Guide

Which Pages Drive Clicks on Your Start Free Button? Ask Your Agent

Your agent finds the page that gets visits and loses the Start free click. You change that one page. Prompts and a demo answer.

Which Pages Drive Clicks on Your Start Free Button? Ask Your Agent

Your agent ranks every page by how often visitors click Start free, and names the page losing signups. You get that answer from a few prompts, then you change the page.

You need Agent Analytics connected to your agent, and these reads are on the Pro plan.

The example answers below are made-up demo data from an invoicing app, with round numbers.

Step 1: track every Start free button

A page view tells you who arrived. A click on Start free tells you who wanted the offer.

Find every Start free button and link on my site. Track clicks on each one, and note where it sits: nav, hero, pricing, or footer. Change nothing else. Show me the diff and wait for my OK before you push. After it ships, confirm that the clicks arrive in Agent Analytics with the page recorded.

Read the diff. This change goes on your live site. After it ships, your agent does the reading.

Step 2: rank pages by click rate

Raw click counts favor the busiest page. Click rate, clicks divided by visitors, shows which pages send people through the button.

For my site, last 30 days: for each page, show the unique visitors who viewed it and the unique visitors who clicked Start free on it. Work out the click rate per page and rank by rate. Leave out pages with fewer than 100 visitors and say how many you left out.

Example answer, demo data:

Start free clicks by page, last 30 days (demo data, made up)
Page                           Visitors   Clicked   Rate
/pricing                            400       100    25%
/features                           500        50    10%
/                                 2,000       100     5%
/blog/late-payment-templates      3,000        30     1%
/docs                               800         0     0%
Left out, under 100 visitors: 14 pages

Each visitor counts once per page. Someone who views three pages is counted in all three rows. Do not add the Visitors column up.

Step 3: find the pages that lose people

A low rate has two causes. The page has a Start free control and people skip it, or the page has no control. Your agent can tell them apart by reading the repo.

For each page in the bottom half of that ranking, open the page source in the repo and tell me whether it has a Start free control, where it sits, and what the copy says. Then look at session paths for my site, with a Start free click as the goal, last 30 days. Tell me which entry pages get the most sessions and how many of them reach the click in the same session. Give me the one page most worth changing and the reason.

The click rate credits the page where the click happened. Session paths credit the page where the session started. A blog post can show a 1% click rate on the post itself and still start sessions that end in a click on /pricing. Ask for both before you call a page a loser.

Example answer, demo data from the same app:

Largest gap: /blog/late-payment-templates
  3,000 visitors, 1% click on the page itself
  Source has one Start free link, in the footer, below the last section

/docs: 800 visitors, 0 clicks
  Source has no Start free control

Session paths, entry /blog/late-payment-templates:
  3,500 sessions, 140 reach a Start free click in the same session
  Most of those clicks happen on /pricing after the visitor leaves the post

Most worth changing: /docs (no button). Second: the blog post (button too low).

The answer names the page, the button, and the line of source to change.

Step 4: check one page before you change it

Before you change a page, ask your agent to confirm the leak on that page alone:

Build a funnel for my site where step 1 is a view of /features and step 2 is a Start free click on that same page. Count by session, last 30 days. Show both counts and the rate.

A session counts when the click on that page follows the view in the same session. Run this on the one or two pages the ranking flagged.

Step 5: change one page, then read it a week later

Pick the page from step 3 and ship one change:

Fix the top page from the ranking with one change only: [describe it, for example add a Start free button under the hero on /docs]. Make it on a branch, show me the diff and wait for my OK before you push.

Tag the date in project context. A week later, read the result with the keep, revert, or extend rule from Let Your Agent Ship a Landing Change, Then Read It 7 Days Later. Ask your agent to measure Start free clicks on the changed page, leave out the day you shipped, and say which page it measured.

What these numbers cannot tell you

  • A click and a signup are different. A page can send many clicks and few completed signups. Once you know the pages, ask your agent to follow a page view, then a Start free click, then a completed signup. Question 4 in Six Questions to Ask Your Agent About Your Product covers which entry page sends signups.
  • A path only covers one session. A visitor who reads a post today and clicks next week still counts as a click. The path only includes a click in the same session as the post.
  • Small pages are noisy. A page with 20 visitors and 5 clicks is a lead. Your agent should say when a page is under your minimum.
  • One window is one window. Ask for the same numbers over 30 days and over the latest 7. A page that moved shows up in the gap.

Keep it running

The Grok Bot guide shows how to run a saved prompt on a schedule. A weekly version of step 2:

Every Monday at 9:00, run the Start free click rate by page for my site for the last 7 days, compare it with the 7 days before, and tell me the one page whose rate changed most and by how much. If nothing moved by more than 20% relative, say so in one line.

That schedule lives in Grok Bot.

For your agent

The details behind these prompts:

Related posts