Grow a Side Project With Your Existing Audience: 3 Tests

Grow a Side Project With Your Existing Audience: 3 Tests

In short6 min read

A developer behind a daily movie game ran three growth experiments after years of flat traffic. Videos and creator outreach flopped; one menu link and one email moved the needle.

Most side-project growth advice tells you to go find new people: post on TikTok, recruit creators, chase the algorithm. A developer from Hungary who runs the daily movie game Actorle tested exactly that playbook this year and kept numbers on everything. The result is one of the clearest arguments you will find for a simpler idea: before you buy attention somewhere new, squeeze the audience you already have.

In this breakdown you will see what the side project earns, which three experiments the founder ran, why two of them produced almost nothing, and how to run the one that worked on your own product this week.

In this article

  1. The project behind the numbers
  2. Experiment 1: Automated short videos
  3. Experiment 2: Custom pages for creators
  4. Experiment 3: A new game for people who were already there
  5. Why the menu link beat TikTok
  6. A quick framework: which growth bet to make first
  7. The cautionary tale: traffic without a list
  8. A note on monetization
  9. Test growth with your current audience

The project behind the numbers

The founder built Actorle in three days in March 2022, at the peak of the Wordle craze. Players guess an actor from their filmography. Launch day was rough: the top Reddit comments called it “insanely hard to win” and “literally unplayable,” partly because a well-known actor was missing from the database. The founder says they nearly deleted it that weekend.

A week later a small British humor site, b3ta, featured it in a newsletter. Coverage on Twitter and Mashable followed, and daily users went from 500 to tens of thousands. Ad income eventually matched the founder’s salary, so they quit their job.

Here is where things stand today, according to the founder:

Metric Figure
Daily visitors, main game Close to 7,000
Daily visitors incl. two sister sites Around 12,500
Revenue Around $4,000/month from display ads
Maintenance A few hours a month
Direct traffic share 63% of visits
Win rate on the “insanely hard” game 69%
Monthly visitors before the experiments Flat between 144k and 172k

That last row is what started everything. After years of flat traffic, flat began to feel like a ceiling, so this summer the founder tried three ways to break through it.

Growing a Side Project From Its Existing Audience: Read your referrers, Capture returning users, Build for current players, Launch with menu link, Fix from feedback

Experiment 1: Automated short videos

The conventional wisdom says daily games grow on TikTok. The founder built a pipeline that generated a puzzle video twice a day and posted it to TikTok, YouTube Shorts and Facebook Reels.

Over ten weeks the pipeline produced 218 videos. Total results: about 51,000 views, 107 likes, 4 comments and 4 YouTube subscribers. TikTok gave each video 250 to 300 views for four days, then nothing on every video after that. In the last 30 days, analytics showed 25 visitors from all of it.

The founder’s own verdict: automated content gets automated results. If they try video again, it will be them on camera, “badly.”

Experiment 2: Custom pages for creators

The second idea was a partnership play. The founder built custom puzzle pages for movie YouTubers and TikTokers, essentially “your own Actorle” tuned to a creator’s audience, plus an application form. Applications so far: zero.

This is a common trap. A creator program asks busy strangers to do work for you, and a form adds friction on top of that. Without a direct, personal ask, an open invitation tends to sit unanswered.

Experiment 3: A new game for people who were already there

In September the founder built Cinections, a Connections-style game built around movie posters. They did not launch it anywhere. It got a link in the site menu and one email to registered players, roughly 4,000 addresses, which produced one spam complaint.

The results, as the founder reports them:

  • First day, behind a sign-in wall: 150 new accounts, five times a normal day
  • By October: about 1,640 people finishing it every day
  • It is now the second most visited page on the site
  • Monthly visitors rose from 163k in August to 183k in September to 201k in October, after a year of flat traffic

The existing players also acted as free QA. Three people wrote in on the same day because one board had five Denzel Washington films when a group needs four. Three others asked to solve the final group themselves instead of having the game auto-complete it. Both fixes shipped within two days.

It is tempting to read this as “short video doesn’t work.” A better reading is that the three experiments asked for very different levels of trust.

Experiment Who it targeted What it asked of them Result reported
Automated videos Strangers scrolling a feed Stop, watch, then leave the app ~25 visitors in 30 days
Creator pages Busy strangers with audiences Apply, then promote your product 0 applications
New game via menu + email People with a daily habit on the site Click one more link ~1,640 daily finishers

The third option worked because it rode an existing habit. One commenter in the thread put it well: people who bookmark a daily game have a habit, and a habit is worth more than any search ranking. The 63% direct-traffic figure is that habit in numbers.

There is also a referrer lesson. The founder notes that a single directory of daily games sends 8,600 visitors a month, more than Bing, Facebook, Reddit and X combined, for zero ongoing effort. Their analytics were pointing at “bookmarks and directories” the whole time. They went to TikTok anyway.

A quick framework: which growth bet to make first

Use this as a sanity check before you spend ten weeks on any channel. Score your project honestly on each question.

  1. Do you have a way to reach past users? An email list, accounts, a Discord, or even a site menu that returning visitors see. If yes, a new feature or sister product launched to them is your cheapest test.
  2. What do your top referrers say? If directories or bookmarks dominate, more listings may beat a brand-new channel.
  3. Does the new channel match how people found you? A game discovered through newsletters and directories may not translate to a swipe feed.
  4. Is the content automated or personal? The founder’s experience suggests automated video is easy to produce and easy to ignore.
  5. Can you measure it within 30 days? The founder knew videos had failed because they checked visitors, not views.

If you answer yes to the first two, start there. New channels are worth testing, but usually after you have harvested the warm audience.

The cautionary tale: traffic without a list

The founder shares one more story. During a fuel shortage in Hungary, they built a crowdsourced map of which stations still had petrol in a day or two. It hit national news and drew more than 100k visitors a day, more than Actorle ever had. When the crisis ended, everyone left. There were no emails and no plan for what those people might need next. The site is dead now.

The takeaway is blunt: when a lot of people show up, decide what you want to keep before they leave. A sign-in wall or an email capture can feel like friction on day one. For Cinections, the roughly 4,000-person list was the launch channel.

A note on monetization

The income here comes from display ads, which means revenue tracks page views. In the comments the founder clarified that the site sees around 33k daily page views with about $200 in daily ad income, roughly a $6 CPM. That math explains why a second game matters: every extra daily page from a returning player adds inventory without new acquisition cost. Your own ad rates will depend on niche, audience location and ad setup, so treat these figures as one data point, not a benchmark.

It is also worth remembering that the original launch was lucky. The founder says plainly that they have no formula. What is repeatable is the habit of measuring, checking referrers, and building for people who already care.

Test growth with your current audience

  • Open your analytics and list your top five referrers for the last 90 days, including direct traffic.
  • Identify every way you can reach past users today: email list, accounts, community, in-app or site menu.
  • If you have no list, add a lightweight sign-up or account option this week.
  • Pick one small new feature, mode or sister product your current users already ask for.
  • Launch it quietly: one menu link and one plain email to your list. Skip paid promotion.
  • Invite feedback and fix the first reported issues within a few days.
  • If a directory or newsletter already sends you traffic, find similar listings and send one short pitch to each.
  • Before testing any new channel, set a 30-day visitor target and kill the test if it misses.

Pull 90 days of referrer data and list the sources

Today, pull your last 90 days of referrer data and write down where your visitors actually come from. Then pick one small feature your current users would want and plan to announce it with a single menu link and email.

Source: My 3-day side project has paid my bills for 4 years (~$4k/month). This fall I finally tried to grow it on purpose. Two of three experiments flopped.

Get new AI automation playbooks by email

One email when new guides and case studies go up. No spam, unsubscribe anytime.

Responses

  1. […] Grow a Side Project With Your Existing Audience: 3 Tests […]

  2. […] Grow a Side Project With Your Existing Audience: 3 Tests […]

  3. […] you want more ideas for getting first buyers without ads, see these three tests for growing a side project with your existing audience. For an honest look at what early revenue often looks like, read $700 in 70 days from solo apps, […]

  4. […] fast, mediocre output. Automating a process you have refined produces leverage. Our guide on growing a side project with your existing audience covers stage two in more […]

Leave a Reply