How-To Guide
How to shortlist YouTube creators for a 20-minute Steam horror demo
Use a 30-slot worksheet to screen YouTube creators for a 20-minute Steam horror demo, recording gameplay fit, evidence and submission routes.
By the Superdirector editorial team.
At a glance
How do I find 30 small creators for my 20-minute Steam horror demo without pitching channels that only cover AAA horror?
Start with channels whose actual videos fit the experience your demo offers. Save the evidence, check the creator's stated submission route and remove duplicates before drafting a pitch. You may finish with fewer than 30 candidates. This guide contains fictional screening examples and blank research slots; it does not recommend any real creator or promise coverage.
A manual research workflow with a 30-slot worksheet, not a list of 30 vetted creators
Start with channels whose actual videos fit the experience your demo offers. Save the evidence, check the creator's stated submission route and remove duplicates before drafting a pitch. You may finish with fewer than 30 candidates. This guide contains fictional screening examples and blank research slots; it does not recommend any real creator or promise coverage.
Write the playable offer before searching
Our fictional teaching demo, The Door at Number Nine, is a roughly 20-minute English-language PC horror game. The player inspects an apartment corridor, notices changes and chooses which door to approach. For this example, assume its demo is publicly playable on Steam and contains sudden loud sounds and flashing lights. These are invented game facts, not observations about a released title. Replace them with facts checked against your build, including relevant content notes.
Record: actual Steam destination; supported platform and language; approximate play length and how estimated; repeated player action; horror style; access instructions; content notes; any real publication restrictions. Do not imply a review key is necessary when the public demo already provides access.
Search by comparable play, then inspect the channel
Choose three real games that share the relevant interaction or format with your actual demo. Search each game's name with “demo,” “first impressions” or “full gameplay.” Open the videos and record what you inspected. A horror thumbnail alone does not show whether a creator enjoys short observation-based games.
For an initial pass, you might choose a 90-day upload window as your own research convenience, not a universal standard. Write the chosen window and inspection date at the top of the sheet. If a channel publishes infrequently, look further back, record why and check its current stated focus. Do not call an older upload “recent” without its date.
Look for language compatibility, short-PC-horror coverage, commentary format and enough room in the channel's usual format for this particular demo. Audience size is secondary to the reason you think the game fits. Missing evidence stays unknown.
Define “small” for your campaign before filtering. For example, you might screen channels against a subscriber ceiling that fits your research capacity, then inspect typical recent views and format. Record your chosen ceiling and why; no universal subscriber cutoff predicts fit or response. A channel above it can remain a documented exception, not an unnoticed change in the rules.
Make inclusion and rejection decisions explicit
Worked acceptance-for-drafting row, entirely fictional: Candidate: Example Channel A, an invented label, not a real YouTube account. Evidence: invented Video A is a narrated playthrough of a short corridor-observation horror game; invented Video B is a short standalone PC horror demo. In the teaching scenario both fall within the chosen 90-day review window. No real dates or URLs are asserted. Fit: the player describes environmental changes and completes short demos in a single video. Language: English in the invented examples. Contact evidence: the teaching scenario assumes an explicit public game-submission form. No real form or address has been supplied. Duplicate/contact check: no previous record within this fictional example. Decision: acceptable as a drafting exercise. In a real tracker, move to “ready to draft” only after entering inspected URLs, dates and the verified submission route. This example is not ready for outreach.
Worked rejection row, entirely fictional: Candidate: Example Channel B, an invented label. Evidence: its invented current series consists of multiplayer survival-build tutorials; the only invented standalone-horror playthrough falls outside the research window and does not reflect that series. Fit problem: the proposed single-player observation demo does not match the channel's current tutorial job. Decision: reject for this game. No contact attempted. Record the reason instead of treating “horror” in the account description as enough.
Explicit AAA-only rejection, entirely fictional: Example Channel C has a horror label, but every upload in the chosen review window covers a major-publisher franchise and none shows small standalone demos. Decision: reject for this campaign because indie-demo fit is unsubstantiated. Do not infer willingness from horror audience size or send a speculative pitch to fill the thirtieth slot.
A third possible decision is “no contact route.” A suitable video does not authorize guessing an email, using private accounts or bypassing a stated no-submissions preference. Keep the fit notes but do not contact that candidate.
Observation play: inspect environmental changes.
Survival-building tutorials: a different current format.
Hypothetical teaching art, not game footage or creator evidence. The observation folder illustrates a closer play-format match for the fictional corridor demo; the survival-building folder illustrates a different tutorial job. Real inclusion decisions still require inspected videos and a verified contact route.
AI-generated teaching illustration. Illustrated format comparison only. Real fit decisions still require inspected videos and a verified public contact route.
Original visual by the Superdirector editorial team. Created 2026-10-20.
Use separate research and contact states
Research state: unreviewed; needs evidence; ready to draft; rejected. Contact state: not contacted; no contact route; do not contact; draft prepared; sent; replied; declined; requested sponsorship. Coverage: unknown; none observed during stated check; coverage observed with URL/date. Keep this separate from replies: a creator may publish without replying.
Before any sending, search your existing tracker by channel URL or stable channel ID, plus the verified submission destination. Merge duplicates, preserve previous actions and honor declines or no-contact preferences. One channel appearing under three comparable-game searches is one candidate, not three outreach opportunities. Record the last action and date; silence is not permission for repeated messages. A sponsorship request does not become unpaid acceptance.
Complete email draft
Replace every bracketed field with verified facts. This is an editable draft, not authorization or an instruction to send.
Subject: [Game name]: short observation-horror demo for consideration
Hi [creator name],
In your [specific video title], you [one precise, observed detail about how they played or explained the game]. That is why I thought [game name] might fit your channel.
I am its developer. It is a roughly [checked duration]-minute [platform/language] horror demo about [one-sentence player action and premise]. The demo is [actual access status].
Play: [verified Steam demo destination and any necessary access instructions] Content notes: [accurate notes from the build] Press materials: [verified, accessible link to materials you own or may share] Publication conditions: [actual conditions, or a verified statement that no embargo applies]
I am offering it for unpaid editorial consideration, not a sponsored placement. There is no obligation to cover it or give a positive opinion. If it is not a fit, no reply is needed.
Thanks, [developer name] [verified project/contact identity]
Do not invent the personalization from a title or summary. Watch enough to substantiate the sentence. Confirm access links and any claimed permission before using the draft. A press kit should not contain rights you do not own or private development credentials.
Copyable 30-slot worksheet
Research date: ____ Chosen initial upload window and reason: ____ Low-frequency exception rule: ____ Game/build/language/access facts checked: ____
For each slot below, complete the same record: Channel URL/stable ID: ____ Evidence video 1 URL, upload date, inspected detail: ____ Evidence video 2 URL/date/detail if available: ____ Inspection date and any recency exception: ____ Language, format and fit reason: ____ Unknowns or exclusion reason: ____ Public contact-source URL and stated submission preference: ____ Duplicate match or prior-contact record: ____ Research state: ____ Contact state: ____ Last action/date: ____ Coverage URL/date if actually observed: ____ Next decision: ____
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
- Unreviewed; not contacted; evidence record: ____________________
Leave unfilled slots blank rather than adding weak matches to reach 30. The useful output is a defensible decision for each inspected candidate and a truthful draft for those that qualify. No outreach, paid research or software integration is necessary to use this worksheet.
Source context: the source study inspected CreatorScout, CreatorFetch and Market Your Games. Their public pages package discovery, shortlisting and outreach as a sequence; the study did not test contact accuracy or produce a verified horror shortlist. This manual worksheet keeps evidence and exclusion decisions visible without requiring a subscription.
The original prototype cited a no-budget creator-discovery question at https://www.reddit.com/r/gamedev/comments/1qvjcav/indie_game_dev_with_0_marketing_budget_how_do_you/ and a developer's outreach account at https://www.reddit.com/r/gamedev/comments/1mvkdac/i_contacted_102_content_creators_it_resulted_in/ . These are supplied discussion leads, not independently reverified here. Their existence does not establish a reply rate, creator availability or the effectiveness of this workflow. No reported numerical results are reused as benchmarks.