What does a hook test measure?

It measures interest, not fun. A hook test asks one question: do strangers stop and click when they see the idea? It cannot tell you if the full game is fun or if players will stay. Those need a playable build.

Measure two steps. The first is the click, which the hook wins. The second is the signup or wishlist, which depends on whether the game looks like something the player already likes. Chris Zukowski watched shoppers look at Steam pages, and the first thing they said was usually the game it reminded them of, not what was new about it. A concept that wins clicks but loses at the second step is novel but not yet clear.

Sources: How To Market A Game, How to make your game acceptable (5 April 2020); Codecks (8 October 2025).

Which cheap tests can I run before I build?

Start with the cheapest test that can fail. Each test below needs only a clip or a small build, and each gives a number you can compare between concepts.

A Steam page is the last step, not the first. Valve needs branding images, a description and a review that you should request at least 7 business days ahead. It also warns that players who wishlist early can be confused if the game changes a lot.

Cheap hook tests, what they need and the number each gives.
TestWhat you needCostNumber to compare
Social clip testA GIF or short clip per conceptTime onlyShares, comments, link clicks
Landing pageOne page per concept with a signupDomain and hostingSignup rate per visit
Small paid testOne ad per concept, same audienceA fixed budgetCost per click, signup rate
Portal playtestA small web buildFree on Poki and CrazyGamesPlaytime, quits, retention
Steam Coming Soon pageCapsules, screenshots, a reviewThe Steam Direct feeWishlists per visit

Sources: How To Market A Game (14 November 2021); Squeaky Wheel (21 June 2024); Poki Playtesting; CrazyGames Basic Launch; Steamworks Coming Soon documentation.

How do I run a social clip test?

Post clips often on channels that allow it, then reuse only the winner on strict channels. Chris Zukowski calls this beta testing your marketing. In his example, the developer of Neon Cyber Desert posted a GIF that got 308 retweets and 2,260 likes before the game had a Steam page. A month later the same GIF and nearly the same headline went viral on Imgur, and the post brought 400 wishlists.

For a hook test, make one clip per concept and keep everything else the same: length, time of day and caption style. Put a tagged link under each clip, so clicks count per concept and not only likes.

Show the hook in the first second of the clip. Do not start with a logo, a menu or a slow pan. Post each concept two or three times on different days, because one post can win or lose on timing alone. Read the comments too. If people ask what the game is, the idea is not clear yet. If they name a game it reminds them of, write that name down. It tells you which players the concept speaks to.

Sources: How To Market A Game, Beta test your marketing (14 November 2021).

How do I run a small paid test?

Show each concept to the same audience with the same budget, and compare the cost per click and the signup rate. Squeaky Wheel ran $500 of Reddit ads for Glyphica in one week, with a maximum cost per click of $0.79. Wishlists rose from 227 in the control week to 386 in the ad week, a lift of 159. Steam UTM data credited 91 of them to the ads.

The ad groups also show how to read a test. Broad Steam subreddits turned 148 tracked visits into 40 wishlists (27%), and roguelite subreddits turned 134 into 41 (30%). A desktop-only group turned 41 into 6 (14%). That is $3.14 per wishlist on the lift and $5.49 on UTM wishlists (our arithmetic).

Glyphica Reddit ad test, control week against ad week.
PeriodWishlistsPer day
Control, 19 to 26 May 202422728
Ads, 27 May to 2 June 202438655
Lift159Not given

Sources: Squeaky Wheel, How many wishlists can $500 worth of Reddit ads get you? (21 June 2024).

How do I test with real players for free?

Put a small web build in front of portal players. Poki Playtesting shows your build to a small group of players who opt in, and each test gives 10 video recordings with button inputs, play duration, country and device. Poki says the first recordings often arrive within minutes, and a baseline test measures how long players stay.

CrazyGames Basic Launch shows a game to a limited audience for 7 to 21 days, the SDK is optional at that stage, and its dashboard has players, average playtime, gameplay conversion and retention. On itch.io, the Gambler's Table prototype got thousands of views from its tags with no description at all.

Sources: Poki Playtesting; Poki Working with Poki; CrazyGames Basic Launch and Full Launch; CrazyGames Requirements introduction; Gambler's Table devlog (May 2025).

How do I read the result?

Compare concepts with each other, and set the pass rule before you start. One number alone means little, because each channel and audience has its own normal. Codecks suggests testing several ideas at once: if one gets 500 likes and another gets 100, you have a reference point.

Give each concept its own tagged link, so every click and signup has a concept label. Then read the rates side by side.

How to read the two steps of a hook test.
Click rateSignup or wishlist rateWhat it means
HighHighBuild a prototype of this concept
HighLowThe hook works, but the game is unclear
LowHighThe idea fits, but the clip or image is weak
LowLowDrop the concept or change the core idea

Sources: Codecks (8 October 2025); Squeaky Wheel (21 June 2024).

How do I count clicks and signups for each concept?

Manalens is the analytics and growth engine built for games. Create one short link per concept with its own utm_content, add the Web SDK to the landing page, and send a signup event. Then compare clicks and signups for each concept in one funnel, and keep the winner.

Sources: Manalens product documentation.

Frequently asked questions

How much should a hook test cost?

Social clip tests and portal playtests cost only time. For paid tests, pick one fixed budget per concept and spend the same on each. Squeaky Wheel spent $500 in one week on Glyphica. Smaller budgets can rank concepts by cost per click, but the signup numbers will be small.

Can I put up a Steam page just to test an idea?

You can, but it is a slow test. Valve reviews the page, asks for at least 7 business days, and warns that early wishlisters can be confused if the game changes a lot. Use clips, landing pages and portal playtests first, and open the Steam page once the concept is set.

What counts as a good result?

There is no universal pass mark, because every channel has its own normal. Set the rule before you start, then compare concepts in the same test. A concept that wins clicks and signups against the others is the one to build. A concept that loses both should change or stop.

Does a hook test tell me if the game is fun?

No. It tells you if strangers want to look. Fun and retention need a playable build. Poki Playtesting and CrazyGames Basic Launch measure playtime and retention, so use them after the hook test, when you have a small build of the winning concept.

Sources and implementation