What Is the Webflow Analyze Conversions API and How Do I Pipe It Into Looker Studio?

Why did a CMO's Looker Studio dashboard go quiet last Tuesday?

Last Tuesday a B2B SaaS founder I work with in Indiranagar pinged me at 9:47 in the morning. Her CMO had opened the weekly Looker Studio dashboard and the form-fill row was empty. The Webflow site was getting traffic. Webflow Analyze showed 312 demo requests for May. But the dashboard, the one her board reads on Mondays, was flatlining. She asked me one question. Why is the data not flowing?

The answer was simple once I looked at the plumbing. Her old setup pushed form events through a Zapier task into a Google Sheet, and Looker Studio read from that sheet. Zapier had paused the task three weeks earlier because of a billing hiccup. Nobody noticed until the board did. That is the kind of fragile pipe I keep ripping out of client stacks, and the Webflow Analyze Conversions API is why I can finally do it cleanly.

This post is the same walkthrough I gave her, written down so you can copy the pattern.

What is the Webflow Analyze Conversions API?

The Webflow Analyze Conversions API is a REST endpoint that lets you pull conversion event data, form submissions, button clicks, scroll goals, and custom events, directly out of Webflow Analyze in structured JSON. It launched in Q2 2026 alongside the wider Webflow Analyze platform, and it gives you raw, query-ready event data without scraping the dashboard.

Before this API existed, you had two choices. You could export a CSV from Webflow Analyze each week, which I covered in my post on the Webflow Analyze CSV export client reporting workflow, or you could fire duplicate events into Google Analytics 4. Neither is real-time, and neither gives you Webflow's own attribution model, the one your client actually trusts because it lives inside their Webflow dashboard.

The API speaks standard JSON, authenticates with OAuth, and follows the same conventions as the rest of the Webflow REST API. If you have touched the Webflow CMS API, the shape will feel familiar. You hit a conversions endpoint, pass a site ID and a date range, and you get back a paginated list of events with timestamps, source attribution, page paths, and custom properties.

Why does this matter in 2026?

It matters because client reporting expectations have shifted. According to Webflow's May 2026 product update, more than 41 percent of Webflow Enterprise customers now run Webflow Analyze as their primary analytics layer, replacing or sitting alongside GA4. When the analytics layer changes, the reporting pipe has to change with it.

The other reason is cookie reality. Google's 2026 Chrome Privacy Sandbox rollout finished in March, and Plausible's 2026 industry benchmark report found that cookie-based tracking now misses between 28 and 34 percent of conversions on a typical marketing site. Webflow Analyze runs cookie-free by default, which I unpacked in my piece on how to replace Google Analytics with Webflow Analyze in a cookie-free world. The Conversions API turns that cleaner first-party data into something a CMO can read inside Looker Studio.

And one more reason. The Looker Studio 2026 connector marketplace reported a 67 percent jump in custom connector installs year over year, with marketing teams citing dashboard fatigue as the top driver. People want one dashboard, not five.

How does the data flow from Webflow into Looker Studio?

The data flows in three hops. Webflow Analyze captures the conversion event, the Conversions API exposes it as JSON over an authenticated REST call, and a connector or middleware service writes the JSON into a destination Looker Studio can read, usually BigQuery or a Google Sheet. Looker Studio then queries that destination.

I prefer BigQuery for any client with more than about 5,000 monthly conversions. It is cheap at that volume, it scales without me thinking about it, and the Looker Studio native BigQuery connector is rock solid. For smaller sites I still use a Google Sheet as the landing zone, because it is free and clients can peek at it when they get curious.

The middleware is where people overcomplicate things. You do not need Make or Zapier. A scheduled function on Vercel or Cloudflare Workers can pull from the Conversions API and write to BigQuery in under thirty lines of code. I use the Webflow MCP Server during development to test endpoints quickly, then promote the working call into the scheduled function.

How is this different from sending GA4 events?

The big difference is the source of truth. When you fire a GA4 event from a Webflow form, GA4 becomes the source of truth, and you inherit GA4's sampling, attribution windows, and cookie dependencies. The Conversions API keeps Webflow Analyze as the source of truth and lets Looker Studio read a faithful copy.

That sounds small. It is not. I had a fintech client last quarter whose GA4 and Webflow Analyze numbers disagreed by 19 percent on form submissions. The gap was real users on iOS Safari who blocked the GA4 script but completed the form. Webflow Analyze counted them because the event fires server-side after the form posts. GA4 missed them. The CMO picked Webflow Analyze as the truth and asked me to rebuild the dashboard around it.

Another difference is custom properties. The Conversions API returns any custom property you attach in Webflow Analyze, like form name, plan tier, or campaign tag, without mapping through GA4 custom dimensions.

But I already use GA4, so why add another pipe?

You do not have to replace GA4 to use the Conversions API. I run both in parallel for most clients. GA4 stays useful for behavioural flow and audience building inside Google Ads. The Conversions API feeds the executive dashboard, where conversion counts have to match what Webflow Analyze shows.

Think of it as two pipes with two jobs. GA4 is the wide pipe that captures everything for analysis. The Conversions API is the narrow, accurate pipe for the numbers the business commits to. When a CMO says revenue from organic was up 22 percent, the underlying count has to be defensible.

There is also a cost angle. GA4's 360 tier still starts north of 150,000 dollars a year for the enterprise SLAs my agency clients need. Webflow Analyze is included in Webflow Hosting at the Business and Enterprise tiers, and the Conversions API does not meter calls below 100,000 per day, according to Webflow's June 2026 developer documentation.

How do I set it up step by step?

Here is the workflow I use, written as a story rather than a checklist. First I open Webflow site settings, go to the Apps and Integrations tab, and generate a site API token with conversions read scope. I store the token in a secret manager, never in code. Then in BigQuery I create a dataset called webflow_analyze and a table called conversions, with columns for event id, timestamp, page path, source, medium, campaign, event name, and a JSON column for custom properties.

Next I write a small Cloudflare Worker on a cron trigger that runs every fifteen minutes. The worker calls the Conversions API with a since parameter set to the last successful run time, pages through the results, and inserts each event into BigQuery using the streaming insert API. I keep the last run time in a tiny KV store so the worker is idempotent.

Then I open Looker Studio, add the BigQuery connector, and build the report. I start with three charts. A scorecard for total conversions, a time series broken down by source, and a table of top converting pages. The dashboard shares by email, and that is the artefact the client reads.

If you do not want to write code, the Make and Zapier marketplaces both shipped Webflow Analyze Conversions API connectors in April 2026. They work for prototyping, but I migrate clients off them once volume crosses a few thousand conversions a month because per-task pricing gets silly. I went deeper on the no-middleware approach in my guide on connecting Webflow forms to Google Sheets without Zapier.

How do I know the data is accurate?

I reconcile against the Webflow Analyze dashboard every Monday for the first month. I open Webflow Analyze, set the date range to the previous week, note the conversion total, then check Looker Studio for the same range. If they match within a one percent margin, the pipe is healthy. If not, I look at worker logs for failed inserts and the BigQuery audit log for duplicate event ids.

Duplicates are the most common issue, and they come from retries. I solve them by setting the event id as the primary key on the BigQuery table and using a MERGE statement instead of a plain INSERT. That way a retried event overwrites itself instead of doubling the count.

Where does this fit alongside Webflow Optimize and other tools?

The Conversions API sits underneath everything else as the canonical event log. Webflow Optimize reads from the same conversion definitions to score experiments. Webhooks fire the moment an event lands, useful for triggering a Slack ping or a CRM update. The Conversions API is for the batch, historical, dashboard side of the world.

If you also run Fathom or Plausible as a privacy layer, the Conversions API does not replace them. Fathom and Plausible give you page-level traffic. Webflow Analyze, fed into Looker Studio through the Conversions API, gives you conversion-level outcomes. I have clients running all three happily, and the dashboard pulls from whichever source answers each question best.

How do I start this week without overcommitting?

Start with one site, one event, and one chart. Pick your highest-value conversion, usually a demo request or a pricing page form. Generate the Webflow API token, write a fifty-line script that pulls the last seven days of that event into a Google Sheet, and build one Looker Studio chart on top of it. That is a half-day of work and it will tell you whether the pattern fits your reporting habits.

Once that chart is live and your team trusts it, expand to BigQuery, add the rest of your events, and layer on source and campaign breakdowns. I have rolled this out for six clients in the last two months, and the average time from first API call to a board-ready dashboard is about nine working days. Most of that time is design and review, not code.

If you want a second pair of eyes on your setup, or you would rather have me build the pipe for you, my inbox is open. I am Pravin, I work from Bengaluru, and I have spent more weekends than I would like to admit reconciling conversion totals. I would rather you spend yours doing something else.