seven signals, each converted to a lift z-score versus a 28-day pre-flight baseline (or matched-control region where available), capped at ±3, rescaled to a 0-100 sub-score, composed weighted into the BLI.
Weights (Stable-set defaults, fixed across the book): Branded search 22% · Non-brand category search 14% · Direct + organic traffic 18% · Paid signal 12% · Press chatter 9% · Online sentiment 16% · Wikipedia + Reddit pulse 9%. Weights are not adjusted in flight to make a brand's BLI look better. Per-brand re-weighting requires brand sign-off and is logged in the methodology refresh memo. The default applies unless the change is documented.
Confidence rules. Sub-scores grey out and are excluded from the blend if (a) baseline period < 14 days, (b) signal source unavailable for the brand, or (c) signal volatility exceeds the noise floor. Index reweights surviving signals.
Signal sources, all public or first-party. Google Trends · GSC · GA4 · UA + Meta + YouTube reporting · Google News + RSS press feeds · Stable-built sentiment ingestor over Reddit + X + YouTube + Wikipedia talk pages · Wikimedia REST API + Reddit API.
The pitch to UA brand partners. Every brand on UA gets Brand Pulse out of the box. Free. Always-on. Directional. If you need a Disqo-grade aided-recall number for the board next quarter, that's a separate $80K-$200K spend with a vendor we'll happily layer on top. If you need a weekly read that tells you whether the brand is moving, you already have it.