Where to Find Candidates: 20 Minutes a Day, 10 Candidates
Three starter channels, 20 minutes daily, steadily 3-5 new game candidates into a candidate table. A thin pool makes every later gate useless.
A true scene first
Two beginners start the same day. A thinks of a game and builds immediately — two weeks later the site is done and almost nobody searches the game. B spends three days filling a 12-row candidate table, validates row by row, and picks a rising new game at row five. Three weeks later A is writing the eighth article for a site nobody visits; B’s site is already being indexed.
The difference is not diligence. It is the candidate pool. Judging one game takes 10 minutes; what actually decides your output is how many options you hold. So lesson one of picking is not judging — it is collecting.
Step 1: scout three channels daily (20 minutes)
| Where | What to watch | Pick |
|---|---|---|
| itch.io newest (filter: Play in browser) | New games daily, fewest watchers, least competition | 3-5 |
| Steam New & Trending / wishlist charts | High quality but crowded with scouts — move fast | 3-5 |
| Roblox Discover (sort: Rising / Up-and-Coming) | Where “redeem codes” searchers are densest | 3-5 |
One picking standard only: does the game have enough mechanics, items, and bosses to write guides about? Pure match-three games have no guide demand — skip. Unsure? Log it anyway; the next lesson verifies with data.
Step 2: log the candidate table
Open a table (paper or spreadsheet — it becomes the workbench for every later lesson):
Game | Channel | Date found | One line: why worth a lookThe date is mandatory — later lessons keep asking “how long has this game been out”; without dates you cannot answer.
Rhythm: 20 minutes daily, a steady intake of 3-5 rows. Do not build anything until the table holds 10 rows. A thin pool turns “validation” into “talking yourself into it”.
Three classic mistakes (made for you in advance)
- Logging only what you like. Your taste is not search demand. This lesson only collects; “do people search it” belongs to the next lesson’s data.
- One candidate and go. A single option means no comparison — you cannot know the one you hold is the better one. Ten rows minimum.
- Scouting one day only. The pool is a pipeline, not a one-time purchase. Twenty minutes daily keeps selection supplied forever.
Three words to know (just these)
- Candidate pool: every game awaiting validation — your selection inventory.
- Channel: a place new games surface. The three above are starters; they last a long time.
- Golden window: the 2-8 weeks after a game takes off, holding the bulk of its lifetime searches — the date you logged is how you judge the window.
Going deeper: once the three starter channels are habit, mine harder (SteamDB player curves, low-follower social anomalies, aggregated portal sitemaps — nine channels with full steps) in the repo reference docs/sourcing.md, section “9 sourcing channels”. Not yet — fill the 10 rows first.
✅ Acceptance (all must hold)
- Candidate table ≥10 rows, each with channel and date found
- Each row answers “why logged” in one sentence
- ☐ The 20-minute scout is now a daily habit
Next lesson
Ten candidates in hand — which are real demand? Next lesson: five minutes per candidate in Google Trends to confirm people actually search, not that it merely looks hot. Go to Lesson 4 · Reading Google Trends