Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

CRM & Marketing Automation › CRM for Marketers

Syncing the CRM and the Marketing Platform

How to sync a CRM with a separate marketing automation platform: which records sync, who owns each field, conflict rules, deletions, consent and sync errors.
Guideintermediate12 min readUpdated

Two systems, one record

The marketing automation pillar sets the rule this guide depends on: the CRM is the system of record, and when the CRM and the email platform disagree, the CRM wins and the integration that should have told the other system is broken. The rule is easy to agree with and hard to implement, because a sync is not one connection. It is dozens of small decisions, one per field and one per kind of event, and most teams accept whatever the connector does by default.

Those defaults produce a recognisable set of symptoms. A lifecycle stage set by sales reverts overnight. A person who unsubscribed in the marketing platform is still marked as reachable in the CRM, so a salesperson adds them to a manual sequence. Two records for the same person exist in the marketing platform because the CRM merged them and the marketing side did not hear about it. A score calculated in the marketing platform arrives in the CRM hours after the salesperson has already called. None of these raises an error that anyone reads.

The fix is a short written agreement covering the parts below. The CRM fundamentals guide already asks that integrations authenticate as their own named users and that field names say which system writes them; this guide adds the rest of the contract. The CDP versus CRM comparison is the place to decide whether a third platform belongs between the two at all; the contract here assumes there are two.

Deciding which records sync

The first decision is which records cross, and it is usually made by accident: whatever the connection was set up to do on the day it was switched on. As general practice, write the rule as a condition on fields the CRM owns, so anyone can read it and test it.

Records that normally sync to the marketing platform are the people it may legitimately market to: contacts with a consent record for the relevant channel or another basis the organisation has decided applies, leads in the stages marketing works, and customers enrolled in lifecycle programmes. Records that normally stay behind are internal users, test and seed records that are not meant for marketing, records held for service, billing or legal reasons only, and contacts whose only role is as a name on a deal. Companies and deals sync only as far as the marketing platform needs their fields for conditions and segments, often as a few values copied onto the contact rather than whole records.

Two classes of people need care. Contacts who have objected or unsubscribed still have to be known to the marketing platform, as suppressed records, so that a later import or form submission cannot create them as new and reachable; the suppression list entry explains why suppression and deletion are different things. And contacts with no consent record at all are better kept out of the marketable set until the record exists, rather than synced on the assumption that someone will check later. The principle of data minimisation in the GDPR glossary entry points the same way: a copy of Personal Data held in a second system for no purpose is a copy with risk and no benefit. Agree the rule with whoever owns data protection; this guide is not legal advice.

Field ownership and sync direction

Every synced field needs exactly one owning system, one accountable team and one direction. The RACI matrix framework is a convenient way to write it, because it insists on a single accountable party per row. A field that two systems may write has no owner, and its value is whatever was saved last.

Field or groupOwning systemAccountable teamDirectionNotes
Email address and nameCRMSales operationsCRM to marketingChanges made in a form update the CRM first, then flow back
Record ownerCRMSales operationsCRM to marketingRead by the marketing platform for sender names and alerts
Lifecycle stageCRMMarketing operations, with sales sign-offCRM to marketingThe marketing platform requests a change through a CRM workflow, not by writing the field
Deal stage and amountCRMSalesCRM to marketing, as summary values on the contactThe marketing platform never writes deal data
Original source and campaign parametersCRM, written once at creationMarketing operationsMarketing to CRM on creation onlyNever overwritten by a later sync
Email engagement and form historyMarketing platformMarketing operationsMarketing to CRMRead-only in the CRM
Lead scoreMarketing platformMarketing operationsMarketing to CRMRead-only in the CRM; thresholds agreed with sales
Consent and unsubscribe statusWhichever system recorded the eventMarketing operations, with the data protection ownerBoth ways, most restrictive value winsCovered in its own section below

A few rows deserve comment. Lifecycle stage is the field most often written by both sides, because a scoring workflow on the marketing side is often built to advance it while sales moves it in the CRM. Pick one writer. If the marketing platform needs to promote a lead, it sets its own field and a CRM workflow applies the change under the CRM's rules, so the stage history stays in one place. The lead score is the reverse case: the marketing platform calculates it and the CRM displays it, and the model, its decay and its thresholds belong to the lead scoring guide. Original source is written once and protected, for the reasons the fundamentals guide gives.

Name fields so the owner is visible, keep the table beside the automation inventory, and make a new synced field a reviewable change, as the automation governance guide does for anything shared. Before such a change goes live, the CRM sync change checklist runs the owners, the consent rules and the seeded records as one tickable gate. The automation inventory shows which workflows read each synced field, so a change to a field's owner or direction can be checked against them first.

Conflicts, merges and deletions

Even with one owner per field, conflicts happen: a person edits their details in a preference centre while a salesperson edits the same record, or a workflow and a sync run against the same field at once. As general practice, decide the rule per field rather than relying on the connector's default, and prefer rules that can be explained: the owning system wins, or for consent the most restrictive value wins. "Most recent edit wins" is the rule that produces flip-flopping, because each system's next save counts as the most recent edit. When a value someone corrected is back to the old one the next morning, the diagnosis for CRM values changing back after a sync reads the record's history to name the second writer, whether it is the sync, a workflow, an import or a migration.

Accuracy is one of the GDPR's principles, as the glossary entry states, and people have a right to have their data corrected. That turns a stale value written back over a correction into more than a reporting problem: if a person corrects their job title or email address and the sync restores the old one, the business is now holding inaccurate data the person has already asked it to fix. Treat corrections made by the person, or by a colleague on their behalf, as values the sync must never overwrite.

Merges need the same care. Deduplication in the CRM follows the merge rules in the fundamentals guide, but a merge in one system does not automatically merge the other's copies. Decide which system merges first (usually the CRM, as the system of record), how the surviving record's identifier is passed to the marketing platform, and what happens to the losing record's engagement history. Without that, the marketing platform keeps two records, sends to both and reports half the engagement on each.

Deletions are the hardest case, because a sync that faithfully copies a deletion can remove the very record that was keeping someone suppressed. Write down what a deletion in each system means. A record removed from the CRM for retention reasons may still need a suppressed entry in the marketing platform; a record deleted on a person's request may need to be removed from both, with only what the organisation's policy says to keep. Those are decisions for the data protection owner; the sync's job is to carry them out consistently, and never to delete suppression entries as a side effect.

Consent is the one group of fields that must travel in every direction. An unsubscribe recorded in the marketing platform has to reach the CRM before a salesperson's manual email goes out, and an objection a person makes on the phone to sales has to reach the marketing platform before the next send. The suppression list entry and the data operations guide it links make the general point: an objection must be applied in every channel it covers, and suppression lists kept separately in each tool drift apart. The UK regulator's direct marketing guidance, listed in this page's sources, is the authority the suppression list entry refers to for honouring opt-outs promptly.

The GDPR glossary entry states that the right to object to direct marketing is absolute: once someone objects, their data is no longer processed for that purpose. That is why the conflict rule for consent fields is the most restrictive value wins, whichever system recorded it and whenever. A sync that lets an older "subscribed" value overwrite a newer objection is the clearest example of a sync failure that is also a data protection failure.

Practical rules follow. Record consent per channel with its source and date, as the fundamentals guide asks, and sync the record rather than a single yes or no. Check suppression before every send in both systems, not only when a record enters a workflow. Keep a seeded test contact that has unsubscribed and objected, in both systems, and check that it stays suppressed after every change to the sync. If that contact, or a real person who opted out, is mailed again, the diagnosis for unsubscribed contacts still receiving email works back from the message to the sync rule or tool that let it through. Outcome uploads to ad platforms follow the same principle, and the guide to connecting the CRM to ads and analytics covers checking consent state before anything leaves for an ad platform.

Watching the sync: errors, counts and drift

A sync that is not watched is a sync that is broken and nobody knows. Three checks cover most failures, and the data quality guide describes the same habits for event data.

The error queue. Where the connection keeps a record of what failed to sync and why (a required field missing, a value not in a picklist, a duplicate identifier, a permission refused), read it on a schedule with a named owner, group errors by cause and fix the cause, not the record. A queue nobody reads grows until the errors are the normal state.

Counts. Compare the number of synced records on each side, by the rule that decides which records sync, and the number of suppressed records on each side. The counts will not match exactly, but the gap should be stable and explained. A gap that grows is the earliest sign that a rule, a field or a credential has changed.

Drift. Sample a handful of records each review and compare owned fields side by side: stage, owner, consent, score. Values that differ on a field with one owner point at a stuck sync; values that keep changing point at two writers. The mechanics behind most drift (append against replace, schema changes, identifier mismatch) are explained in the data flow guide rather than repeated here.

Some failures start upstream, at a lead source rather than in the sync itself. The diagnosis of LinkedIn leads not syncing to the CRM shows how to trace one source's field mapping and connection, and the same reading applies to any form or ad platform that writes into either system. When the CRM itself is replaced, every rule in this guide has to be rebuilt and tested, which is one of the steps in the tutorial on moving to a new CRM.

An accountancy trainer untangles its sync (illustrative)

Illustrative example — not a real result

Oskivar Accountancy Training untangles its sync

Oskivar Accountancy Training, an invented provider of accountancy courses for firms and individuals, runs a CRM for its sales team and a separate marketing platform for email. Sales complains that 140 leads it had marked as sales-qualified in one month were back at marketing-qualified the next morning. The investigation finds that both systems write lifecycle stage: the marketing platform recalculates it nightly from score, and its value overwrites the CRM's. The team builds a field table of 34 synced fields and finds 9 with two writers. Each gets one owner; lifecycle stage becomes CRM-owned, and the marketing platform now sets a separate request field that a CRM workflow reads. The record rule is rewritten too: 3,800 contacts with no consent record had been synced as marketable, and they stop syncing until a record exists, while 620 unsubscribed contacts stay synced as suppressed. A weekly review of the sync error queue is assigned to one named person; the first review finds 210 errors, most from a picklist value added in the CRM and never added on the marketing side. Over the following month the stage reversals stop and the error queue settles at a few new items a week.

Common mistakes

  • Accepting the connector's defaults. The defaults decide ownership, direction and conflicts for you; write the contract instead.
  • Letting both systems write the same field. The value depends on who saved last. One owner per field.
  • "Most recent edit wins" for everything. It produces flip-flopping and lets old consent overwrite new objections. Use owner-wins, and most-restrictive-wins for consent.
  • Syncing every record. People with no consent record, internal users and service-only contacts do not need a marketable copy.
  • Deleting instead of suppressing. A deletion synced across can remove the entry that kept someone from being mailed again.
  • Merging on one side only. The other system keeps both records and splits the history.
  • Overwriting corrections. A value a person corrected must not be restored by the next sync.
  • Nobody reads the error queue. Assign it to a person, review it on a schedule and fix causes, not records.

Frequently asked questions

Should the CRM and the marketing platform sync in both directions?

The connection usually runs both ways, but each field should not. Treat direction as a property of the field rather than of the integration: the CRM owns the fields sales and the pipeline depend on, such as owner, stage and deal data, and the marketing platform owns the engagement fields it observes, such as email activity and the score it calculates. A field that both systems are allowed to write is a field whose value depends on which system saved last, and that is the most common cause of values flipping back and forth.

Should every contact in the CRM be synced to the marketing platform?

No. The marketing platform needs the people it may legitimately market to, plus whatever it needs to recognise and suppress the rest. Contacts with no consent record or a recorded objection, internal users, test records, and records kept in the CRM only for service or finance reasons rarely need a marketable copy there. Fewer synced records also means fewer duplicates to reconcile, and the data minimisation principle the GDPR sets out points in the same direction. Agree the rule with whoever owns data protection rather than deciding it inside the integration settings.

What is the first thing to check when the two systems disagree?

Whether the field has one owner. If both systems can write it, the disagreement is a design problem and the fix is to choose an owner and make the other side read-only. If it has one owner, look at the sync error queue for the record, then at whether the record exists twice on either side, then at whether a workflow changed the value after the sync ran. Record what you find in the change log, because the same cause usually affects many records, not one.

Terms used on this page

Frameworks

Where this page is used

Also related

Sources

  1. Direct marketing guidance — Information Commissioner's Office (UK)
  2. Regulation (EU) 2016/679 (GDPR) — EUR-Lex

Related