AD PERFORMANCE INTELLIGENCEv2.4 · MAY 2026

Know exactly why your Meta campaigns moved.

Stop spending mornings in Ads Manager. Sentrum detects what changed, explains why, and tells you the next move — before ROAS craters.

Start free trial See live demo7-DAY TRIAL · 2-MIN SETUP

DetectDiagnoseRecommendAct

✦WHAT'S NEW

Full changelog →

NEW · MAY 2026

Introducing Blueprint

Decode your account, layer by layer. The seven-section AI diagnostic walks the full cause-tree from spend to signal quality — in one screen.

See Blueprint

NEW · MAY 2026

Signal Quality on every account

EMQ, CAPI coverage, and AEM event health surfaced for every campaign. Andromeda's most under-diagnosed CPA inflation cause, finally called out.

How it works

NEW · MAY 2026

Built for the MCP era

Run Sentrum from Claude Code, Claude Desktop, Cursor, or any MCP client. Diagnose campaigns from inside your AI agent — without context-switching.

Connect to Claude

WHAT YOU GET

Not a reporting tool. A decision engine.

Every screen answers the same four questions: What changed, why, so what, now what. Plus the one thing Ads Manager and generic AI never give you: a learning record of your own account.

WHAT HAPPENED?

Rolling anomaly flags

Real-time anomaly detection on CPM, CTR, ROAS, frequency, and 40+ derived signals. Catch creative fatigue and pacing drift before they cost you — not in next week’s review.

WHY DID IT HAPPEN?

Multi-pass diagnostics

Sentrum walks the dependency tree across audience, creative, and bid signals. Causes ranked, evidence cited. Specific drivers, not "the algorithm is learning."

WHAT DO I DO?

Confidence-scored fixes

Every diagnosis ships with a recommended action tied to the diagnosed cause, plus an estimated impact window. Not generic advice. Not "test new creative."

WHAT WORKED LAST TIME?

Account playbook intelligence

Sentrum builds knowledge on your historical performance — what works, what doesn’t, what to repeat per account. Context that pasting CSVs into a generic LLM can’t replicate.

WHAT DID I JUST DO?

Audit trail you can hand a client

Every change tracked, reversible, optionally human-approved. No black box. Defensible. The log is the report.

HOW ARE WE DOING?

Reports on your schedule

Weekly, monthly, and campaign-wrap reports — auto-generated, email-delivered. Hours of spreadsheet labor compressed to inbox-ready decks.

SEE THE LOOP IN ACTION

HOW IT WORKS

Four steps from signal to shipped fix.

STEP 01 · DETECT

Catch shifts the moment they happen

Real-time anomaly detection across CPM, CTR, ROAS, frequency, and 40+ derived signals. No waiting for the weekly review.

Per-campaign and per-ad-set baselines
Statistical, not threshold-based
Slack & email alerts you can actually act on

CPM · MIDNIGHT PHASE2

$12.40 +38.2%

ANOMALY

APR 24MAY 10 · NOW

STEP 01 · DETECT

Catch shifts the moment they happen

Real-time anomaly detection across CPM, CTR, ROAS, frequency, and 40+ derived signals. No waiting for the weekly review.

CPM · MIDNIGHT PHASE2

$12.40 +38.2%

ANOMALY

APR 24MAY 10 · NOW

STEP 02 · DIAGNOSE

Find the why, not just the what

Sentrum walks the diagnostic order — spend, efficiency, creative resonance, mix shifts, signal quality — and names the driver. Not a vibe, not a chart, an actual cause.

SAMPLE INSIGHT · ASSET FATIGUE

UGC-v2 hit creative saturation: ROAS −38% with consistent spend.

WHAT

UGC-v2 (28 days in market, no refresh). Spend held at $420/day. ROAS 4.2× → 2.6× over 14d. CTR −45% PoP. CPM held flat.

WHY

Asset fatigue. CTR + VTR both decay while CPM holds = creative-side problem, not auction pressure or audience saturation.

SO WHAT

Severe — top-spend creative compounds the loss. ~$1.7K wasted spend / week if untouched. Confidence: high.

NOW WHAT

Pause UGC-v2. Brief 2 net-new archetype variants this week (UGC testimonial + before/after demo). Watch: cost-per-purchase recovery in 72h.

STEP 03 · RECOMMEND

Get a fix tied to the diagnosed cause

Every recommendation maps to the specific driver Sentrum just diagnosed. Includes the watch-metric to verify the move worked.

SAMPLE INSIGHT · AUDIENCE OVERLAP

Consolidate 3 prospecting ad sets — 40% auction overlap.

WHAT

LAL-1%, LAL-2%, Interest:Coffee bidding against each other 40% of the time. Combined CPMr +28% vs your 90-day baseline. Frequency at 4.8x.

WHY

Audience overlap. Three sets target overlapping seed pools — algorithm bids against itself, inflating CPM without adding unique reach.

SO WHAT

~$1.2K/week burned on duplicate auction wins. Frequency saturation imminent. Confidence: high.

NOW WHAT

Consolidate into one ad set with combined audience exclusions. Or migrate to ASC if catalog-driven. Watch: CPMr −15-20%, unique reach +30% in 5d.

STEP 04 · ACT

Push the change without leaving the dashboard

Approve a fix and Sentrum writes the budget shift, audience exclusion, or creative rotation back to Meta — with an audit trail.

AUDIT LOG · LIVE

14:02

Excluded LAL-1% on Midnight Phase2

by sentrum

APPLIED

14:02

Notified Slack #ads-meta

by sentrum

APPLIED

14:01

Approved fix (1-click)

by a.arango

HUMAN

13:58

Flagged anomaly · confidence 92%

by sentrum

SYSTEM

SEE IT IN ACTION

Four screens, every diagnostic the team runs.

DAILY PULSE

One-screen morning brief — what moved, what to fix, what to scale

Real-time anomaly detection on CPM, CTR, ROAS, frequency, and 40+ derived signals.
Funnel balance scoring (TOF/MOF/BOF) with spend allocation per stage.
Automatic "Needs Attention" and "Scaling Opportunity" flags with next steps.
Rolling baselines per account so what counts as "abnormal" is your normal.

Greenfield Coffee Co.

Daily Pulse

Funnel Health

TOF 65%

$5,478 · 198 results

MOF 20%

$1,685 · 52 results

BOF 15%

$1,264 · 34 results

Spend Efficiency TrendsCPM trending up >15%

CPM

CPC

CPR

Campaign Health

48 ad sets, 156 ads

ad sets <50 conv/7d

18d

avg since refresh

87%

budget utilization

Needs AttentionACTIVE

Subscription Box - Retargeting

Spend

$980

Results

Cost/Result

$35.00

VS KPI AVG

+42%

CPR 42% above account avg — $18.50 vs $13.02 avg for Purchases. → Identify least efficient ad sets and pause or restructure.

Performing Well

TOF - Broad Prospecting

Scaling opportunity — stable performance. Consistent $12.40 CPR over 14 days. → Test 15-20% budget increase over 3 days.

CAMPAIGN DEEP DIVE

Tie every campaign change to its performance shift

Optimization log overlay — every budget edit, audience tweak, and creative swap, mapped to the metrics that followed.
AI connects creative changes to performance shifts — not just that CTR dropped, but which ad rotation caused it.
Surfaces which optimizations actually moved the needle, and which didn't.
Demographic breakdown with efficiency scoring by age and gender.

Greenfield - TOF - Broad ProspectingACTIVE

Jan 15 - Feb 12, 2026Had DeliveryLast 14 days

3 of 4 ad sets are in learning phase — consider consolidating

Frequency is 3.2x — audience may be approaching saturation

Total Spend

$24.3K

-5.3% vs prev

Purchases

812

+12% vs prev

CPA

$29.93

-8% vs prev

CPM

$18.42

+1.7% vs prev

Impressions

1.3M

-10.4% vs prev

Frequency

3.21

-3.8% vs prev

Performance Over Time

SpendReachCPRCPMCTRImpressions

Spend

Purchases

CPA

Optimization Log

2026-02-08

New creative launched — "Cold Brew Morning Routine" by Sarah M.

2026-02-05

Budget increased: $400 → $500/day

2026-02-01

Ad set paused — "Interest: Latte Art" low conversion rate

Age Breakdown

AgeSpendReachCPR

18-24

$2,180

142K$34.50

25-34

$8,940

386K$24.12

35-44

$6,720

298K$28.40

45-54

$3,890

178K$31.20

55-64

$1,820

84K$38.70

65+

$750

32K$42.90

Female — 75%

Spend: $18.2K · CPR: $26.40

Male — 25%

Spend: $6.1K · CPR: $34.80

CADENCE REPORTS

Reports your clients ask for, generated on schedule

Weekly, biweekly, monthly, and campaign-wrap reports.
Campaign metrics with PoP comparison + creative / demographics / placement analysis.
Bulk AI insight generation across all active campaigns.
Email delivery to your team or clients on your cadence.

Saved Reports

Report	Range	Campaigns	Status	Last Generated
Weekly Performance	Feb 3-9, 2026	All (5)	Generated	Feb 10, 9:00am
Monthly Summary	Jan 2026	All (5)	Generated	Feb 1, 9:00am
TOF Deep Dive	Feb 3-9, 2026	Broad Prospecting	Scheduled	Pending

Greenfield Coffee Co.

Weekly Performance Report — Feb 3-9, 2026

Auto-generated every Monday at 9am · Delivered via email

Spend

$8,427

+12% vs prev

Purchases

284

+14% vs prev

CPA

$29.67

-8% vs prev

ROAS

3.2x

+6% vs prev

Campaign	Spend	Results	CPA	ROAS
TOF - Broad Prospecting	$3,241	112	$28.94	3.4x
Summer Cold Brew Launch	$2,180	84	$25.95	3.8x
MOF - Retargeting	$1,560	52	$30.00	3.1x
BOF - Purchase Converters	$980	28	$35.00	2.6x

ASK AI

Plain-English questions, grounded in your account history

Natural language Q&A over your real performance data — not a pasted CSV.
Sentrum references your historical baselines, patterns, and optimization log.
Specific actionable recommendations — not boilerplate.
Ask about any metric, time period, or comparison.

Ask AI

Greenfield Coffee Co.

Ask anything about your campaigns...

Which ads are working best on Reels and why?

Your top 3 Reels ads drive 67% of purchase volume and share one pattern: creator UGC with product before/after in the first 3 seconds.

(1)Hook rate 18.4% on UGC-creator format vs 11.2% account avg
(2)Catalog carousels are underperforming on Reels — long videos aren't showing UVP + CTA fast enough for the format
(3)Best performer (“Sarah's morning routine”): 4.9× ROAS, 31% CTR, 54s avg watch time

Brief: 3 net-new creator UGC ads, before/after structure, 15s + 30s cuts, full vertical, no music. Avoid: studio product shots, voiceover-only, 45s+. Watch: hook rate above 15% within 7d.

What format is winning?Top hook in last 14d?Catalog vs videoBest creator ROAS

VS META'S BUILT-IN AI

Meta's AI optimizes for Meta. Sentrum optimizes for you.

Meta Built-in AI

Single-platform (Meta only)
Generic optimization, no diagnostic depth
Acts on your account silently — no audit trail
Optimizes for Meta's objectives, not your margin
Black-box — can't inspect or override logic

Sentrum

Multi-platform diagnostic (Meta + Google + more)
Reads cause + effect, not just outcomes
Every action logged, reversible, human-approvable
Profitability-aware (CAC, COGS, contribution margin)
Transparent — every diagnosis cites the data

BUILT FOR THE MCP ERA

Run Sentrum from Claude, Cursor, or any MCP client.

Sentrum ships a first-class MCP server. Diagnose campaigns, pull creative analysis, and ask follow-ups — all from the AI assistant you already use. No new dashboard to log into.

Five diagnostic tools: getCampaignDiagnostics, getCreativeAnalysis, getAccountContext, getRecommendations, askQuestion
JSON-RPC 2.0, API-key authenticated, scoped to your workspace
Works with Claude Code, Claude Desktop, Cursor, and any MCP-compatible client

Connect to Claude Read the walkthrough

CLAUDE CODE · TERMINAL
> claude mcp add sentrum
✓ Connected. 5 tools available.
> Why did ROAS drop on Campaign 8421 yesterday?
↳ getCampaignDiagnostics
ROAS −34% PoP. Driver: CPM +28% (auction pressure, broad placement mix shifted toward Reels). CTR stable. Recommendation: pin manual placements to Feed + Stories for 48h, monitor.

PRICING

Enterprise-grade diagnostics, indie pricing.

Plans from $19/mo. 7-day trial. No setup fees. Cancel anytime.

See full pricing Start free trial

START FREE

Stop chasing yesterday’s numbers. Catch them while they shift.

Connect your ad accounts in two minutes. Sentrum starts diagnosing on day one — no implementation cycle, no consulting hours.

Start free trial Book a demo

7-DAY TRIAL · 2-MIN SETUP