The data MH2 ingests, scores, and surfaces at every phase — from raw customer records to executive KPIs, creative strategy, and optimization logs.
Data Layer
Intelligence Layer
Execution Layer
Customer Data (raw)
One row per customer · ingested at kickoff
Customer IDUnique identifier
First Purchase DateAcquisition date
RevenueTotal lifetime revenue
Orders# of purchases
AOVRevenue ÷ Orders
Margin %Gross margin
Channel SourceAcquisition channel
Product CategoryPrimary product
RegionGeography (optional)
Value Signal
High / Mid / Low clusters scored by MH2
Cluster NameHigh / Mid / Low Value
CriteriaRules (e.g. LTV > $300)
Customer Count# in cluster
RevenueTotal from cluster
Avg LTVAvg revenue per customer
Avg MarginAvg margin %
Repeat Rate% with >1 purchase
Creative Strategy
Profile-to-channel ad strategy matrix
ProfileLinked Value Profile
HookOpening concept
AngleStrategic message
OfferDiscount / bundle / CTA
FormatVideo, static, UGC
ChannelMeta, Google, etc.
HypothesisWhat you're testing
Value Customer
Filtered subset · same schema, high-value only
Customer IDUnique identifier
RevenueTotal lifetime revenue
AOVRevenue ÷ Orders
Margin %Gross margin
Channel SourceAcquisition channel
Product CategoryPrimary product
Voice of Customer
Extracted quotes and behavioral themes
Response IDUnique entry
Customer IDLink to customer
Key PhraseExtracted quote
Themeprice, quality, speed…
ClusterLinked value cluster
Media Performance
Channel-level efficiency tracked weekly
ChannelMeta, Google, etc.
SpendTotal spend
CACCost per acquisition
LTVAvg LTV from channel
CAC:LTVEfficiency ratio
Value DensityProfit efficiency
% High-Value% from top cluster
↓
Feeds MH2 Intelligence
Data Layer outputs flow into clustering, scoring, and profiling
Value Profiles
VoC-driven personas — core intelligence output
Profile NameHuman-readable name
ClusterHigh / Mid / Low
Key TraitsBehavioral + economic
Core MotivationWhy they buy
Pain PointsWhat they care about
Top SignalsBest-performing signals
Creative AnglesMessaging direction
Optimization Log
Every change tracked with outcome + impact
DateWhen change occurred
ChangeWhat was adjusted
ProfileValue Profile impacted
HypothesisExpected outcome
ResultWhat happened
ImpactRevenue / CAC / LTV Δ