LinkedIn Conversions API for B2B attribution

B2B attribution breaks differently from consumer ecommerce. A buyer might click a LinkedIn ad on a phone, read the page from a laptop, ask a colleague to evaluate the product, book a demo from a work calendar, and become an opportunity weeks later inside a CRM. The browser event that started the journey is useful, but it rarely represents the business outcome sales and finance care about. LinkedIn Conversions API gives your server, CRM, or tracking layer a direct way to send those later events back to the conversion rule that campaigns optimize against.

This matters because LinkedIn campaigns are often judged on lead quality, target-account penetration, pipeline created, and meetings with the right buying committee, not only raw form volume. A pixel-only setup can miss blocked scripts, delayed consent banners, cross-device journeys, CRM qualification, and long sales-cycle updates. CAPI does not solve weak strategy, but it does let you report cleaner evidence: who converted, when the lifecycle step happened, which account it belonged to, and what value the business assigned to it.

Build the integration in narrow layers. Prove the endpoint and headers, then add the required fields, then add normalized identifiers, then connect CRM events and logging. That sequence keeps every failed request tied to one clear change.

POST https://api.linkedin.com/rest/conversionEvents
Authorization: Bearer LINKEDIN_ACCESS_TOKEN
LinkedIn-Version: 202601
X-Restli-Protocol-Version: 2.0.0
Content-Type: application/json

{
  "conversion": "urn:li:sponsoredConversion:123456",
  "conversionHappenedAt": 1776943226000
}

{
  "conversion": "urn:li:sponsoredConversion:123456",
  "conversionHappenedAt": 1776943226000,
  "eventId": "demo_98172_booked",
  "value": {
    "amount": "2500.00",
    "currencyCode": "USD"
  },
  "user": {
    "userIds": [
      {
        "idType": "SHA256_EMAIL",
        "idValue": "0d741ab04ce9..."
      }
    ],
    "userInfo": {
      "firstName": "f52fbd32b2b3...",
      "lastName": "a94a8fe5ccb...",
      "companyName": "acme software"
    }
  }
}

import crypto from "node:crypto";

function sha256(value: string) {
  return crypto
    .createHash("sha256")
    .update(value.trim().toLowerCase())
    .digest("hex");
}

const userIds = [
  { idType: "SHA256_EMAIL", idValue: sha256("buyer@example.com") },
  { idType: "SHA256_PHONE_NUMBER", idValue: sha256("+14155550123") }
];

{
  "conversion": "urn:li:sponsoredConversion:123456",
  "conversionHappenedAt": 1776944209000,
  "eventId": "opp_44381_sql",
  "value": {
    "amount": "18000.00",
    "currencyCode": "USD"
  },
  "user": {
    "userIds": [
      { "idType": "SHA256_EMAIL", "idValue": "0d741ab04ce9..." }
    ],
    "userInfo": {
      "firstName": "b6d767d2f8ed...",
      "lastName": "3e23e8160039...",
      "companyName": "acme software"
    }
  }
}

POST /rest/conversionEvents
Authorization: Bearer LINKEDIN_ACCESS_TOKEN
LinkedIn-Version: 202601
X-Restli-Protocol-Version: 2.0.0
Content-Type: application/json

{
  "conversion": "urn:li:sponsoredConversion:123456",
  "conversionHappenedAt": 1776944209000,
  "eventId": "lead_77821_submit",
  "user": {
    "userIds": [
      { "idType": "SHA256_EMAIL", "idValue": "0d741ab04ce9..." }
    ]
  }
}

Match rate estimates vary by region, consent rate, job seniority, work email quality, and how often buyers use personal identities while researching business software. Treat the table as a prioritization model rather than a universal benchmark.

Account-based marketing needs a wider lens than person-level conversion counting. One LinkedIn member might research the product, another might attend the webinar, and a third might book the meeting. If those events are treated as unrelated people, the campaign can look weak even while the target account is moving forward. Your server-side event envelope should therefore keep account domain, CRM account ID, company name, lifecycle stage, opportunity ID, and campaign membership alongside the LinkedIn payload.

LinkedIn CAPI should receive only supported, policy-compliant fields, but your own logs can preserve the account graph that explains the journey. That lets media, demand generation, sales operations, and revenue operations reconcile LinkedIn reporting with Salesforce, HubSpot, or a warehouse model. For ABM, the practical win is not just more conversions. It is the ability to see whether paid LinkedIn activity is creating qualified movement inside named accounts.

Campaign Manager validation should happen before the first live campaign depends on server events. Test a realistic B2B journey, including form submit and CRM stage changes, because many CAPI issues appear only after the browser event hands off to marketing automation or sales operations.

LinkedIn attribution windows can extend up to 90 days, which is unusually important for B2B. A buyer who clicked an ad in January may not become a sales-qualified opportunity until March. If your server event uses the lifecycle timestamp and lands inside the configured window, LinkedIn has a better chance of connecting campaign exposure to the business result. If your integration sends only the import timestamp, the event can look artificially late and fall outside the period the campaign should receive credit for.

The operational rule is simple: preserve the moment the buyer actually performed the relevant action. For demo requests, use the submit or booking time. For MQL and SQL, use the CRM stage-change time. For opportunities, use the created date or accepted sales date agreed by revenue operations. Long windows are valuable only when timestamps are honest and event IDs stay stable through retries.

Most LinkedIn CAPI failures are integration contract problems: wrong account, wrong rule, missing permission, stale token, malformed URN, invalid timestamp, weak hashing, or duplicate sends from CRM imports and browser tags. Fix them by logging the final request body and response, not by guessing from reporting screens alone.