Should Marketing Automation Live Inside Your CRM or Beside It?

If Salesforce is your CRM and will stay your CRM, marketing automation that runs inside it has fewer moving parts to maintain than a platform that sits beside it and syncs. The connector explains most of the gap: how often it syncs, which records it may touch, whose API allowance it spends, and what it does when two records share an email address. Both vendors document those rules, so the comparison can happen on paper before anyone signs.

To keep this concrete, the piece follows a hypothetical team: 25 sales reps working in Salesforce Sales Cloud, two marketers, and one Salesforce admin who also looks after quoting and reports. They are choosing between Account Engagement (the product formerly called Pardot), which runs inside Salesforce, and HubSpot Marketing Hub, which would sit beside it. Which CRM to run is its own decision, covered in our HubSpot vs Salesforce comparison, and for this team the CRM is fixed.

What decides built-in CRM automation versus a separate platform?

The integration trade-offs between marketing automation built into the CRM and a separate platform come down to four behaviors you can look up: sync timing, sync scope, API consumption and duplicate handling. Features still matter, and the wider category is mapped in our guide to what marketing automation is. Features get compared in demos, though, while connector behavior surfaces months later as a lead that reached a rep late, a contact that crossed into Salesforce when it shouldn't have, or a field that keeps overwriting itself.

Native automation and a connected platform, side by side

Native CRM marketing automation and a third-party integration each have pros and cons, and the 2026 documentation states most of them plainly. The table compares the two connectors on the points that create admin work, with the sources for each row in the sections that follow.

What HubSpot documents about syncing contacts and leads

The official HubSpot Marketing Hub and Salesforce integration documentation for 2026 covers how contacts and leads sync, and three rules shape the daily work. The first is scope. You pick an active segment as the inclusion segment, and HubSpot says records outside it won't sync between the two systems. With no segment selected, all HubSpot contacts sync to Salesforce. Plan for one exception: if your settings create HubSpot contacts from new Salesforce records, a newly created Salesforce lead or contact creates or matches a HubSpot contact even outside the segment, and later updates wait until the record meets the segment criteria.

The second is record type. In its integration settings, HubSpot lets you choose whether a new or updated HubSpot contact creates or updates a contact or a lead in Salesforce, and whether a Salesforce update creates or updates a HubSpot contact or does nothing. The same settings page is where you allocate a number of Salesforce API calls to the integration, and its Sync Health tab shows calls used over the last 24 hours next to the sync errors that are blocking records.

The third is timing and cost. HubSpot checks both systems for new or updated information approximately every 15 minutes, and its sync triggers article advises having the segment's record count times three in available Salesforce API calls before a segment resync. Day to day, HubSpot states that the integration has a daily API call limit and that a single contact sync can take up to four API calls per record, though the number varies.

Duplicates follow email. When a HubSpot contact is first created and several Salesforce leads or contacts share its address, HubSpot syncs with the first record Salesforce returns, and Salesforce returns them in no particular order. After that it follows the most recently updated match. HubSpot warns this can leave HubSpot contacts overwritten again and again, and its fix is to merge the Salesforce duplicates while keeping the currently syncing record as the primary.

What Salesforce documents for Account Engagement

Account Engagement's connector does the same job with different defaults. The Salesforce Connector v2 implementation guide, last updated July 22, 2026, says syncs can occur as often as every 2 minutes, with larger volumes taking longer. The connector is created paused, and when the integration user is the connector user it syncs all records for the objects shared between the two systems. To sync only a subset, the guides offer two routes: a connector user with narrower Salesforce permissions, or Marketing Data Sharing, which the Salesforce setup guide ties to the Advanced and Premium editions. Marketing Data Sharing is the nearest equivalent to HubSpot's inclusion segment, and it sits on a higher tier.

Matching runs on CRM ID. New business units allow multiple prospects with the same email address by default, and in those units the connector matches on CRM ID in both directions; when it detects merged or deleted records in Salesforce, it updates the matching prospects and moves deleted ones to the recycle bin. Email still matters at the edges. A prospect without a CRM ID syncs to the first Salesforce record it finds with that address, and importing by email in a shared-email business unit creates duplicate leads or contacts in Salesforce, so Salesforce says imports there must use CRM ID. The connector also won't create prospects from existing Salesforce leads and contacts by itself, which makes the first sync an import.

Two smaller points help the native side. Changes made by the connector are attributed to an integration user that doesn't use a standard Salesforce user license, so the audit trail separates sync edits from human ones. The setup guide also recommends connected campaigns, which link Account Engagement and Salesforce campaigns for features such as campaign influence attribution. Whichever connector you choose, have your admin compare the org's Salesforce API usage before and after go-live.

Who carries the integration work in best-of-breed and all-in-one stacks?

The best-of-breed versus all-in-one question is mostly about who carries the integration burden, and that burden grows with every tool that writes to the CRM. The 2026 Connectivity Benchmark Report from MuleSoft, a survey of 1,050 IT leaders published in February 2026, found the average organization manages 957 applications and only 27% of them are connected. IT teams in the same survey spent an average of 36% of their time designing, building and testing new custom integrations. That figure covers new builds across all of IT, and marketing connector upkeep sits on top of it.

The same survey found 86% of IT leaders agree that without proper integration, AI agents can add more complexity than value. If you plan to let AI agents work across marketing and sales records, one platform gives them one permission model and one data shape, while a connected stack gives them two of each plus a sync delay. Whichever you choose, decide who approves agent writes before they start; our governance operating model sets that out.

An upkeep worksheet for each connected tool

Total cost of ownership across multiple SaaS tools includes integration maintenance that never appears on a quote, so count it in hours, per connected tool, every month. Each row below ties to a behavior documented above. The hours are assumptions for the hypothetical team; replace them with your own time logs.

At an assumed loaded admin cost of $85 an hour, 11 hours a month comes to 132 hours and $11,220 a year for the connected option, against 84 hours and $7,140 for the native one. The $4,080 gap matters for a team this size, though it belongs next to your two license quotes before anyone treats it as the deciding number. The worksheet earns its place as the stack grows, because every extra tool that writes to the CRM gets its own column and the totals add up across all of them. If you are also weighing automation or agent spend, the AI ROI calculator is a place to set those hours against expected savings.

When is HubSpot Marketing Hub with Salesforce worth it?

HubSpot Marketing Hub with a Salesforce CRM is worth the integration when your marketers would ship noticeably more in HubSpot's tools, can own their changes without queueing behind a Salesforce admin, and can live with a sync check roughly every 15 minutes. It also helps when you want rule-based control over which contacts reach Salesforce without buying a higher automation tier, since HubSpot's inclusion segment comes with its Professional and Enterprise plans while Account Engagement's Marketing Data Sharing needs Advanced or Premium. Check two numbers first: your Salesforce org's spare API capacity at up to four calls per record, and how many Salesforce leads and contacts share an email address, because that second count predicts your overwrite problem.

The native option wins when reps live in Salesforce all day, the admin is already the bottleneck, and lead routing depends on marketing data reaching a record within minutes. It also wins when your database has many shared emails, since CRM ID matching avoids most of the collisions HubSpot warns about. Handoffs between marketing and sales gain the most from a short sync, and our lifecycle email playbook covers how to design them.

For the hypothetical team, the native option comes out ahead: one admin, 25 reps who rarely leave Salesforce, and a worksheet showing four extra hours a month that the admin doesn't have. A team with a dedicated marketing operations lead, an email-unique database and marketers who already know HubSpot could run the same sheet and land the other way. For help mapping your own connectors, see our agentic AI and automation service.

Sources

• Salesforce, Salesforce Connector v2 Implementation Guide
• Salesforce, Account Engagement Setup Implementation Guide
• HubSpot, Sync records between HubSpot and Salesforce
• HubSpot, Manage your Salesforce integration settings
• HubSpot, Import and sync Salesforce records
• HubSpot, Connect HubSpot and Salesforce
• HubSpot, Resolve duplicate Salesforce records syncing with HubSpot
• MuleSoft, Agentic Transformation Reaches New Heights: 2026 Connectivity Benchmark Report Insights