What is Identity Resolution?

Identity resolution follows a four-stage pipeline. First, identifiers are collected from every available source: website pixels, CRM records, form fills, email engagement, product analytics, ad-platform data, and offline signals like event attendance or direct mail responses.

Second, a matching engine applies rules to connect those identifiers into unified profiles. Deterministic matching links records using confirmed, exact data — a shared work email address, a user-verified phone number, a CRM ID. It is highly precise but requires that the identifying data exist in the first place. Probabilistic matching fills the gaps: if two devices consistently connect from the same corporate IP, share browsing patterns, and hit the same product pages, statistical modeling assigns them to the same account with an explicit confidence score.

Third, the resolved identities are written into an identity graph — a centralized data structure that maps all known attributes per person and per account. In B2B, this means linking individuals to their employer, to their buying-committee peers, and to firmographic data like company size, industry, and tech stack. The fourth step is activation: pushing verified, resolved records into CRM workflows, ad-audience segments, ABM orchestration platforms, or sales sequences.

Deterministic matching is exact. It relies on verified first-party identifiers — a hashed work email, a logged-in user ID, a confirmed phone number — to link records with near-certainty. LiveRamp's deterministic graph has been independently tested to achieve 95–98.9% precision. The trade-off is coverage: if a visitor never logged in or submitted a form, there is nothing to match deterministically.

Probabilistic matching trades precision for reach. Algorithms analyze co-occurrence patterns — two devices sharing an IP address, a device fingerprint consistent across sessions, behavioral similarity — and assign identity links with a stated confidence level. This approach scales to anonymous traffic but introduces a margin of error, and error rates compound when probabilistic matches are stacked or used for high-stakes one-to-one personalization.

Sophisticated B2B platforms (6sense, Demandbase, Bombora) blend both: deterministic resolution anchors known contacts to accounts, while probabilistic inference extends coverage to the anonymous majority of traffic. The best systems attach explicit confidence scores to every resolution so downstream teams know which records to trust for direct outreach versus broader audience targeting.

The B2B buying journey is structurally invisible. According to 6sense's 2024 Buyer Experience Report, 81% of buyers have already selected a preferred vendor before making first contact with a salesperson, and nearly 70% of the buying journey happens before buyers engage with any vendor at all. Without identity resolution, sales teams see only the 2–3% of visitors who fill out a form — the rest is a black box.

This invisibility is compounded by buying-committee complexity. Gartner reports that a typical B2B purchase involves 6–10 decision-makers. A single form fill from one stakeholder tells almost nothing about where the full committee sits in the buying journey. Identity resolution connects the dots: an anonymous product page visit from a VP of Engineering at a target account, combined with a webinar signup from the CFO and a pricing page hit from the procurement lead, collectively signal high readiness — but only if all three are resolved to the same account.

The market consequences are significant. The Forrester Total Economic Impact study commissioned by LiveRamp (June 2025) found that enterprises using data collaboration infrastructure anchored in identity resolution achieved a 313% ROI over three years, driven by a 15% improvement in paid-media efficiency and over $18 million in incremental revenue from more accurate measurement and audience activation.

An identity graph is the data structure that identity resolution produces and maintains. It is a centralized, continuously updated map that links every known identifier for a person or account — email addresses, device IDs, cookie values, phone numbers, CRM IDs, hashed PII — to a single persistent node. The identity graph is the output; identity resolution is the process that builds and refreshes it.

In B2B, identity graphs operate at two levels simultaneously. At the person level, the graph tracks an individual's identifiers across their devices and channels. At the account level, it links individual nodes to employer records and to one another within a buying committee. This dual-layer structure is what separates B2B identity graphs from B2C equivalents, which are typically household-centric.

Bombora's identity graph maps more than 3 billion digital identifiers to business attributes across thousands of business domains, covering the B2B co-op data of nearly 4.8 million unique domains. LiveRamp maintains PII on 245 million US individuals in its deterministic graph. The scale of the underlying graph directly determines match rate: a resolution platform can only identify a visitor if that visitor's identifiers are already represented in the graph.

Third-party cookies were the original backbone of probabilistic identity resolution across the open web, but their reach has been eroding for years. Safari's Intelligent Tracking Prevention and Firefox's Enhanced Tracking Protection mean that 34.9% of US browsers already block third-party cookies by default (eMarketer, 2026). More than half of all mobile impressions now carry no user identifier at all, and the share is growing.

The industry response has been a shift to durable, consent-based identifiers. Universal IDs like LiveRamp's RampID and The Trade Desk's UID2 are built from authenticated first-party data — typically a hashed email — that persists across channels without relying on cookies. Data clean rooms, adopted by 66% of US data and ad professionals (IAB and BWG Strategy, 2024), allow brands and publishers to match identity graphs without sharing raw PII, enabling attribution and audience collaboration under privacy constraints.

For B2B specifically, IP-to-company matching has become the primary probabilistic fallback. In ideal enterprise conditions — office-based employees on static corporate IP blocks — match rates reach 60–80% at company level. In practice, remote and hybrid work patterns reduce accuracy significantly, which is why the most resilient B2B identity strategies layer IP matching with first-party CRM data, product telemetry, and gated content to anchor probabilistic signals to confirmed identities.

Komo — the AI Revenue Engine — operates in the space where identity resolution delivers its highest B2B value: converting anonymous or fragmented buying signals into qualified, contextual outreach. When a target account's employees visit your site, engage with content, or show intent signals across third-party networks, Komo monitors those signals and resolves them to specific contacts and accounts in your CRM, so reps see an enriched, deduplicated picture of who is showing up and why.

Rather than asking reps to manually cross-reference a de-anonymized visitor list against their CRM, Komo automates the signal-to-research-to-draft pipeline. A resolved identity — a VP of Sales at a Series B fintech who just hit your pricing page for the third time — triggers research, a tailored draft, and a queued task for the rep to review and send. The human is in the loop on every send that matters; Komo handles the repetitive stitching work in between.

This matters because identity resolution on its own produces a list. Acting on that list with the right message, to the right stakeholder, at the right moment requires the layer Komo provides: signal monitoring, buying-committee research, contextual drafting, and human judgment — all wired together so that resolved identity becomes revenue, not just data.