B2Metric · Identity Resolution

One customer. One profile. Every channel.

Your customer is one person — your data says they're six. B2Metric stitches cookies, device IDs, emails, phone numbers and CRM records into a single resolved identity that every segment, journey and model can trust.

duplicate records before resolution: 3.4×; duplicate records before resolution
match precision on deterministic keys: 99.2%; match precision on deterministic keys
to resolve an identity at event time: <200ms; to resolve an identity at event time

1 profile · 6 identifiersUn-merge

Webanon_8f21c4iOS appidfv_3d9e17Androidaaid_7b02d5CRMu_44192Emailsha256:9c1d…Call centertkt_20714

Cookie ID
IDFV
AAID
Customer ID
Hashed email
Phone E.164

4.8Capterra
4.9Product Hunt
4.6G2
5.0Gartner

Why it matters

Fragmented identity quietly breaks everything downstream

Every analysis, campaign and model inherits the mistakes in your identity layer.

Inflated audience counts
The same person counted three times turns every reach, conversion and CAC number into fiction.
Contradictory messaging
A customer who just bought on the app still gets the acquisition campaign meant for strangers.
Models learn the wrong thing
Churn and LTV models trained on fragmented records learn from noise instead of behavior.
Consent gets missed
An opt-out captured in one system doesn't reach the others — a compliance risk, not just a UX one.

How it works

From scattered records to one resolved person

01
Ingest every signal
Web, mobile, CRM, call center, POS and warehouse events land in one stream — no schema rewrite required.
02
Match on your rules
Deterministic keys first, then optional probabilistic matching for the records exact keys can't reach.
03
Build the golden profile
Survivorship rules pick the winning value per field, so one person resolves to one trusted record.
04
Activate everywhere
The resolved profile feeds segmentation, journeys, churn models and every downstream channel.

Matching engine

Your rules, your confidence threshold.

Start with deterministic keys — user IDs, normalized emails, E.164 phone numbers — where a match is a fact. Then decide whether probabilistic matching should reach the records exact keys can't. Toggle a rule and watch the identity count settle in real time.

Order rules by precedence; higher-confidence keys win first.
Tune the probabilistic threshold instead of accepting a black box.
Every merge is logged with its rule, score and timestamp — and reversible.

Match rule set

Applied in order

Exact user ID (applied)user_id = user_id100%Exact
Normalized email (applied)lower(trim(email)) match99%Exact
Phone number (E.164) (applied)+90… normalized match98%Exact
Device ↔ login session (applied)device_id seen with user_id96%Exact
Name + address similarity (off)jaro_winkler > 0.9287%Fuzzy
Household grouping (off)shared address or network74%Fuzzy

Resolved people

44,400

Duplicates collapsed

65.4%

from 128,400 raw records

Person by person

What resolution actually looks like

Each of these customers arrived as several disconnected records. After resolution, one profile — with the full history attached.

resolved
Elif Demir
Champions · ₺18.4K LTV
6records1profile
- Web cookie
- iOS IDFV
- CRM ID
- Email
- Phone
- Loyalty
resolved
Can Aydın
Loyal · monthly buyer
4records1profile
- Android AAID
- Email
- POS card
- Support ticket
resolved
Zeynep Kaya
New · first session
3records1profile
- Web cookie
- Email
- Push token
resolved
Mert Çelik
At risk · dormant 34d
5records1profile
- Web cookie
- iOS IDFV
- Email
- Phone
- CRM ID

Golden profile

When sources disagree, survivorship decides

Two systems hold two different emails for the same person. Pick the policy — the winning value updates everywhere the profile is used.

Survivorship policy

Most trusted source
Most recent update
Most complete value

Email
3 sources
- elif.demir@corp.com.trCRM · 118d agowinning value
- elif.d@gmail.comWeb form · 4d ago
- elif.demir@corp.com.trMobile app · 27d ago
Phone
2 sources
- +90 532 118 44 07Call center · 9d agowinning value
- 0532 118 4407Web form · 61d ago
Full name
2 sources
- Elif DemirCRM · 210d agowinning value
- Elif Demir YılmazMobile app · 12d ago
City
3 sources
- AnkaraCRM · 300d agowinning value
- İstanbulWeb form · 6d ago
- İstanbul / KadıköyPOS · 18d ago
Marketing consent
2 sources
- Granted — email onlyCRM · 95d agowinning value
- Granted — email, pushMobile app · 8d ago

Elif Demir

1 person · 6 identifiers · 3 devices

person_0x4fa19c

Champions
LTV ₺18.4K
Push opted-in
iOS + Web
TR / Istanbul

One canonical ID every downstream system references.
Field-level lineage: see which source won and why.
Re-resolves automatically as new evidence arrives.

Downstream

Resolve once, benefit everywhere

Identity isn't a feature you look at — it's the layer everything else quietly depends on.

Segmentation
Every micro-segment counts people, not records — so reach numbers finally match reality.
Flowly journeys
One journey state per person, whichever device or channel they show up on next.
Churn & LTV models
Models train on complete behavioral histories instead of three partial ones.
Campaigns & push
Frequency caps, suppression and consent apply to the human being, not the cookie.

Governance

Built for regulated data teams

Hashed & encrypted
PII is hashed at collection; matching runs on hashes, never raw identifiers.
Consent-aware
Consent and opt-out states resolve with the profile and propagate to every channel.
KVKK & GDPR ready
Right-to-be-forgotten deletes the whole identity graph, not just one record.
Auditable
Every merge keeps its rule, confidence score and timestamp — and can be unmerged.