Analytics for builders
Set up your first goal when you only get a few signups
By Theo Marsh · Oct 30, 2026 · 9 min read
Why a goal is worth defining at three signups a month
The worry is fair. If 3 people sign up out of 150 visits, that is 2%, and next week it could be 1% or 4% with no change on your site. A rate built on three events is mostly noise.
But a goal is not a rate. It is a list of the people who did the thing, and each one is traceable: which page they landed on, where they came from, what they read first. Three signups a month is three small case studies, and you cannot read them if nothing is counting.
Defining a goal early also costs less than waiting. In Peeko a goal is a definition rather than a counter, so a goal written today counts over months before it was written, and a site can hold twenty of them. You are not starting a clock by writing one. You are choosing how to read events that are already being recorded.
Small segments can also matter more than their traffic share suggests. Ahrefs reported on its own site that AI search accounted for 0.5% of traffic over 30 days and 12.1% of signups over the same 30 days. That is one large site, measured with its own analytics, and Ahrefs says the result "might be somewhat unique to us." The page does not report how many visits or signups sit behind those percentages. Read it as an example of why a thin channel is worth counting, not as a benchmark for your site.
Source: Ahrefs, Does AI Search Traffic Convert Better Than Traditional Search? For Ahrefs, Yes: 0.5% of Visitors Drove 12.1% of Signups, Jun 16, 2025. Retrieved Oct 27, 2026.
You will never see a gap like that if the only number you look at is total visits.
Pick one outcome
Choose the single action that comes right before money or a real conversation. For most small products that is one of these:
- A signup or account creation.
- A demo request or a contact form submission.
- A click on the checkout or pricing button.
Do not start with ten goals. With a handful of conversions, each extra goal splits the same few people into smaller piles. One goal, named in plain words, is enough for the first month. The limit of twenty per site is there for later, when one outcome has grown into several.
If you are choosing between two, pick the one that happens more often and sits closer to the visit. A demo request that you get twice a quarter will teach you nothing in a month. A click on the "Book a demo" button might happen 15 times in that month, and every click is still a person who decided to talk to you.
Count it with no code, or with one line
Peeko records five kinds of click with no code at all: outbound, phone, email, whatsapp and form. If your outcome is a form submission, an email link or a phone number, it is already being recorded once the script tag is on the page. The script is one tag with a site key, it sets no cookie, and it needs no configuration once it is on the page. If you have not installed it yet, install analytics and prove it works in one agent session walks through the check.
For anything else, name the event yourself. There are two ways, and both take one name and one optional string, with no properties object.
The first is an attribute on the element:
<a href="/signup" data-peeko-event="signup-click">Start free</a>
The second is a call from your own script, for example when a form submits:
<script>
document.querySelector('#demo-form').addEventListener('submit', function () {
window.peeko.event('demo-request', 'pricing-page');
});
</script>
The optional string is where you put the one detail you would want later, such as the page the form was on. Keep names short and stable. If you rename signup-click next month, you will have two names to merge in your head.
Then define the goal in the dashboard: give it the event name and a plain label. The five built-in clicks are recorded from the day the script is on your page, so if you are going to use one of them, put the script on first.
Read tiny numbers as counts
With a small site the most useful habit is to stop computing rates and start reading counts.
Take 20 visits from one source. One signup is 5%. Two is 10%. A single person moved the rate by five points. Nothing about your page changed between those two results, and a person who arrived in a good mood is enough to explain them.
So use three rules.
Compare counts, not percentages. "Four signups from search, one from a newsletter" is a statement you can check. "Search converts at 4 times the rate of email" is built on five people.
Do not act on a difference of one or two. Going from 3 signups one week to 5 the next is not a trend. A change you made on Tuesday cannot be credited for it. A difference worth acting on is one that repeats across several weeks, or one that is large next to the total: 12 against 2, not 4 against 3.
Keep a longer window than feels natural. Because goals are definitions, you can look at 90 days of data the day you create one. Use the long view for judgments and the short view only to confirm the goal is firing.
A measurement note: all of this is arithmetic on a handful of events, and it carries no confidence interval. If you want to claim that one page beats another, you need far more traffic than most small sites have. What you can do is read each conversion and notice patterns that are too consistent to be chance, such as three of your four signups reading the same post first.
What a goal can tell you without cookies
A goal needs to know that the person who clicked is the same person who arrived earlier. Peeko handles this without a cookie. A visitor is the first part of a one way keyed hash of their address and browser string, and the address itself is never written to a database or a log. The count of People uses a hash that mixes in the date and is re-salted every day, so nobody can be followed from one day to the next.
A second hash without the date is what tells a returning visitor from a new one and strings a person's page views into a journey. It is the identity behind goals, funnels and revenue. That is why you can ask "which source did this signup come from" and get an answer without a consent banner for a cookie that does not exist.
The limit is the one you would expect. A person who signs up from a phone after reading on a laptop is two visitors to a tool that has no cookie and no login to join them. Expect some conversions to show up as direct or without a source.
Use a two-step funnel when the last step is too thin
If signups are the final step and you only get three, a funnel on top of them will show almost nothing. Add the step before it.
A funnel holds two to five steps, with a completion window of 1 hour, 1 day, 7 days or 30 days, and it covers at most 90 days. It follows a person rather than a visit, which matters on a small site where one reader often comes back two or three times before acting. A site can hold ten funnels.
A useful first funnel for a small product has two steps:
- A view of your pricing page.
- Your signup or demo event.
With 120 people reaching pricing and 6 signing up, you have a number to hold still while you change the page. With 6 signups alone, you do not. Set the window to 7 days if your product is something people think about, and to 1 day if it is a quick tool.
If you ship with an AI coding agent, it can read the results of your goals and funnels without opening the dashboard. Connect Claude Code or Cursor to your site analytics over MCP shows the setup. The command line needs no install: npx @usepeeko/cli speaks the same MCP endpoint an agent does, so a command and a tool call return the same numbers.
Where the few conversions come from
When conversions are rare, source matters more than volume. Visits that arrive from ChatGPT, Claude, Perplexity, Copilot and Gemini are counted from ordinary page views by their referrer, with no middleware involved. If one of your four signups this month came from an assistant, that is worth knowing. Count visits from AI assistants by referrer explains how that count works, and follow an AI referral to signup without cookies shows how to tie it to a goal.
If a signup becomes a paid plan, revenue arrives one of two ways: a one line window.peeko.revenue call from your own site, or a verified Stripe webhook. The two totals are never added together, and amounts in different currencies are listed, never converted. Report revenue from your site or let Stripe confirm it covers which to pick. Skip this for the first month unless money is the thing you are trying to count.
A weekly routine that fits in ten minutes
- Open the goal and read the count for the last 7 days and the last 90 days.
- For each conversion, look at the journey: source, entry page, pages read before the event.
- Write down, in one line, what the conversions had in common. If the answer is "nothing," write that.
- Change at most one thing on the page you think mattered, and note the date.
- Next week, look at the count again. Do not compare it to last week alone. Compare it to the 90 day weekly average.
That is the whole practice. It works with 3 conversions a month and it keeps working with 300, because the goal, the event names and the habit do not change as the numbers grow.
What this approach does not do
It does not tell you whether a change worked. At these volumes it cannot. It does not replace asking the people who signed up how they found you, which is still the best source for a handful of conversions. And it does not count conversions that happen off your site, such as a reply to a cold email, unless you route them through a page that carries the script.
What it does is give you a definition on day one, so that when your traffic grows you already have months of counted history behind the goal.
If you have not started, the trial is fourteen days with no card, it starts when you create your account, and it covers one site. Install the script, define one goal, and look at it next week.