GUIDE

Connecting your CRM to your website: what to sync and what to leave alone

· 5 min read

Someone in sales asks why website enquiries take a day to show up in the CRM, and someone in marketing asks why the CRM has three records for the same person, and the answer to both is the integration that was set up with a Zapier account two years ago and hasn’t been looked at since. Connecting a website to a CRM is one of the first integrations most businesses do and one of the least thought about. It’s worth thinking about, because it decides whether the sales team trusts what comes out of the website, and a lot of the trouble comes from connecting too much rather than too little.

Quick answer: Sync form submissions into the CRM as contacts or leads, with the source data (landing page, campaign, click ID) attached, and route them to an owner. That’s the integration that pays. Be careful with two-way sync of contact records, don’t push marketing lists back and forth automatically, and don’t let the website create deals. Test the whole path with a real submission once a week, because these integrations fail silently and nobody notices until the pipeline looks thin.

What should flow from the website to the CRM?

Every enquiry, with enough context for a salesperson to act without asking. The form fields, obviously. Then the things the visitor didn’t type: which page they submitted from, where they arrived from (referrer, UTM parameters, the Google Ads click ID if there is one), and ideally which pages they looked at before enquiring. Most CRMs have fields for the first three; the browsing history needs the CRM’s own tracking script on the site, which HubSpot, Pipedrive, and the others all provide.

The form that matters most is the one that indicates buying intent: request a quote, book a call, ask about pricing. Newsletter signups and gated downloads are worth capturing too, but as contacts with a marketing status rather than as leads for sales to chase. Sending every download to the sales queue is how salespeople learn to ignore the queue.

Routing is part of the integration, not an afterthought. A lead that lands unassigned in a shared view gets picked up when someone remembers to look. Assign by territory, by product line, or round-robin, and set a notification that reaches a phone. The value of a fast response to a web enquiry is one of the better-documented findings in sales, and the integration is where the speed comes from.

What should you be careful with?

Two-way sync of contact details. It sounds tidy: update the phone number in the CRM and it updates in the email tool and the website account. In practice, two systems that can each change the same record produce conflicts, and the sync tool resolves them by whichever it saw last, which is how a corrected email address reverts to the typo a week later. If you need two-way, pick one system as the master for each field, and let the other only read.

Marketing lists. Pushing email-tool segments into the CRM and CRM lists back into the email tool, on a schedule, generates duplicates and consent problems. Someone who unsubscribed in the email tool gets re-added from the CRM. Keep subscription status in one place and have the other system respect it.

Deals. Letting the website create a deal or opportunity for every form submission fills the pipeline with records that have no value, no stage, and no owner, and it makes the pipeline report meaningless. Create contacts from the website; let a person create the deal once they’ve spoken to the contact.

Duplicates deserve their own line. Match on email before creating a contact, and decide what to do with the second submission from the same address: update the existing record and log a new activity, rather than creating another contact. Most native integrations do this; many Zapier-style setups don’t unless someone configured the lookup step.

Which integration method should you use?

Native, if your CRM has a plugin or embedded form for your platform and it does what you need. HubSpot’s forms and tracking, for instance, handle source capture and dedupe without any configuration, and the trade-off is that you’re using their form rather than yours.

An automation tool (Zapier, Make, and similar) when the native option doesn’t exist or doesn’t map the fields you need. These are quick to set up and they’re the ones that fail silently: a field renamed on the form, an expired connection, an exceeded task quota, and submissions stop arriving with no error anyone sees. If you go this way, build the failure notification at the same time and put a monthly check on the calendar.

A small custom integration when the rules get specific: lead scoring based on your own criteria, routing logic the automation tool can’t express, or a form that has to create records in the CRM and the accounting system and a project tool in one go. This is a few days of development and it’s the option that tends to be still working three years later, because it doesn’t depend on a third-party tool’s plan limits. The custom software vs off-the-shelf framework applies in miniature here.

How do you know whether it’s working?

Submit a test enquiry every week and follow it: did it appear in the CRM, with the right source fields, assigned to someone, within a few minutes? Put it on someone’s Friday list. Then reconcile monthly: the count of enquiry-form submissions on the website against the count of contacts created from the website in the CRM. They won’t match exactly, for the reasons in our post on why GA4 leads don’t match the CRM, but the gap should be explainable and stable.

And set an alert on silence. If the integration normally creates ten contacts a week and creates none for three days, someone should get an email. That one alert catches most of the failures that otherwise get noticed at the end of the quarter.

Businesses that have been keeping enquiries in a shared inbox or a spreadsheet before this point usually find that the integration is the moment they discover how many leads were falling through; the spreadsheet workflow post covers what tends to happen next.

If you’d like help choosing the method or building the custom version, this is a standard piece of AI and systems integration work for us, usually a week including the routing rules and the monitoring.

Frequently asked questions

Which CRM works best with a WordPress site?

Any of the common ones. HubSpot has the most complete free WordPress integration, and Pipedrive, Zoho, and Salesforce all have supported paths. Choose the CRM for how the sales team works; the website connection is solvable for all of them.

Should the website use the CRM’s own form?

If the CRM’s form does what you need and you can style it acceptably, it’s the lowest-maintenance option. If you need conditional logic, file uploads, or a specific design, keep your form and integrate, and accept a little more upkeep.

How do we capture the Google Ads click ID?

A small script on the landing page reads the gclid parameter from the URL, stores it in a cookie, and writes it into a hidden field on every form. The CRM then holds it on the contact, which is what lets you report closed deals back to Google Ads later.

What about phone enquiries?

They go into the CRM by hand, with a source. If phone volume is high, call tracking numbers per channel can automate the source. The point is that every lead, whatever the channel, ends up in the same system with the same fields.

Can we sync the CRM back to the website for logged-in customers?

Yes, carefully. Showing a customer their account manager or their contract details on the website is useful and read-only. Letting the website write back to the CRM is where the two-way problems start; keep writes to a defined set of fields with the CRM as master for everything else.

How often do these integrations break?

Often enough that the weekly test matters. Common causes are a form field being renamed, a CRM API key expiring, a plugin update changing the submission event, and an automation tool’s quota being hit. None produces a visible error on the website.