Funnels in 5 Minutes: Where Your Visitors Drop Off (5 Real Examples)

A funnel is a sequence of pages or events you expect a visitor to pass through. The dashboard shows you how many made it to each step — and, more importantly, where the rest went. This is the single fastest way to find what's actually broken in your conversion path.

What a funnel actually shows

You define a sequence — typically 3 to 5 steps — using URL paths (and/or events). Logly counts, for each step, how many distinct visitors reached it. The output is a stack: 100% at the top, smaller bars below, with the drop-off between each pair labelled.

The number that matters is not the final conversion. It's the largest single drop. That step is your bottleneck, and fixing it usually moves the bottom number more than any new feature would.

How to set one up in Logly

In the dashboard, open the Funnels panel, click + New funnel, give it a name, and add steps. Each step is either a URL path (/pricing, /signup, /welcome) or a custom event name (signup_completed). Save. The chart populates from your existing traffic — no historical wait, no extra script.

Paths support exact matches. Avoid query strings in the steps; visitors arriving with UTM tags would not match otherwise.

Five funnels worth copying

1. The SaaS signup funnel

The classic. If you have one funnel set up, make it this one.

Step 1: /pricing
Step 2: /signup
Step 3: signup_completed   (event)
Step 4: /onboarding
Step 5: /dashboard

What to look for. The 2→3 drop (form opened vs account created) is the form's fault — too many fields, broken validation, password rules, or a slow API. The 4→5 drop (onboarding entered vs dashboard reached) is your onboarding's fault — usually one step too many.

A healthy ratio in 2026 for a self-serve B2B tool: pricing→signup around 8–15%, signup→completed around 50–70%, onboarding→dashboard above 80%. Numbers below those are conversion problems, not traffic problems.

2. The e-commerce checkout funnel

Step 1: /product/*       (any product page — use a representative path)
Step 2: /cart
Step 3: /checkout
Step 4: /thank-you

What to look for. Cart→checkout is the abandonment step everyone knows about (shipping cost surprise, account-required walls, payment trust). Checkout→thank-you is the payment-failure step — if you see more than 15% drop here and you offer only one payment provider, that's your weekend project.

3. The lead-magnet funnel

If you give away an ebook or a template in exchange for an email, this funnel tells you whether the content brief is right.

Step 1: /blog/article-slug
Step 2: /resources/lead-magnet
Step 3: newsletter_subscribed   (event)
Step 4: /thanks

What to look for. A high blog→landing rate combined with a poor landing→subscribed rate means the article promised more than the lead magnet delivers. A low blog→landing rate means the in-article CTA needs work (placement, copy, or a friendlier alternative further down).

4. The docs / activation funnel

For developer tools, "did they ever fire a real event?" is a better activation signal than "did they sign up?".

Step 1: /docs
Step 2: /docs/install
Step 3: /docs/first-event
Step 4: first_event_received   (event, fired server-side by your API)

What to look for. Install→first-event drop is your getting-started friction. If half the people who read install never reach first-event, the install snippet is missing something or the first-event example is too abstract.

5. The multi-step form funnel

Long forms (mortgage applications, quote requests, account onboarding) live or die on each step. Use events at each step, not URLs — most multi-step forms keep the same URL.

Step 1: form_step_1_viewed   (event)
Step 2: form_step_2_viewed
Step 3: form_step_3_viewed
Step 4: form_submitted

What to look for. The first big drop tells you which question is killing the form. Common offenders: phone number on step 2 (too soon), salary on step 3, file upload on any step.

How to read a funnel honestly

The bottom number is a vanity metric. It compresses three or four problems into one. Look at the step deltas — the percentage you keep between each pair of steps. That's where the actionable signal lives.

Compare against your own past, not benchmarks. Industry conversion benchmarks vary by 10× between sources. Your own funnel a month ago, before you changed the form, is a much better baseline.

Don't optimise the step that already converts at 90%. The marginal gain there is tiny. Spend the same effort on the step that converts at 25%.

Three pitfalls to avoid

Putting too many steps in. A 9-step funnel always looks broken — every step compounds the drop. Three to five steps is the sweet spot.

Mixing required and optional paths. If /welcome is sometimes skipped depending on user type, including it in the funnel will exaggerate drop-off for the cohort that legitimately skips it. Either build two funnels or remove the optional step.

Forgetting that not every visitor enters at step 1. Funnels don't filter for "started at step 1" — they count anyone who hit any step in order. Returning users who jump straight to /dashboard still register on later steps. Usually fine, occasionally surprising.

Funnels need events to be useful

The funnels in this post all reference custom events at some step — signup_completed, newsletter_subscribed, form_submitted. If you haven't set those up yet, that's the prerequisite. The events guide below walks through it.