# How to Set Up Analytics for a Side Project You'll Actually Check
> Set up analytics for a side project you'll actually check: one job, five numbers, a privacy-friendly tool, UTMs, and a Monday 15-minute review.

```yaml
url: "https://sidehunt.io/studio/how-to-set-up-analytics-for-a-side-project-youll-actually-check"
markdown: "https://sidehunt.io/studio/how-to-set-up-analytics-for-a-side-project-youll-actually-check.md"
title: "How to Set Up Analytics for a Side Project You'll Actually Check"
slug: how-to-set-up-analytics-for-a-side-project-youll-actually-check
category: Growth
tags:
  - analytics
  - side projects
  - Plausible
  - Umami
  - Cloudflare Web Analytics
  - UTM
  - privacy
  - web analytics
published_at: "2026-10-13T11:00:22.502Z"
updated_at: "2026-10-13T11:00:22.838Z"
featured_image: "https://txmhk1zrnc.ufs.sh/f/xSkWTCqmKWx93i3Kwb1duKC2cTLjUHtkW4FmRlen8BZa9JpY"
```

## Article

Set up analytics for a side project you will actually check by picking one job for the data, tracking five numbers at most, installing a privacy-friendly tool in one evening, and reviewing it for fifteen minutes on Monday. Skip the daily dashboard habit. Most side project analytics installs fail because they ask for more attention than a day job leaves you.

This guide is for employed builders who want honest signals about traffic and use, not a second job staring at charts. You will decide what the numbers are for, pick a tool with facts checked on vendor pages in October 2026, tag your distribution links, and run a Monday ritual that fits a weeknight. Nothing here is legal advice about cookies or consent. If you serve EU visitors and use cookies, read your own counsel or a primary regulator guide before you ship a banner.

## Why do most side project analytics installs fail?

Three patterns show up again and again:

- Too many events. You wire forty custom events on night one, then never look at them. The setup feels productive. The dashboard becomes noise.
- Daily checking guilt. A red realtime map on your phone at lunch trains you to open analytics when you should be doing your job. When you miss a day, you feel behind, so you stop opening it at all.
- Cookie banner theater. You paste a heavy consent stack "just in case," visitors bounce before the page loads, and half your traffic never gets counted. You then "fix" the product with incomplete data.

The fix is not a prettier dashboard. It is a smaller job for analytics, a short list of numbers, and a fixed review time that does not compete with your day job.

## What do you need analytics for?

Pick **one** primary job. Write it in one sentence and stick it at the top of a note you will open on Mondays.

- Distribution habit feedback. You post or share once a week and want to know which source sent real visitors and signups. The SideHunt guide on [stealing one weekly distribution habit for a side project](https://sidehunt.io/studio/how-to-steal-one-weekly-distribution-habit-for-a-side-project) is the habit side. Analytics is the scorecard.
- Conversion path. You care about landing page to signup (or signup to first useful action). You need page and goal counts, not forty events.
- "Is anyone using this?" You shipped something quiet and want a weekly pulse: visitors, one core action, and one failure signal. Enough to know whether to keep evenings on it.

If you try to do all three at once, you will rebuild Mixpanel in your head. Pick one for the next month. You can change the job later.

## Which five numbers are worth tracking on a side project?

Cap yourself at five. A deeper product-metrics framework (activation, tiny cohorts, what to ignore before 100 users) lives in the IndieHunt guide on [tracking product metrics before your first 100 users](https://indiehunt.io/studio/how-to-track-product-metrics-before-100-users). This post does not rehash that. For side project analytics you check after work, these five are enough:

- Visitors (unique people on the marketing site or app, weekly).
- Signups or activations (accounts created, or people who completed the first useful action if signup is free and empty).
- One core action (the thing that means they got value: published a list, sent an invite, exported a file).
- One source attribution (which utm_source / referrer sent this week's visitors or signups).
- One failure signal (404 spikes on key pages, payment errors, or "empty results" events. Pick one. Do not turn this into full error tracking; SideHunt already covered monitoring separately).

Write the five names on a sticky note. If a sixth number keeps asking to join, it waits until next quarter.

## Which analytics tool should a side project use?

You want something you will open, not something that needs a certification. Below are privacy-friendly options with facts checked on vendor pages in October 2026. Prices change; recheck before you budget.

### Plausible (Cloud or self-host)

Plausible is a simple, privacy-friendly analytics product. Its homepage and [data policy](https://plausible.io/data-policy) say it uses no cookies and no persistent identifiers, and that visitor data is processed in the EU on European-owned infrastructure. The [cookieless analytics page](https://plausible.io/cookieless-web-analytics) states that because no personal data is collected, no consent banner is required under most privacy regulations including GDPR and CCPA (that is Plausible's claim; it is not your legal advice).

Cloud pricing shown on the homepage in October 2026 starts at **$9/month Starter**, **$14/month Growth**, and **$19/month Business** at the 10k pageview tier. [Plausible's subscription docs](https://plausible.io/docs/subscription-plans) (updated Sep 29, 2026) say all plans share the same pageview tiers, Starter is for one site and solo use, Growth adds multiple sites and sharing, Business adds funnels, revenue, custom properties and API access, and you get a 30-day free trial with no credit card. Yearly billing saves two months. There is also a free self-hosted Community Edition you operate yourself.

**Pick Plausible Cloud** if you want a one-page dashboard and are fine paying a small monthly fee. **Pick the Community Edition** only if you already like running servers for fun.

### Umami (Cloud or self-host)

Umami is open source and can be self-hosted. Its [FAQ](https://docs.umami.is/docs/faq) says it does not collect personally identifiable information, does not use cookies in the tracking code, and that you do not need to display a cookie notice for the tracker alone. [Umami Cloud pricing](https://umami.is/pricing) (checked October 2026) includes a free **Hobby** plan with up to 100K events per month, 1 website, and 6 months retention; **Pro** at $20/month with 1M events; and higher tiers. The pricing FAQ says Umami does not use cookies or collect personal data, and positions the cloud product as not requiring a cookie consent banner. Self-hosting is free of Umami subscription fees; you pay for your own server and maintenance. Cloud servers are in the US and EU per their FAQ.

**Pick Umami Hobby** if you want free managed analytics for one low-traffic site. **Self-host** if you want unlimited retention and already run Postgres.

### Cloudflare Web Analytics

If your site already sits on Cloudflare, [Cloudflare Web Analytics](https://developers.cloudflare.com/web-analytics/about/) provides free, privacy-first analytics. Cloudflare's docs say it does not collect or use visitors' personal data, and that you can enable it with a JavaScript beacon (or automatic setup when proxied) without changing DNS for the non-proxied path. It is available on all plans. This is a strong "I already pay Cloudflare nothing extra" default for basic page views and visitors. It is thinner on product goals than Plausible or Umami, so pair it with a simple server log of signups if conversion is your job.

### When is GA4 overkill?

Google Analytics 4 is free and powerful. For a nights-and-weekends side project it is usually the wrong shape: property setup, consent modes if you use cookies, and reports that reward exploration more than a fifteen-minute Monday. Use GA4 when you already live in Google ads and Search Console and need that ecosystem. Otherwise start with Plausible, Umami, or Cloudflare Web Analytics and graduate only when a concrete question requires it.

## Do you need a cookie banner for analytics?

**Not legal advice.** In plain words: tools that set cookies or build persistent visitor profiles often need a consent step for analytics in the EU. Tools that claim to be cookieless and to avoid personal data (Plausible, Umami's tracker, Cloudflare Web Analytics as described in their docs) argue that an analytics consent banner is not required for that tracker alone. Your site may still use other cookies (auth, payments, chat). Treat vendor marketing as a starting point, not a courtroom answer. If you are unsure, use a cookieless option and keep your privacy policy honest about what you collect.

![](https://txmhk1zrnc.ufs.sh/f/xSkWTCqmKWx9acIhHAZJsOEgxD9oIMj2XZ36Y1qQH7rah5db)

## How do you install analytics in one evening?

Block two hours. Do not open a second tool "for comparison."

- Create the account on your chosen product (or enable Cloudflare Web Analytics on the site).
- Add the site and copy the snippet they give you. It is usually one  tag in the site layout or a tag manager entry. Paste it once in the shared layout, not page by page.
- Deploy to production (or a preview URL you will hit from your phone).
- Open the live site in a private window , click three pages, and confirm a visit appears. If nothing shows, check ad blockers on your own browser first.
- Define one goal if the tool supports it: signup thank-you URL, or one custom event for the core action. Stop there.
- Write the Monday note with your one job and five number names. Put a calendar reminder for next Monday.

High-level only on purpose. Each vendor's install docs change; follow the snippet they show you tonight, not a tutorial from 2023.

## How should you tag distribution links with UTMs?

UTMs turn "someone showed up" into "this newsletter / this board / this weekly post showed up." Google's Analytics Help page on [URL builders and campaign parameters](https://support.google.com/analytics/answer/10917952) says you should always use `utm_source`, `utm_medium`, and `utm_campaign`, and that parameter values are case sensitive. Plausible and Umami both report UTM campaigns as well, so the same links work across tools.

Keep a tiny convention that matches your weekly distribution habit:

- utm_source = the place (for example reddit , newsletter_name , sidehunt )
- utm_medium = the type ( social , email , referral )
- utm_campaign = a short week label ( 2026w42_packlist )

Lowercase everything. One spelling per source. Paste the tagged URL into the place you publish, and save the plain mapping in the same Monday note. If you have not picked the weekly habit yet, do that first with the SideHunt distribution guide linked above, then add UTMs on week two.

## What does a Monday 15-minute review look like?

Same time every week. Phone on Do Not Disturb for the side project folder if you must. Open only the analytics tab and your note.

- Minutes 0-3: Write last week's visitors, signups/activations, and core actions. No screenshots of realtime maps.
- Minutes 3-7: Check the top source. Did your weekly distribution action show up? If yes, keep it. If no, either the UTM was wrong or the post did not land.
- Minutes 7-11: Glance at the failure signal. One spike worth a ticket, or "quiet, good."
- Minutes 11-15: Write one sentence: keep / tweak / stop for this week's distribution or product focus. Close the tab.

If the review regularly runs over fifteen minutes, you are tracking too much. Delete a chart, not your evening.

![](https://txmhk1zrnc.ufs.sh/f/xSkWTCqmKWx9KQdbZ250h7ago5xuVX3wrym1lZ9DEkY8UWez)

## What should you ignore?

- Bounce rate theater. On a one-page landing site a "bounce" can mean they read the whole page. Do not optimize for it alone.
- Realtime world maps. Fun on launch day. Useless every Monday.
- Forty custom events. You will not read them. Track the core action and one failure signal.
- Comparing yourself to SaaS Twitter dashboards. Different traffic, different jobs, different hours in the day.
- Daily open streaks. A weekly habit beats a broken daily one.

## A made-up example: Ravi and Trailmint

Everything in this example is made up: the person, the product name, and every number. It is here so you can see the steps with concrete details.

Ravi works a full-time job and built **Trailmint** on weekends: a packing checklist for weekend hikers that remembers what you actually used last trip. His one analytics job: distribution habit feedback. He posts one helpful packing tip in a hiking forum each Sunday and wants to know if that tip sent signups.

He skips GA4. He already uses Cloudflare for DNS, so he turns on Cloudflare Web Analytics for page views, and adds Umami Hobby for one signup goal on the thank-you URL. Five numbers on his Monday note: weekly visitors, signups, "first checklist saved" (core action, logged in his own database and pasted into the note), `utm_source` from the Sunday post, and 404s on `/pricing`.

Install evening: thirty minutes for the snippets, ten minutes to confirm a private-window visit, twenty minutes to write UTM links for the next four Sundays. After three Mondays the note shows: week 1 = 40 visitors / 2 signups (forum post), week 2 = 18 / 0 (he forgot the UTM), week 3 = 55 / 5 (better tip + fixed tag). He keeps the Sunday habit and stops opening analytics on weeknights.

## A weekend and 14-day setup checklist

Times below are planning guesses, not measurements.

When

Task

Done when

Weekend evening 1 (45 min)

Write the one analytics job and the five number names

Note exists at the top of your Monday doc

Same evening (60-90 min)

Pick Plausible, Umami, or Cloudflare Web Analytics; create account; paste snippet; deploy

Private window visit appears

Same evening (20 min)

Add one goal or thank-you URL; ignore every other event

One conversion shows after a test signup

Next weeknight (30 min)

Build UTM links for the next 2-4 distribution posts; save the mapping

Tagged URLs ready before you publish

Monday 1 (15 min)

First review ritual; expect thin data

Five numbers written; one keep/tweak/stop sentence

Monday 2 (15 min)

Compare to Monday 1; fix missing UTMs

Source column is readable

Day 14 (15 min)

Decide whether the tool and the five numbers are enough for another month

No new events added "just in case"

## Common traps

- Installing two analytics tools "to compare." You will trust neither and maintain both.
- Tracking signed-in product behavior only in the web analytics tool. For early products, a database count of the core action is often clearer. Analytics watches the marketing site and acquisition.
- Changing UTM spellings every week. Reddit and reddit split your report.
- Building a consent banner for a cookieless tool because a blog said every site needs one. Match the banner to cookies you actually set.
- Opening analytics during work "just for a second." That is how daily guilt starts. Wait for Monday.
- Treating a quiet week as product death. Side projects move in weekend pulses. Look at four weeks, not one.

## FAQ

### Can I self-host and forget cloud pricing?

Yes, if you enjoy ops. Plausible offers a Community Edition and Umami is fully self-hostable. Budget your own time for updates and backups. For most employed builders, Umami Hobby or Plausible Cloud is cheaper than an evening of broken Postgres.

### What if I already have GA4 installed?

Leave it for a month if Search Console or ads need it. Add a simple privacy-friendly dashboard for the Monday ritual, or strip GA4 down to the same five numbers and hide everything else. Do not run three overlapping event schemas.

### Should product analytics and web analytics be the same tool?

Not at side project scale. Web analytics answers "who arrived and from where." Your database answers "who activated." Connect them with a week label and UTMs, not a warehouse.

### Is a cookie banner required if I use Plausible or Umami?

Those vendors say their trackers do not use cookies and do not need an analytics consent banner. That is their position as of the pages cited above, not legal advice for your site. Other scripts on your pages may still need consent.

## Where to start this weekend

Write the one job. Pick five numbers. Install one tool. Tag next week's links. Put fifteen minutes on next Monday's calendar. That is a complete analytics setup for a side project you will actually check.

If you are still hunting for distribution ideas after the numbers are quiet, browse launches and builders on [SideHunt](https://sidehunt.io/) and steal one weekly habit worth measuring.

## Links

- [Studio post (HTML)](https://sidehunt.io/studio/how-to-set-up-analytics-for-a-side-project-youll-actually-check)
- [AI-friendly Markdown](https://sidehunt.io/studio/how-to-set-up-analytics-for-a-side-project-youll-actually-check.md)

## Explore

- [Studio index](https://sidehunt.io/studio)
- [Blog](https://sidehunt.io/blog)
- [Launches](https://sidehunt.io/launches)

_HTML version: https://sidehunt.io/studio/how-to-set-up-analytics-for-a-side-project-youll-actually-check_