This guide is for marketing operations leads, digital leads and CTOs who run a company website on a CMS and want leads, consent and campaign data to reach the right tools without manual exports. It sits under Netbase's web application development services, and the API work behind it is part of systems and API integration.
In this guide
- Which system should own each record?
- What breaks when nobody owns the data?
- Integration patterns compared
- Build sequence: eight steps from audit to monitoring
- Where AI fits
- What delivery record exists, and what does not
- Alternatives: when not to build a custom integration
- How do you know the wiring works?
- Limits of this guide
- Frequently asked questions
- Next step
Which system should own each record?
Most website integrations fail on ownership, not on code. When two tools can both edit a contact's lifecycle stage or consent, the last sync wins and nobody can explain the result. The fix is a written matrix that names one system of record per record type, the only writer allowed, and the direction data moves.
-
Pages, articles and form definitions
- System of record
- CMS
- Who may write it
- Editors and developers
- Sync direction
- CMS to site only
-
Form submission (raw payload)
- System of record
- Integration layer log
- Who may write it
- The website, append-only
- Sync direction
- Website to log, then to CRM
-
Contact and company
- System of record
- CRM
- Who may write it
- Sales, the integration layer by upsert
- Sync direction
- Website to CRM; CRM to marketing automation
-
Consent and preferences
- System of record
- Consent store or CRM consent fields
- Who may write it
- The visitor through the banner or preference centre
- Sync direction
- One direction out to every tool
-
Lifecycle stage and lead status
- System of record
- CRM
- Who may write it
- Sales rules and marketing automation scoring
- Sync direction
- CRM to marketing automation and reporting
-
Campaign membership and email activity
- System of record
- Marketing automation
- Who may write it
- The marketing automation tool
- Sync direction
- Marketing automation to CRM
-
Page views, clicks and conversions
- System of record
- Analytics
- Who may write it
- Tags and server-side events
- Sync direction
- Website to analytics; summary back to CRM
-
Campaign source (UTM values)
- System of record
- Captured at first touch, stored on the contact
- Who may write it
- The website at submission time
- Sync direction
- Website to CRM, never overwritten by later visits
Three rules make the matrix work. The CMS never stores personal data beyond what it needs to render a form. The CRM is the only place where a person's commercial state lives. Consent travels out from one store and is never re-entered by hand in a second tool.
What breaks when nobody owns the data?
The same failures appear on most company websites that grew their integrations one plugin at a time. Each one maps back to a missing row in the matrix above.
- Duplicate contacts. Two forms create two records because one matches on email and the other does not match at all.
- Lost campaign source. UTM values sit in the landing-page address but are not copied into a hidden field, so the CRM records every lead as direct traffic. Google's own guidance asks for one standardised UTM naming scheme across all campaigns.
- Consent drift. A visitor withdraws consent in the preference centre, but the marketing automation tool still holds the old flag because consent was copied once at signup.
- Silent failures. A CRM API change or an expired token stops form delivery, the website still shows a thank-you message, and leads are lost for days.
- Overwritten sales data. A nightly sync writes a website value over a field a salesperson corrected by hand.
Article 7 of the GDPR requires that a controller can demonstrate consent and that withdrawing consent is as easy as giving it; that is hard to show when consent lives in four tools with no single record.
Integration patterns compared
Pick the pattern per data flow, not per project. A site can send forms through a small integration layer and still use a native connector for email activity.
| Pattern | Strength | Risk | Choose it when |
|---|---|---|---|
| Native connector or plugin | Fast to switch on; maintained by the vendor | Limited field mapping and error handling; hard to audit | One CMS, one CRM, standard forms and low change frequency |
| Form-to-API through a small integration layer | Validation, deduplication, retries and a payload log in your control | One more component to host and monitor | Leads matter commercially, forms are custom or several tools receive the same submission |
| Tag manager plus server-side events | Analytics and ad events survive browser limits; consent applied in one place | Event naming and consent state must be designed up front | Measurement accuracy matters and conversions happen after the visit (calls, signed contracts) |
| Middleware or iPaaS | Many systems, shared mappings, monitoring dashboards | Licence and governance overhead; mappings spread across a separate tool | Three or more systems exchange data and an operations team will own the platform |
Server-side events have limits worth knowing. Google states that its GA4 Measurement Protocol is meant to augment tag-based collection, not replace it, so offline conversions sent from the CRM complement the website tags rather than standing alone.
Build sequence: eight steps from audit to monitoring
-
Inventory every form and tag
List each form, the fields it collects, where submissions go today and which tags fire on which pages.
-
Write the ownership matrix
One system of record, one writer and one sync direction per record type, approved by marketing and sales owners together.
-
Define the contact match key
Usually a normalised email address, with rules for personal versus company domains and for updates to an existing contact.
-
Design consent first
Set default consent states before any tag fires and update them from the banner, as Google consent mode describes, and send the same state to the CRM and marketing automation.
-
Capture campaign source at first touch
Store UTM values and the landing page in hidden fields or the session, and write them to the contact only when the fields are empty.
-
Route submissions through one layer
Validate, log the raw payload, upsert the contact, then notify marketing automation; use idempotency keys so a retry never creates a second record.
-
Name events once
Agree a short event list (form submitted, demo requested, document downloaded) shared by analytics and marketing automation.
-
Monitor and reconcile
Alert on failed deliveries and token expiry, and compare weekly form counts in the log with new contacts in the CRM.
Where AI fits
AI helps around the integration, not inside the consent rule. An assistant can draft a field mapping from the form inventory and the CRM field list, suggest test cases for duplicates and odd inputs, and summarise a reconciliation report in plain language. Netbase works with the major commercial and open-source AI tools and models, chosen per project.
Engineers approve every mapping, match rule and consent setting; no model writes contacts or consent to a production CRM on its own.
- In delivered work
WhatsApp AI chatbot with CRM integration
An AI chatbot for a client that is not named, synchronised with its CRM for lead capture, customer data and workflow automation; a conversational channel, not a CMS integration.
- Available capability
AI and data services
AI-drafted field mappings and plain-language reconciliation summaries are a way Netbase approaches the work, not yet tied to a published case.
What delivery record exists, and what does not
Netbase deploys a help desk and B2B CRM for business partners with lead capture, scoring and nurturing, workflow automation, and support and sales performance analytics. Its ERP and CRM work includes API sync of product catalogues, inventory, orders and customer data between e-commerce stores and ERP systems, and a Vtiger CRM to ERP upgrade covering customer experience management, sales orders and support tickets. The commerce platforms it has delivered include WooCommerce, Magento 2, Laravel and headless commerce.
For a client that is not named, Netbase built an AI chatbot synchronised with the client's CRM for lead capture and workflow automation. Its named AI delivery is the 4over4 recommendation engine, which suggests products from browsing history and previous purchases, and it has delivered other AI projects, including RAG knowledge assistants, for clients that are not named.
Netbase plans security and API design from the start and delivers in a six-step lifecycle from discovery and architecture planning to rollout and ongoing support. Most Netbase projects are delivered on fixed-price contracts. No published Netbase record describes a CMS-to-marketing-automation project end to end, and no lead volume, conversion, cost or time figure is claimed here.
Alternatives: when not to build a custom integration
A custom layer is not always the right answer. Use these criteria before you commission one.
- An all-in-one suite. If the website is small, marketing and sales sit in one team and nobody needs a separate CMS, a suite where the CMS, CRM and email share one database removes the integration entirely. The cost is lock-in and weaker content modelling.
- Native connectors only. If you have one CMS, one CRM, standard forms and fewer leads than a person can check each week, a vendor connector plus a weekly export check is enough.
- A scheduled export. If leads are rare and reviewed by hand, a daily CSV export with a named owner is cheaper and easier to audit than any live sync.
- Warehouse first. If several products, stores and apps already feed a data warehouse, send website events there and push audiences to the CRM from the warehouse, with the guardrails in reverse ETL to the CRM.
Back-office records such as orders, invoices and stock follow different rules; ERP, CRM and e-commerce integration architecture covers them. Once leads reach the CRM, CRM workflow automation handles routing and follow-up.
How do you know the wiring works?
Run five checks before launch and repeat them after every CMS, plugin or CRM update. Submit each form with a new email, an existing email and a malformed value, and confirm one contact per person. Open a tagged campaign link, convert on a later page, and confirm the UTM values reach the contact. Withdraw consent and confirm every tool shows the change. Revoke the API token in a test environment and confirm an alert fires. Compare a week of logged submissions with new CRM contacts and explain every gap.
Limits of this guide
- It covers company websites and lead flows, not order, payment or stock data.
- It names patterns, not vendors; check each tool's API limits and consent features in its own documentation.
- It is not legal advice; confirm consent wording and retention rules with your privacy counsel.
Plan the next step with a Netbase consultant
Frequently asked questions
The CRM should own contacts and their commercial state. The CMS defines pages and forms and passes each submission on, keeping no more personal data than it needs to render the form and confirm delivery.
It is often enough for one CMS, one CRM and standard forms with low volume. Move to a small integration layer when leads matter commercially, forms are custom, several tools need the same submission or you need retries and an audit log.
Capture the UTM values and landing page on the first visit, keep them in the session or hidden form fields, and write them to the contact only when its source fields are still empty, so later visits never overwrite the first touch.
Next step
Netbase can review your forms, tags and CRM fields and propose an ownership matrix, an integration pattern per flow and a monitoring plan: book a solution review.
Related services and solutions
Web application development with AI built in where it converts Netbase provides web application development for product and commerce teams that need fast, maintainable web apps and storefront front ends, with AI search, assistants and automation added where they measurably help. Since 2012 Netbase has worked with 500+ businesses across 80 countries; according to its published case studies, PrintLeo's page load improved 35% and 4over4's conversion rose 48%.
Learn More
API integration services that keep your systems in agreement and ready for AI Netbase provides API integration services for commerce and operations teams that need e-commerce, ERP, CRM, shipping and payment systems to agree, so data moves once and AI services can act on it, with retries, monitoring and reconciliation on every API and webhook. According to its published client record, Netbase has worked with 500+ businesses across 80 countries since 2012.
Learn More
CRM workflow automation with AI lead scoring, from first lead to paid invoice CRM workflow automation connects your sales pipeline to the work behind it, so leads, proposals, approvals and invoices move on rules instead of spreadsheets and email threads, with AI scoring leads and summarising conversations. Netbase builds it for sales and operations teams in growing companies from its CRM and B2B sales engine and workflow automation toolkit, behind Cloodo Workspace.
Learn More
Work with Netbase remotely
Every Netbase engagement is delivered remotely, so clients in any country and any industry can work with us; Netbase has worked with clients in 80 countries. Contracts and invoices are issued by Netbase JSC, and payment is made by bank transfer to Netbase JSC's company account or PayPal Business. Talk to the CEO, Huy Nguyen (David), on WhatsApp or by email.
Discuss a project
Netbase JSC helps organizations design, build, modernize, and operate digital products and AI-enabled business systems.project@netbasejsc.com
+84 937 869 689
91 Nguyen Chi Thanh, Dong Da, Hanoi, Vietnam
Get in touch
Tell us what you want to build, modernize, or operate.