Are your conversion goals answering the right question?

Short answer: a conversion goal should answer a business question — "does this page earn its keep?" — not count taps. Most goals fail because they track button clicks instead of completed outcomes, or because nobody defined what "done" looks like. Define the outcome first, track the smallest set of events that proves it, and review each goal every month.

A goal is a question, not a counter

Every time I audit a small site's analytics, I find the same thing: a list of goals that nobody can explain. "Button click — header." "Form submit — old." "Purchase???" with three question marks, set up by an agency two redesigns ago. The numbers exist, the meaning is gone.

In my experience, the confusion always starts in the same place: people track what is easy to track instead of what they need to know. A button click is easy to track. But "did this page earn its keep?" is the question your business actually needs answered, and a click count will never answer it.

So here is the distinction I would nail to the wall of every marketing team:

• An event records that something happened: a click, a scroll, a download, a video played.
• A goal declares that a specific event — or sequence of events — counts as a business result.
• A metric like conversion rate is just a goal divided by the traffic that could have completed it.

Events are the raw material. Goals are the verdict you attach to them. When everything is a goal, nothing is — you get conversion rates of 40% that mean nothing and dashboards nobody opens, because nobody believes them.

5 goal mistakes that fill your reports with noise

Before the method, the diagnosis. If any of these sound familiar, your goals are counting taps, not answering questions.

1. Tracking the click instead of the outcome

The most common one. The goal fires when someone clicks "Get a quote" — not when the quote request actually arrives in your inbox. Every abandoned form, every validation error, every visitor who changed their mind at the last field still counts as a conversion. Your conversion rate looks healthy while your pipeline stays empty.

2. Counting every micro-step as its own goal

Page viewed, button clicked, step two reached, step three reached, confirmation shown — five goals for one signup. Each one inflates some rate somewhere, and together they make it impossible to say what the site's conversion rate even is. Intermediate steps are useful context for a funnel. They are not goals.

3. No definition of "done"

What counts as a lead? Any form submission — including spam, test entries from your own team, and the "I clicked by accident" ones? A goal without a written definition of the outcome drifts within months. I would write one sentence per goal — "a quote request with a valid phone number that reaches the CRM" — and treat anything else as noise.

4. Goals nobody looks at

Set up once, never reviewed. Then the site gets redesigned, the thank-you page URL changes, the goal silently stops firing — or starts firing twice — and six months of reports are wrong in a direction nobody noticed. An unreviewed goal is worse than no goal: it lends false authority to bad numbers.

5. One goal for the whole site

A newsletter signup on the blog and a checkout on the shop answer two different questions, but many small sites track a single blended "conversion" across both. The result is a number that moves for reasons you can't attribute. One offer, one question, one goal.

How to set a goal that answers something: 5 steps

This takes an afternoon, including the cleanup of your existing goals. Do it with the person who owns revenue, not alone at midnight — the definitions are business decisions, not technical ones.

1. Write the question first. One sentence, with money or pipeline attached: "Do visitors who land on the pricing page request a demo?" If you can't phrase the question, you don't need the goal yet.
2. Define "done" precisely. What exact event proves the outcome? An order confirmation page shown, a lead record created in your CRM, a trial account activated — not the click that started the process. Write the definition next to the goal, where future-you will find it.
3. Track the outcome; keep the steps as context. Instrument the funnel steps as plain events so you can see where people drop off — that leak analysis is exactly what finding where visitors leave is for — but only the completed outcome gets goal status.
4. Name it like you'll read it in six months. "quote_request_received" beats "Goal 4". Include the outcome and the direction: received, completed, activated. Your future self, and anyone who inherits the account, will thank you.
5. Put a monthly review on the calendar. Thirty minutes: does each goal still fire, does it still mean the same thing, did any redesign or form change break it? Calendared beats intended. If a goal's trend suddenly jumps, read it per page type before celebrating — bounce vs exit explains why a single blended rate misleads.

Concretely, here is what the cleanup looks like on a typical small site:

Notice the pattern: every "after" is a completed outcome with a verifiable end state. That is the whole trick.

Three goals worth copying

Adapt these to your site this week. Each one is a single goal with a precise definition.

A local service business: the qualified enquiry

Goal: contact or quote form submitted and containing a reachable phone number or email that passes validation. Not the click on "Contact us", not the page view of the contact page. Question answered: "Does my site generate enquiries I can actually follow up?" Review monthly by spot-checking five recorded leads against what arrived in your inbox.

A small online shop: the completed order

Goal: order confirmation page shown, one count per order ID. Keep add-to-cart and checkout-started as funnel events so you can find the leak, but never as goals. Question answered: "Which pages and sources produce paying customers?" The order count then feeds the number that ranks those pages and sources by what they actually pay: value per visit. When the number moves, compare the same goal before and after the change — the before-and-after discipline from working without A/B tests applies here directly.

A SaaS or newsletter: the activated signup

Goal: account created and email confirmed, or first meaningful action completed — not the signup form submission alone. Unconfirmed signups are intentions, not conversions. Question answered: "Does my landing page create users, or just form-fillers?" This is the goal most teams overcount, often by a factor of two or more.

One goal, three views: the only tooling you need

A goal alone gives you a number. To act on it, you need three views of the same visitors:

• Analytics for the what: traffic, sources, pages, and the goal trend itself.
• Session replays for the why: what visitors who didn't convert actually did.
• Heatmaps for what gets ignored: whether anyone even sees the path to your goal. If you misread these, here is how to read a heatmap without fooling yourself.

That is the gap SessionInsight fills: analytics, session replays, heatmaps and goals in one lightweight script, with a cookie-less mode for basic analytics. You can see how it works or check the free plan — one website, 50,000 pageviews a month, 10 goals included.

If you are still deciding which analytics tool should run underneath all this, I laid out the trade-offs in GA4 alternatives for small sites — including the honest reasons to stay on GA4.

FAQ