Quick answer: A goal funnel in Google Analytics maps the ordered pages or screens you expect a user to hit before a conversion. In Universal Analytics you define a goal and optional funnel steps, then use the Funnel Visualisation report to see where people abandon. Fix the step with the largest drop first. Funnels do not replace thinking about intent; they show friction in a path you already believe matters.
Goals first, funnels second
A goal is the completion event: thank-you page, form confirmation URL or (with extra setup) an event. A funnel is optional scaffolding that describes the path toward that goal. Without a solid goal, funnel charts are theatre.
If you are still defining what “converted” means, start with our Google Analytics goals walkthrough, then return here to interpret steps.
This article assumes Universal Analytics property settings as used in April 2020. Keep tagging disciplined; messy containers make funnel steps look abandoned when the real issue is a missing pageview. Google Tag Manager best practices help keep names consistent.
When a funnel helps
Good use cases:
- Multi-page checkout
- Multi-step lead forms (each step a URL)
- Demo request paths with a pricing middle page
- Event registration with a review screen
Weak use cases:
- Single-page apps without virtual pageviews
- Paths where users legally skip steps (and you still force the funnel to require them)
- Goals completed mostly from email links that land mid-funnel when you required step 1
Building a usable funnel in Universal Analytics
- Admin > View > Goals > New goal
- Choose a template or custom
- Type: Destination (common for thank-you URLs)
- Enter the destination path
- Turn Funnel on
- Add steps in order with names humans understand (“Cart”, “Shipping”, “Payment”)
- Mark early steps as required only when users truly must pass them for a valid path
Use request URI paths carefully. Trailing slashes and query strings can split steps. Prefer clean thank-you URLs without leftover cart parameters when you can.
Verify with the goal validation feature where available, and complete a real test transaction or lead in a session that is not filtered out of the view.
Reading Funnel Visualisation without fooling yourself
The chart shows entrances to each step, exits and the final conversion count. Large exits at a step mean friction or intentional leave. Before redesigning the page:
- Check that the step URL is correct and firing
- Segment mobile vs desktop
- Compare a short date range to a longer one
- Exclude internal traffic so staff testing does not invent drop-offs
People entering mid-funnel are normal for some businesses (bookmarking checkout, for example). If Required Step is on for step 1, those users may not appear as completions even when they pay. That is a configuration choice, not a moral failing of the customer.
Pair funnels with behaviour reports
Funnels show path abandonment. They do not explain why. Open the step page in Behaviour reports, watch exit rate and look at on-page events if you track them. Event tracking on CTA clicks and field errors often explains a silent drop better than another meeting about button colour.
For ecommerce, Enhanced Ecommerce reports can give richer checkout step analysis than a basic destination funnel. If you already invested there, do not maintain two conflicting stories. See enhanced ecommerce for the richer model.
Common funnel design mistakes
- Thank-you URL reused for multiple goals (newsletter and purchase share
/thanks/) - Steps that are modals without pageviews
- Including the homepage as step 1 “because everyone sees it”
- Changing URLs in a redesign without updating the goal
- Filtering out the thank-you page accidentally in the view
Use unique thank-you URLs per major conversion type.
Turning drop-offs into a backlog
Pick the biggest absolute loss step, not only the biggest percentage if volume is tiny. Create tickets such as:
- Shipping cost surprise at step 3: show estimates earlier
- Form field errors on mobile: enlarge targets, reduce fields
- Login wall before checkout: add guest checkout
Then re-check the funnel two weeks after release. Annotate the date in Analytics so charts make sense later.
Views, filters and hygiene
Run funnels in a reporting view that:
- Filters internal IPs when possible
- Uses consistent hostname casing
- Does not append query parameters that fragment paths unnecessarily
Keep a raw unfiltered view for debugging. Google Analytics views hygiene prevents months of bad funnel history.
What funnels cannot tell you
- Brand lift from people who never convert online
- Call conversions without call tracking integration
- Offline closes after a form
- Soft intent (“researching for next year”)
Use CRM and call tracking when phone still closes deals. Otherwise marketing will over-react to on-site funnel exits that were never the whole business.
A monthly funnel review ritual
- Export or screenshot Funnel Visualisation for the key goal
- Note top exit step and device split
- Confirm tagging still matches production URLs
- Ship one friction fix
- Log the change
Small steady fixes beat annual redesign guesses.
Naming steps so humans understand reports
Call steps “Shipping address” not “step2.php”. When a new marketer opens Funnel Visualisation six months later, labels should still make sense. Rename steps when the site redesigns, and note the date in your analytics diary.
SPA and virtual pageviews
If your checkout is a single-page application, classic destination funnels need virtual pageviews or you will see one step forever. Coordinate with developers on a simple naming scheme (/checkout/shipping, /checkout/payment) fired when views change. Without that, stop forcing a funnel and use event-based analysis instead.
Sharing funnel findings with non-analysts
Translate chart drops into tickets with owners. “42% exit on shipping” becomes “show shipping estimate on product page” with a product owner and a due date. Screenshots of Funnel Visualisation help, but a sentence about the user problem helps more in standups.
Revisit the same funnel after each major UX release. If nobody owns the follow-up, the report becomes wallpaper.
Aligning CRM stages with funnel steps
When sales marks “qualified” in the CRM but Analytics only sees the thank-you page, debates start. Map each funnel step to a CRM stage or ignore CRM for path analysis. Dual systems need a dictionary, not rival screenshots in meetings.
FAQ
Destination goals vs event goals: which supports funnels?
Classic Funnel Visualisation is built around destination goals with optional steps. Event goals are useful but need different analysis patterns.
Should every step be required?
Only when the business truly requires that page. Over-requiring hides valid conversions.
Why do completions look lower than CRM leads?
Duplicate filters, missing thank-you hits, cross-device paths or offline leads. Reconcile definitions before blaming the website.
Can we funnel Instant checkout from ads?
If users land on payment directly, design a funnel that matches reality or track that path separately.
Find the leak, then patch it
Goal funnels in Google Analytics show where planned paths break. Configure steps that match real URLs, read exits with tagging humility and fix one high-loss step at a time. If you want help aligning analytics with the paths that create customers, see Know what’s working.
