64 article notes: 11 HTMAG posts + 47 analyst, practitioner and developer readings. 30 are in the original June–September research window. Sorted by source publication date; verification dates and access limits are shown separately.
13 practical readings. From a clear hook to a credible test.
Game World Observer, Derek Lieu, Game If You Are, IMPRESS and Game Developer—with an earlier set of 26 proposed actions for a five-person demo team. For the new solo and small-team action plans, open Developer Playbooks. All plans are editorial proposals, not proven recipes.
Analyst comparison using Steam review play-device indicators
Carless compares overall handheld popularity with reviews from people who played mostly on Steam Deck. Those measures answer different questions: a large audience and a high handheld share need not identify the same games.
NUMBERS, WITH THEIR DEFINITIONS
The September 25 comparison filters for games above $1m gross revenue and 100 Deck reviews; UFO 50 has a 19.2% majority-Deck review share.
Use it forOur proposal: use relevant handheld comparables to prioritize control and readability tests, then validate the experience on hardware.
Do not concludeThe indicator concerns reviewers with over half their playtime on Deck, not all buyers or revenue. The filtered sample excludes smaller outcomes. Switch performance remains a hypothesis for any new game.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Check controller navigation and small-screen text.
Keep device-specific review share separate from sales forecasts.
Analyst sales estimates and named developer comments
Carless examines a vehicle sandbox whose community creates different reasons to play and watch. Developer comments connect its continuing audience growth with mods, creator videos and an established simulation technology.
NUMBERS, WITH THEIR DEFINITIONS
GameDiscoverCo estimates 1.3 million Steam copies sold in the twelve months preceding its September 22, 2026 article.
Use it forOur proposal: test whether the same mechanic supports several understandable player-made situations before expanding content production.
Do not concludeA mature technology business is not a short-development benchmark. Estimates are not audited sales; interview explanations do not isolate marketing effects. Mods also create support and rights-management obligations.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Capture contrasting uses of one mechanic.
Record creator reach separately from attributable purchases.
Self-selected survey linked to Steam libraries; self-reported discovery methods
Carless compares reported discovery habits across age groups in GameDiscoverCo’s Steam Fan Snapshot panel. The practical value is a question for audience research: whether a planned channel matches the players being sought.
NUMBERS, WITH THEIR DEFINITIONS
The September 18 article describes a panel of over 16,000 worldwide Steam players recruited through online giveaways; respondents could select multiple discovery methods.
Use it forOur interpretation: test channel fit with the intended audience before allocating campaign effort.
Do not concludeGiveaway recruitment is not random sampling. Age-group associations do not establish channel effectiveness, acquisition cost or incremental sales. Subgroup sample sizes are not supplied in the accessible text.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Ask target players where they found recent games.
Separate reported discovery from tracked visits, wishlists and purchases.
Commercial analyst estimates and observational launch analysis
Elliott examines Wardogs’ launch through estimated wishlist conversion, acquisition timing and audience overlap. Its pre-order activity makes the distinction between total copies and conversions from the remaining launch wishlist pool especially relevant.
NUMBERS, WITH THEIR DEFINITIONS
Alinea’s September 16 article estimates an early-access launch pool of about 1.8 million Steam wishlists and 11% conversion after seven days.
Use it forOur interpretation: keep pre-orders, launch wishlists and post-launch conversions as separate measures.
Do not concludeThese are Alinea estimates, not Valve or publisher accounts. A selected large shooter launch is not a small-team benchmark; price, community and wishlist recency are not isolated causes of performance.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Record the starting pool and conversion window.
Compare acquisition-age cohorts without treating timing as a causal explanation.
Commercial analyst estimates with illustrative budget and recoup scenarios
Carless uses a monthly release cohort to examine how production costs and publishing terms change the revenue needed to sustain a studio. The analysis supplies a useful counterweight to individual breakout stories.
NUMBERS, WITH THEIR DEFINITIONS
In its September 15 snapshot, GameDiscoverCo estimates that 16 of 1,230 January 2026 Steam paid releases exceeded $1 million in lifetime gross revenue.
Use it forOur interpretation: evaluate a production plan against several net-receipt scenarios and the time between releases.
Do not concludeThe cohort mixes budgets and experience levels. Cost and recoup assumptions are illustrative; estimated gross and hypothetical net are not audited profit or a particular studio’s success probability.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Model costs and cash timing separately from gross sales.
Compare releases with similar scope and team circumstances.
Developer figures reported in an analyst interview; accessible lead section
The public lead case describes a substantial update and more targeted outreach at Yet Another Zombie Survivors’ exit from Early Access. Returning owners explain why renewed activity must be measured separately from new purchases.
NUMBERS, WITH THEIR DEFINITIONS
Awesome Games Studio’s Diana Biszczanik reports 88,000 Steam copies sold in the first two weeks after the August 20, 2026 1.0 launch, compared with 93,000 in Early Access’s first two weeks.
Use it forGive returning players a clear reason to revisit and explain the new content to relevant creators.
Do not concludeOne successful studio’s self-report does not establish a typical 1.0 result or isolate outreach effects. Paid material after the lead case was not reviewed.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Measure returning owners, new players and purchases separately.
Show the update’s concrete additions in a focused outreach pitch.
Practitioner analysis of commercial revenue estimates and selected examples
Zukowski challenges the rush to finish horror games for Halloween. His monthly release comparison and selected catalogue examples separate the launch decision from later seasonal promotion.
Use it forChoose a launch window around readiness and relevant competition; evaluate seasonal promotion separately.
Do not concludeUnequal release-age windows, a model revenue floor and a revenue-filtered subset limit the month comparison. It does not prove June causes better sales. The author also cautions against moving an already committed October launch because of this post.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Compare similarly scoped releases and retain their release years.
Record launch readiness and later promotional opportunities as separate decisions.
Analyst cohort analysis using proprietary Steam concurrency records
Carless compares demo and full-game peak concurrent players across a filtered release cohort. The distribution makes demo engagement useful for scenario planning while showing why a median cannot forecast an individual launch.
NUMBERS, WITH THEIR DEFINITIONS
GameDiscoverCo reports a 3.01× median full-game/demo peak-CCU ratio for 2,569 Steam games released January 2024–August 2026 whose demos exceeded 50 peak concurrents.
Use it forCompare a range of engagement outcomes for similar games before using demo results in a budget.
Do not concludeThe selected cohort excludes games without a qualifying demo. All-time peaks have unequal observation windows; pricing, promotion and game quality differ. The ratio is observational, and the article’s sales conversions are estimates.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Keep the demo threshold, release window and peak-CCU definition with the comparison.
Separate concurrent players from purchases and revenue.
Commercial analyst estimates for a selected major-publisher release
The Steam section distinguishes two demo measures: the share of buyers who tried a demo and the share of demo players who purchased. Those denominators answer different questions.
NUMBERS, WITH THEIR DEFINITIONS
As published September 8, Alinea estimates that about 40% of Onimusha’s Steam players previously played its demo, while 25% of demo players bought the Steam game.
Use it forName the denominator and observation date whenever reporting demo conversion.
Do not concludeThese are model estimates for an established franchise, not audited sales or an indie benchmark. The figures do not isolate the demo’s incremental effect, and the exact conversion window is not disclosed.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Track demo-to-purchase conversion separately from prior-demo participation among buyers.
Compare matching platforms, regions and acquisition windows.
Public newsletter preview linking developer reports and platform documentation
The accessible news section connects discovery reporting with original developer accounts and platform-documentation changes. It points back to the How Many Dudes? retrospective and flags that creating a third-party Steam event does not guarantee promotion.
Use it forFollow the original evidence behind each campaign claim; record event eligibility separately from promised visibility.
Do not concludeThis is a digest of the accessible preview. The post is marked Paid; its subscriber-only remainder was not reviewed. No paid ranking table or sales estimate is reproduced here.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Keep a dated log of the original developer or platform source.
Check current Steamworks event guidance before committing campaign resources.
Commercial analyst estimates; original cohort analysis
Alinea challenges the idea that a middle revenue rank identifies a mid-budget game. Its analysis of ranks 101–1,000 finds a large catalogue presence, including former blockbusters, so current-year chart position alone is a poor production-budget comparison.
NUMBERS, WITH THEIR DEFINITIONS
Alinea estimates that 83% of the 900 games ranked 101–1,000 by 2026 Steam gross revenue launched before 2026.
Use it forChoose comparables by release age, genre, price and production scale before comparing their current revenue ranks.
Do not concludeAlinea's model is not Valve accounting. The selected revenue band is not a representative sample of all indie releases; chart rank does not establish a game's development budget.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Separate new releases from catalogue games in the comparison set.
Keep lifetime outcomes distinct from current-year estimates.
Carless uses Hearth and Hamlet to examine compact incremental games that emphasize accessible progression and a finite experience. The public section describes a solo developer working alongside a teaching job and publisher support, including regional distribution. It also exposes a positioning problem: the attractive settlement suggests freeform city building, but the layout is predetermined. Some negative feedback came from players expecting a deeper city builder. The useful opportunity is a focused progression loop with a clear fantasy, while the practical warning is to show exactly what players control before a persuasive screenshot creates broader expectations.
NUMBERS, WITH THEIR DEFINITIONS
Headline reports approximately 100k copies; game was nine days old.
Publisher supplied >200k launch wishlists and 7h20 median playtime.
A comparison of top-five genre tags among commercially successful 2021 and 2025 releases shows a shift toward run-based hybrids, life simulation and idlers. Both unit- and revenue-based views are provided.
NUMBERS, WITH THEIR DEFINITIONS
>50k-unit cohorts: 261 games in 2021 versus 469 in 2025.
Their median current US prices: $19.99 versus $16.99.
>$1m-gross cohorts: 196 versus 311 games.
2025 unit-cohort rank changes: Roguelite +39 to #1; Life Sim +31 to #6; Idler +46 to #15.
Use it forCompare both copies and revenue; use a rising tag to locate comparables, not as a design brief.
Do not concludeWinner-only cohorts do not measure success probability or supply. Tags overlap; prices are current, not necessarily launch prices.
A digest of Newzoo's updated global forecast places PC growth in context. It explicitly explains why estimates from providers can disagree when off-store spending and geographic coverage differ.
NUMBERS, WITH THEIR DEFINITIONS
Newzoo forecast: global software/services spending $213.9bn in 2026, +6.1% year on year.
PC forecast: $45.9bn, +5.3%; 977m PC players.
2025 PC growth was revised to 12%.
Full Newzoo report was scheduled for September 10, after this research cutoff.
Use it forUse PC-specific and region-specific measures; never treat global games spending as the addressable market for a Steam niche.
Do not concludeForecast, not final 2026 actuals. Covers software/services beyond Steam and excludes tax, hardware, advertising and gambling.
Aggregator linking publisher disclosures and reporting
A useful index of original milestone announcements across platforms. It puts fast breakout stories beside Arco's much slower recovery, and distinguishes free player acquisition from paid unit sales.
NUMBERS, WITH THEIR DEFINITIONS
Machine Party: >1m copies in two weeks, linked to its Steam announcement.
ReStory: >300k copies in a week and $3m gross by August 10, linked to Steam.
Arco: >100k worldwide copies across platforms and development-cost breakeven after two years, confirmed to Game Developer on August 18, 2026. Available on Steam, Nintendo Switch, Mac App Store and Epic Games Store; most sales came through Steam, but no exact Steam unit count was disclosed.
Moonlighter: >7m free additions in five days; these are not paid sales.
Use it forUse this as a lead finder, then open the original announcement and record paid/free, platform scope and elapsed time.
Do not concludeDifferent underlying evidence types; these figures are not independently audited by the newsletter. Milestones overrepresent winners. Arco's cross-platform milestone must not be labeled Steam-only.
Two very different releases show the need to separate audience fit, pricing and demo evidence. How to Fish combined inexpensive co-op with familiar social audiences; Mortal Shell 2 attracted many players new to the original.
NUMBERS, WITH THEIR DEFINITIONS
Alinea estimate: How to Fish 1.9m units, approaching $8m revenue; launch price $4.95.
How to Fish overlap: PEAK 53%, Content Warning 27%.
Mortal Shell 2: estimated 510k multi-platform units and $23m revenue; 57.6% of units on Steam.
Approximately 600k Steam demo players; 9% of Steam buyers had played the original.
Use it forKeep platform and regional splits explicit; investigate demo engagement and relevant audience overlap instead of relying on franchise familiarity.
Do not concludeNo causal proof that demos drove purchases. Landfall financially supported How to Fish; it is not an unsupported zero-budget comparator.
Practitioner analysis of public trailers; historical view claims
Lieu contrasts playful, player-controlled incidents with staged comedy, slow framing, and visibly rough capture. He finds Dear Passengers appealing as a space for friends to experiment, while questioning whether its montage communicates a coherent loop. The comparison is a craft critique, not an explanation of the campaigns' different reach.
NUMBERS, WITH THEIR DEFINITIONS
At publication, Lieu reported more than 2 million trailer views for Dear Passengers versus roughly 5,000 across multiple channels for Banana Airways; these are not matched campaign measurements.
Use it forShow genuine player agency and readable comic consequences; capture quality remains important even when the tone is deliberately chaotic.
Do not concludeLieu explicitly says he does not know what caused the view gap. Promotion, distribution, art direction, and game content differ. Neither organic attribution nor paid-sales conversion is established.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
by day 5, capture one playable choice-to-comic-consequence sequence in the actual prototype; label any scripted demonstration honestly and avoid building extra features solely to fill a montage.
by day 12, have the marketing manager cut two openings from the same captured session, test comprehension with fresh viewers, and log distribution separately from creative differences.
Big Walk expands social co-op beyond chaos into a shared journey of communication and discovery. A strong reveal and later gameplay showcase generated interest without a public demo.
NUMBERS, WITH THEIR DEFINITIONS
Panic confirmed 1m paid copies across platforms in approximately six days.
Steam peak cited: 46k CCU; estimated launch wishlists: almost 1m.
2023 Game Awards reveal rapidly passed 100k estimated wishlists.
GDCo Steam overlap estimates: PEAK 75%, Lethal Company 67%, Untitled Goose Game 6%.
Use it forDesign moments that require friends to talk; target the social co-op audience rather than assuming the studio's previous audience transfers.
Do not concludeDevelopment began around 2020; established developer and publisher. Multi-platform sales and subscription players are separate measures.
A first-party timeline contrasts an expensive, long project with a fast-to-understand, replayable game. It exposes experiments and actual wishlist history rather than retroactively reducing success to a single viral moment.
NUMBERS, WITH THEIR DEFINITIONS
July 30, 2026 launch, exactly one year after the idea.
173k Steam units sold in the first week.
The studio had fourteen years of prior experience.
A downloadable annotated wishlist CSV accompanies the article.
Use it forKeep your own dated experiment log: hook, asset, channel, cost, baseline, observed change, and alternative explanations.
Do not concludeOne successful studio case, not a typical outcome. Accumulated experience and pipeline matter; not a one-year beginner benchmark.
Tactile controls, mechanical sound and a confined artillery fantasy made raw clips compelling. Mission variety reused one environment. Short video created early attention; playtests, longer videos and the demo broadened the funnel.
NUMBERS, WITH THEIR DEFINITIONS
Two developers; just over one year of development.
Developer data as of day five: >365k copies, $4.7m gross, 4% refunds.
>400k wishlists before the demo; >700k Next Fest demo players.
Demo later reached 876k players with 2h19 median playtime.
Use it forConcentrate art and audio on one satisfying machine, then create missions by recombining procedures instead of building more maps.
Do not concludeNot a three-week completed-game comparator. Inspired by PVKK; the article discusses differentiation and shipping-speed concerns.
A large turn-based-combat cohort shows a severe mean/median gap. The author studies hybrid loops, presentation, pricing and co-op, but the filtered observational sample cannot identify which design choice causes success.
NUMBERS, WITH THEIR DEFINITIONS
1,542 games released since January 2023; snapshot August 13, 2026.
Estimated median gross revenue $7,744 versus average $776,940.
Top 1% took 69.4% of revenue; top 10% took 95.8%.
Games under $500 estimated revenue and NSFW/unreleased titles were excluded.
Use it forLook at median and downside cases before adding a strategy genre to the shortlist; identify a second appealing loop and production cost.
Do not concludeCo-op, price and art correlations are confounded by budget and team quality. Exclusions make the apparent success rate optimistic.
Carless proposes evaluating a streaming showcase before committing scarce marketing resources. The key questions concern how many games compete for attention, whether creators can co-stream, how much of the lineup contains new reveals, participation costs and total runtime. GameDiscoverCo compares shows through estimated wishlist changes, but explicitly notes omissions for games without Steam pages, released titles and console wishlists. The article recommends matching the game’s character to the event audience rather than chasing the highest headline ranking. For small teams, this becomes a practical briefing document for evaluating a reveal opportunity and its production cost.
NUMBERS, WITH THEIR DEFINITIONS
GDCo tracked 385 showcases overall, including 68 in its leading-event group since July 2024.
Typical lineups cited: 35–40 games at Triple-I versus roughly 60–70 at some large shows.
Tears of Metal: estimated >150k units since July 22; demo ranked #9 of 2,966 in October 2025 Next Fest.
Use it forAudience fit matters alongside reach. Co-streaming can extend distribution. Estimated wishlist rankings have coverage gaps.
Do not concludeBefore-and-after wishlist estimates do not isolate an event’s effect or guarantee return on spending.
Apply in a small-team experiment · 4 steps
Our proposed experiment—not a result measured by this article.
Solo: Request cost, runtime, lineup size and co-streaming details.
Solo: Check whether the audience already likes comparable games.
2–5 people: Budget trailer work alongside placement fees.
2–5 people: Prepare a live store page and measurement plan.
Zukowski reconstructs a history of third-party Steam festivals and finds highly uneven participation: some games repeatedly appear while many receive only a handful of placements. Cozy, narrative and building games are prominent among frequently featured titles. The article also presents a useful counterexample: Lost Twins 2 appeared in many festivals yet had comparatively few reviews when the author checked. This makes event access and commercial demand separate research questions. For a small team, the operational lesson is to maintain an application calendar and assess audience fit while measuring whether exposure leads to meaningful interest.
Use it forFestival selection has its own genre preferences. Repeated inclusion does not establish strong sales. Past participant lists help target applications.
Do not concludeIncomplete archives and selected participants prevent population-wide success rates; reviews are not verified sales.
Apply in a small-team experiment · 4 steps
Our proposed experiment—not a result measured by this article.
Solo: Shortlist five events matching the player fantasy.
Solo: Reuse a concise pitch and current demo assets.
2–5 people: Assign deadlines and post-event measurement owners.
2–5 people: Compare wishlist gains with the normal daily baseline.
Lieu explains how the Sheepherds trailer improved by giving the basic herding action room to communicate before introducing decoration and extra modes. The early version interrupted action with a title reveal, moved rapidly to customization and showed uniformly successful play, obscuring both the challenge and the game’s depth. The revised sequence demonstrates herding first, then coloring sheep, environmental complications and visible mistakes. This is useful for small teams because the improvement depends on footage selection and sequencing rather than adding features. A trailer can make a modest system understandable by showing its cause, result and variation.
NUMBERS, WITH THEIR DEFINITIONS
The revised opening demonstrates herding for 10 seconds.
Lieu reported over 100,000 IGN views for the revised trailer at publication; no controlled uplift or sales result is supplied.
Use it forDemonstrate the main action immediately. Show difficulty through authentic imperfect play. Introduce extras after the core is understood.
Do not concludeReported views and craft judgments do not establish that the revision increased sales.
Apply in a small-team experiment · 6 steps
Our proposed experiment—not a result measured by this article.
on day 4, the PM and artist storyboard three shots using one existing mechanic: successful action, complication, and visible recovery; defer customization assets until these shots are readable.
on day 10, show a silent rough cut to five unfamiliar viewers and ask what the player controls and what makes the task difficult; revise confusing footage before polishing titles.
Solo: Capture one uninterrupted action and its result.
Solo: Remove title cards that require insider knowledge.
2–5 people: Ask unfamiliar viewers to explain the objective.
2–5 people: Sequence basic action, complication and variation before cosmetics.
Demand still expanded across successive cohorts, but individual hits rapidly lost category share. Networking/support risk is not captured by the chart.
Use it forRead the original methodology and cross-check the lesson against the case library before using it as a rule.
Do not concludeCohorts and interviews are observational. Public correlation is not Steam algorithm documentation or channel attribution.
A PvE autochess/roguelike hybrid found players asking for that exact combination. Months of weekly balance iteration and specialist creator coverage preceded a same-day reveal and demo launch. Short-form was not a major driver.
NUMBERS, WITH THEIR DEFINITIONS
>275k demo players; 3h55 median playtime; 25% exceeded ten hours.
Approximately 230k wishlists by July 27; demo peak 14k CCU.
Discord grew from 250 early testers to 3k at demo launch.
Team included six full-time staff alongside founders, freelancers and outreach support.
Use it forSearch niche communities for an unmet mode request, then validate with specialists and balance telemetry before mass promotion.
Do not concludePaid creators and ads were also used. Strong craft, experience and funding; not evidence that all acquisition was organic.
Analyst estimates and reported campaign information
A failing-airline co-op fantasy generated substantial pre-release demand through short social clips. The analyst connects low price friction and audience circulation among co-op games, while warning about incoming clones.
NUMBERS, WITH THEIR DEFINITIONS
Article reports >1m Dear Passengers wishlists and no paid marketing, creators or ads to that point.
Alinea estimates: 48% of wishlisters own R.E.P.O.; 47% PEAK.
Estimated R.E.P.O. prelaunch-wishlist conversion: 38% by day seven, 64% by day thirty.
Developer Flexus is a roughly 70-person studio, not a tiny new team.
Use it forTest a shared everyday role under absurd pressure; evaluate group purchase friction and regional pricing before choosing a price.
Do not concludePre-release demand is not sales. Selected successful co-op conversions do not establish a universal rate; studio resources matter.
Practitioner analysis using release intervals as a proxy
Cole argues for smaller projects using a Steam release-history analysis. The useful methodological question is whether time between a studio's releases measures development time: overlapping projects, ports and different teams can break that assumption.
Use it forUse actual team schedules and staffing when estimating scope; treat release intervals as a research lead, not a verified production duration.
Do not concludeThe cohort excludes low-review games and first-project durations are unknown. A single observed release does not establish that a developer quit. The article's review-threshold claim is not adopted here as an official Steam rule.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Record development start, team changes and overlapping projects where disclosed.
Preserve the article's review-count and release-date filters when discussing its sample.
High base + high momentum: 28/34 reached 3K+ festival wishlists; low + low: 3/47. Momentum correlated slightly better than starting size; self-report caveats apply.
Use it forRead the original methodology and cross-check the lesson against the case library before using it as a rule.
Do not concludeCohorts and interviews are observational. Public correlation is not Steam algorithm documentation or channel attribution.
In a self-reported n=119 sample, games entering above 1K wishlists still performed better; the evidence does not support “slop made discovery impossible.”
Use it forRead the original methodology and cross-check the lesson against the case library before using it as a rule.
Do not concludeCohorts and interviews are observational. Public correlation is not Steam algorithm documentation or channel attribution.
Follower-derived estimates and observational cohort analysis
A broad June Next Fest scrape finds pre-existing interest and real demo engagement associated with better results. Starting wishlist totals were more correlated with gains than recent velocity in this dataset.
NUMBERS, WITH THEIR DEFINITIONS
Approximately 279 games lacked wishlist estimates and were excluded.
80% of demos did not reach ten concurrent players.
Estimated median wishlist gain rose from 224 in the 1–9 CCU bucket to 1,284 in the next bucket.
The 2k–5k starting-wishlist cohort was the first with median gains above 1k.
Use it forRelease the demo early enough to fix feedback; monitor natural engagement, not manufactured concurrent sessions.
Do not concludeWishlist proxy was adjusted during the event; pre-event CCU polling was incomplete. Correlation is not an algorithm threshold.
The piece argues Next Fest amplifies demand rather than creating it. Bombanana's legible asymmetric co-op premise snowballed through creators; Nightwater used an incentivized Bongo Cat collaboration, a distinctly different acquisition route.
NUMBERS, WITH THEIR DEFINITIONS
Bombanana page launched 25 days before Next Fest after approximately five months' development.
A creator tweet reached approximately 650k views; 60% of demo players exceeded one hour.
Over the Hill: estimated >400k demo players, roughly 100k event wishlists.
Nightwater: approximately 75k event wishlists; 73% of wishlisters played Bongo Cat that month.
Use it forLabel incentive-driven wishlists separately. Prepare an independently attractive demo and creator campaign rather than relying on festival inclusion.
Do not concludeWishlist conversion uses different cohorts and time windows. Strong claims about event causality are interpretation, not controlled evidence.
The relevant competitive set is narrower than the total festival catalog. Filtering by genre and evidence of existing interest reveals a more useful comparison group, while weak demand in a small genre remains a warning.
NUMBERS, WITH THEIR DEFINITIONS
4,350 participating games at the initial snapshot; 175 excluded for inaccessible follower data.
1,241 of 4,175 had <100 estimated wishlists; 2,146 had 100–2k; 788 exceeded 2k.
Point-and-click cohort: 272 games, only 27 above 2k estimated wishlists.
General wishlist proxy used: followers ×12.
Use it forBenchmark against comparable games with a real audience, while retaining the low-performing cohort to understand risk.
Do not concludeCalling only high-interest games competition is an interpretation. Proxy and age-gate exclusions affect counts; tags are non-exclusive.
The first two days are noisy; the recommendation system starts sorting around Wednesday. Ads/outreach begun after the festival starts are usually too late.
Use it forRead the original methodology and cross-check the lesson against the case library before using it as a rule.
Do not concludeCohorts and interviews are observational. Public correlation is not Steam algorithm documentation or channel attribution.
Observed platform change plus analyst interpretation
Early analysis of the June storefront redesign distinguishes the more selective Popular Upcoming widget from personalized calendar discovery. Historical wishlist rules of thumb no longer describe the same placement.
NUMBERS, WITH THEIR DEFINITIONS
Homepage beta began April 1; public rollout cited as June 4, 2026.
Article reports the Personal Calendar can surface games approximately two months before launch.
Use it forUse current Steamworks documentation and actual traffic-source reports; do not plan launch around an old 7k-wishlist promise.
Do not concludeEarly observations, not Valve's ranking specification. Reported entry ranges are not fixed or guaranteed thresholds.
Irene Borodina, Digital Vortex CMO; page byline Alexander Semenov · Outside original priority window · Checked
Research note by Steam Discovery Research; original article by Irene Borodina, Digital Vortex CMO; page byline Alexander Semenov.
Publisher self-reported campaign column
The publisher describes repeated wishlist spikes from playtesting, a showcase, creator outreach and vertical video, with paid traffic also used.
NUMBERS, WITH THEIR DEFINITIONS
200,000+ Steam wishlists reported May 21, 2026; approximately 70,000 when the publisher joined in autumn 2025.
February 2026: >44,000 additions in one month during IGN Fan Fest plus simultaneous outreach, social and paid traffic.
April 2026: >25,000 additions in the IndieGameJoe posting week; the publisher says those posts were unpaid.
Use it forPackage each real build milestone for several audience-matched channels.
Do not concludeNot an attribution audit. 'Tripled' is approximate, not exact accounting. February IGN Fan Fest precedes claimed first IGN contact in March; sequence unresolved. Total spend unknown.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Days 1–5: the PM and marketer select one demonstrable hook; the two developers make it reliable and the artist supplies a readable before/after state. Capture three short variations from real play.
Days 6–15: the marketer targets creators who recently covered comparable mechanics; the PM records posting dates, any agreed cost, qualified visits and wishlist changes separately. Do not promise a viral result or buy coverage without a budget decision.
A coordinated reveal combined a PAX East presentation, press embargo, creators, trailer and Kickstarter. The lesson is orchestration, not that unknown teams should hide their work for years.
NUMBERS, WITH THEIR DEFINITIONS
March 27 reveal; >200k wishlists and >4k Discord members within one month.
Kickstarter raised €272,214 from 5,869 backers against a €150k goal.
Studio was founded in 2020 with industry veterans; article cites >$7.9m funding.
Closed backer beta began May 15.
Use it forCoordinate genuinely complementary announcement assets and embargoes; only use a late reveal when a strong playable and distribution relationships already exist.
Do not concludeOutside priority window. Veteran, funded team is unlike a three-week micro-team. Campaign spending and attribution are not fully disclosed.
The initial cut communicated cooking and turn-based battles but not their purpose. Lieu's consultation clarified the sisters' conflict and the fast-food empire they oppose, linking cooking to restoring villages and advancing the quest. Revised dialogue and title cards gave existing footage a clearer objective without rebuilding the entire trailer.
Use it forConnect the advertised actions to a concrete goal; remove title cards whose meaning requires insider knowledge.
Do not concludeNo wishlist, retention, sales, or experiment result is reported. Lieu acknowledges his involvement and subjective judgment. The displayed date is April 13 despite April 12 in the URL.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
on day 2, the PM writes one sentence linking the demo's action to its goal, then assigns one existing gameplay shot to prove each half; cut unsupported promises from the brief.
on day 11, the marketing manager reviews every title card alongside its footage and removes cards that a cold viewer cannot explain; test whether viewers can state why the character takes the action.
Lieu describes planning Mandrake's reveal before full trailer production: select recognizable life-sim actions, pair them with the game's darker character, and prioritize the footage and art needed to communicate that combination.
Use it forPlan a reveal around a few playable genre anchors and one distinctive feeling; each narration line should have a clear visual purpose.
Do not concludeCraft guidance from one selected client project, not controlled evidence that editing caused reach or sales. Publication date follows the visible April 4 byline; the URL contains April 3.
Apply in a small-team experiment · 3 steps
Our proposed experiment—not a result measured by this article.
List the familiar actions that identify the genre.
Capture those actions with the game's distinctive tone visible.
Test whether a viewer can describe both the activity and its twist.
Steam/Gamalytic estimates and qualitative positioning analysis
A narrow productivity cohort separates focus companions, customization-driven work rewards, social accountability and parasocial companionship. Marketing sells the feeling of getting work done, not just a timer and a feature list.
NUMBERS, WITH THEIR DEFINITIONS
34 selected games; data collected February 5–6, 2026.
Nine had >500 reviews; three had estimated gross revenue >$1m.
Virtual Cottage 2 had 58,026 estimated wishlists.
Pure idle and bottom-of-screen games were excluded.
Use it forDefine a specific work ritual and audience; test cozy lifestyle creators, customization sharing and relevant cross-promotions.
Do not concludeOutside priority window; small handpicked cohort and estimated numbers. Review counts alone do not establish profitability.
Spirit City: Lofi SessionsVirtual Cottage 2On-Together: Virtual Co-WorkingChill with You: Lo-Fi Story
An event recap turns publisher feedback into practical pitch improvements: put the game's identity, working footage, schedule and financial needs early, then make the build easy for different decision-makers to assess.
Use it forPrepare a concise fact sheet and short gameplay demonstration that explain the player promise and the concrete support being requested.
Do not concludeSelected panel feedback from 2025, not a universal publisher policy or a measured chance of securing a deal. No acceptance-rate or sales-conversion claim is adopted.
Apply in a small-team experiment · 3 steps
Our proposed experiment—not a result measured by this article.
Show actual footage alongside the premise.
State the budget, schedule, current evidence and remaining needs.
Adapt the pitch to the publisher's relevant portfolio.
Event-organizer recap of practitioner panel; secondary case claims
This recap of the IMPACT! 2024 panel highlights game-specific humor in Another Crab’s Treasure, entertaining patch communication in Content Warning, and creator outreach for The Ouroboros King.
NUMBERS, WITH THEIR DEFINITIONS
The panel recap reports The Ouroboros King moving from 500 to 20,000 wishlists. It describes a two-month outreach phase and a two-week surge; the exact timeline is ambiguous and raw data is not supplied.
Use it forMake the game's actual systems the material for its jokes and creator pitch; reserve developer time for useful capture.
Do not concludeThese are selected successes, not a controlled campaign comparison. Spend and attribution are incomplete; the wishlist claim is not independently audited. Platform views, followers and wishlists are different measures.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Capture a genuine setup, player decision and consequence from a working build. Do not imply systems the game lacks.
Turn one real playtest incident into a short clip and readable update. Log the channel, date, spending and wishlist changes without treating coincident spikes as proof of attribution.
Game If You Are; displayed byline: admin · Outside original priority window · Checked
Research note by Steam Discovery Research; original article by Game If You Are; displayed byline: admin.
Agency practitioner guidance, not a measured market study
Practitioner guidance argues for purposeful channel selection: a distinct product promise, compelling footage, relevant creators, planned events and disciplined advertising tests.
Use it forChoose a channel because its audience and format fit the demonstrable game.
Do not concludeNo campaign sample or underlying dataset is provided. Illustrative $1 cost-per-wishlist and creator-price examples are not benchmarks. Its three-to-four-week prerelease review-code advice is not a same-sprint review promise; plan previews. These are dated 2025 opinions, not verified 2026 platform guarantees.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Days 1–3: the PM and marketer write one target-player promise and select one primary channel, with a measurable action such as a qualified Steam visit. Ask the two developers and artist which promised interaction can be shown honestly by day five.
Days 5–15: capture a short playable preview and pitch only relevant preview-friendly creators or press. Keep ads optional and separately budgeted; the PM protects build time from broad event travel or an untestable multi-channel campaign.
Vendor practitioner guidance; no measured campaign result
A useful outreach checklist: make the game and the request clear, supply usable assets and access instructions, and approach creators in relevant batches. Different key-distribution approaches suit trusted contacts and unfamiliar recipients.
Use it forMake an interested creator's next step easy: understand the hook, inspect the footage, and access a reliable build.
Do not concludeNo measured conversion or campaign outcome is supplied. Suggested timing is a heuristic. The text mixes words and characters in its email-length guidance; those numerical limits are not reproduced here.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Days 1–3: the marketer selects ten creators who recently covered the closest three comparable games. Record why their viewers would enjoy our playable hook, not just follower counts.
Days 6–12: the PM and developers prepare a stable ten-minute build, an honest 15-second gameplay clip and a small press kit. Track replies, published coverage and measured Steam visits separately.
This report follows three developers who turned a spare-time plant-decorating game into a full-time project. Urban Jungle combined a readily understood domestic fantasy with repeated demo updates, feedback, relevant showcases and unexpected social sharing. The team had mobile-game experience, and the account distinguishes wishlists accumulated before signing a publisher from later support. That context matters when assessing what another three-person team can reproduce. The repeatable work is building a legible game, improving its demo and finding compatible audiences. The viral Thai Facebook post is a reminder that individual discovery events remain difficult to schedule.
Use it forPrior experience and publisher support affect comparisons. Repeated demos create feedback opportunities. Unexpected sharing is not a reliable plan.
Do not concludeDeveloper recollection does not isolate channel effects or establish final sales, profit or repeatable virality.
Apply in a small-team experiment · 4 steps
Our proposed experiment—not a result measured by this article.
Solo: Choose a fantasy visible in one screenshot.
Solo: Test a small playable loop before expanding.
2–5 people: Divide demo improvement, capture and applications.
2–5 people: Log pre-publisher and post-publisher campaign results separately.
Ryan K. Rigney, interview with Antonio Uribe · Outside original priority window
Research note by Steam Discovery Research; original article by Ryan K. Rigney, interview with Antonio Uribe.
Direct developer interview; historical snapshot
An interview with Arco's co-creator demonstrates that strong reviews and a respected publisher do not guarantee sufficient launch revenue. The marketer also describes initially misreading its visual presentation as a slow narrative game.
NUMBERS, WITH THEIR DEFINITIONS
Arco had 98% positive Steam reviews at the article's snapshot.
The developer said first-month income did not fund six months of team work.
Interview describes 197 main encounters plus sixteen arena encounters.
Use it forCheck whether screenshots and the first seconds of a trailer communicate the actual play experience; include admired commercial underperformers in comparisons.
Do not concludeHistorical counter-case, not 2026 market data. Playtime-share figures in the original concern 2023 and are not used as current estimates.
First-person studio postmortem; historical internal observations
Historical account by Shiny Shoe's CEO: integrated feedback and technical testing helped iteration, but free testers did not anticipate paying customers' objections to cosmetic monetization. Supporting solo and co-op created conflicting design demands. English-only Early Access simplified iteration while leaving a significant localization problem.
NUMBERS, WITH THEIR DEFINITIONS
Early Access began in May 2023; a few hundred people tested before release.
Approximately 60% of runs were solo and 40% co-op; these are runs, not players.
Cooke reports missing Chinese support as the leading Early Access refund reason; no count or rate is provided.
Use it forChoose the supported play mode deliberately and collect contextual feedback; free playtests cannot validate willingness to pay.
Do not concludeNo sales lift, refund denominator, or lost-localization revenue is quantified. This established studio's online infrastructure is not evidence of 15-day feasibility.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
on day 1, select one primary demo play mode and write its acceptance criteria; exclude a second mode, matchmaking hub, and live-service monetization unless they are essential to the tested hook.
by day 6, add a simple feedback link with an optional build ID and consented reproduction details; have the PM separate repeated gameplay issues, language requests, and price expectations rather than merging them into one approval score.
Serg Kharchenko, Evangelist at CM Games; guest writer · Outside original priority window · Checked
Research note by Steam Discovery Research; original article by Serg Kharchenko, Evangelist at CM Games; guest writer.
Developer-authored historical postmortem
CM Games assembled a genre-matched sale, supplied shared promotional assets and coordinated participating communities. Steam placement amplified the event.
NUMBERS, WITH THEIR DEFINITIONS
Reported event: >20 participants and 96 hours of Steam-homepage featuring.
Reported exposure: >11 million homepage impressions and 2 million sale-event impressions.
Steam developer-page subscribers reportedly rose from 100 to >3,000.
Use it forMake collaboration easy with relevant partners and ready-to-use assets.
Do not concludePublished in 2024; planning began in 2022, but exact sale dates are absent. Impression surfaces are not unique audiences and should not be summed as reach. Discounts, platform featuring and promotion were mixed; costs and incremental revenue are undisclosed. A new demo cannot assume equivalent placement.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Days 1–5: the marketer lists eight closely related games and existing community events; the PM screens audience fit and dates. Do not spend the sprint organizing a multi-studio Steam sale.
Days 6–15: the developers freeze a stable demo path while the artist and marketer prepare one small shared-asset kit. Approach two plausible partners about future cross-promotion; the PM keeps this separate from the demo's delivery-critical work.
Historical February 2024 case: a spare-time solo developer refined an introductory demo but left marketing and media preparation late. Feedback exposed compressed GIFs that misrepresented the game's movement. The result illustrates useful learning and modest traction, not automatic festival success.
Reported wishlist endpoints: 420 to 755. Demo-play endpoints: 194 to 691. Median playtime: 6 minutes for an 8–10-minute demo.
Use it forPrepare onboarding, promotional assets, and a measurement plan before the event begins.
Do not concludeWishlist endpoints imply +335, not the claimed +355; a gross/net distinction is unverified. Do not reuse the claimed conversion rates or infer sales. Channel attribution is unproven; historical Steam procedures are not current guidance.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
by day 8, have the PM run cold tests of the opening and the artist review the actual compressed promotional files; address onboarding failures and misleading motion before capturing final assets.
by day 12, prepare the announcement, short clips, and outreach list; maintain separate fields for impressions, visits, gross wishlist additions, removals, net totals, demo starts, time windows, and source definitions.
Historical proxy compilation; definition checked against original survey
A historical catalogue of Steam estimation shortcuts—and a useful example of why metric definitions must be checked against the original research before forecasting.
NUMBERS, WITH THEIR DEFINITIONS
Definition correction: the cited 2020 survey's 0.2 median is total first-week units divided by prelaunch wishlists. It is not a measured 20% conversion of wishlisters into buyers.
The original GameDiscoverCo survey received 50 responses before outlier exclusions; it asked actual wishlist conversion and total first-week sales as separate questions.
Use it forSeparate wishlist-cohort conversion from all sales relative to a wishlist baseline. Label estimates with model, cohort and year.
Do not concludeThe guide conflates those measures. Its historical ratios are not verified 2026 benchmarks; the small, selected survey mixes release contexts. No universal revenue or profit forecast follows.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Day 1: the PM defines separate fields for wishlist additions, removals, net wishlists, wishlist-cohort purchases and total units. Record the time window and source beside each number.
Days 10–15: the marketer reports measured prototype signals, not projected copies from a universal multiplier. Use weak, base and strong scenarios only as labeled planning assumptions.
Historical developer report with analyst interpretation
Power of Ten entered Early Access with roughly twelve thousand wishlists but its developer reported only about three hundred launch-day sales. Carless examines where interest came from: quiet periods produced little daily growth, while festival exposure contributed a large share of the total. The account also includes a player who wished to wait for the finished game. Its value is as a counterexample to treating accumulated wishlists as guaranteed immediate buyers. Teams can use the case to examine baseline demand, campaign cohorts and release expectations while retaining uncertainty about why any particular person did not purchase.
Use it forWishlist totals do not guarantee launch purchases. Campaign spikes need contextual interpretation. Early Access changes some buyers’ timing.
Do not concludeA 2021 anecdote is not a current conversion benchmark or proof that festival wishlists are worthless.
Apply in a small-team experiment · 4 steps
Our proposed experiment—not a result measured by this article.
Solo: Track normal daily growth between promotional spikes.
Solo: Ask target players whether they buy Early Access.
2–5 people: Prepare low, middle and high launch scenarios.
2–5 people: Separate reported conversion from inferred causes.
Game If You Are; individual author not named · Publication date not displayed · Checked
Research note by Steam Discovery Research; original article by Game If You Are; individual author not named.
Agency self-reported client campaign
An urgent Next Fest campaign selected cozy-game creators and prepared rates, briefs, contracts and assets before rapid outreach.
NUMBERS, WITH THEIR DEFINITIONS
Fifteen creators approached; four produced five sponsored content pieces.
Almost 25,000 combined YouTube/Twitch views.
Over 4,000 wishlist additions by week-end, alongside Steam Next Fest promotion.
Use it forAudience fit and operational readiness matter more than a long generic creator list.
Do not concludePublication date and event edition are not displayed. Total budget is undisclosed. The wishlist result combines sponsored content and festival promotion, not isolated creator lift; views are not unique people. Do not calculate cost per wishlist or infer sales.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Days 1–3: the marketer builds a fifteen-creator shortlist from recent comparable-game coverage; the PM prepares the brief, response tracker and a clear unpaid-versus-sponsored distinction. Any paid work needs an explicit budget.
Days 4–10: the developers stabilize one short demo route and the artist delivers approved screenshots. The marketer sends a clear access window and honest build limitations; the PM tracks replies, published pieces and measured traffic without assuming a Next Fest falls inside the sprint.
Game If You Are; individual author not named · Publication date not displayed · Checked
Research note by Steam Discovery Research; original article by Game If You Are; individual author not named.
Agency self-reported positioning and funding case
After a failed Kickstarter, the agency repositioned Hauntii beyond its twin-stick-shooter label around atmosphere, audiovisual identity and possession mechanics, then supported publisher pitching.
NUMBERS, WITH THEIR DEFINITIONS
Three audience personas developed.
Agency reports a Firestoke publishing agreement and the requested funding/support; amounts are undisclosed.
Use it forLead with the distinctive player experience, then support it with a credible production plan.
Do not concludeUndated case; no quantified wishlist, sales or reach results. Funding is not revenue or proof of profitable demand. Organic activity described by the agency does not mean zero total cost or establish causal campaign impact.
Apply in a small-team experiment · 2 steps
Our proposed experiment—not a result measured by this article.
Days 1–3: the PM and marketer write two alternative descriptions of the same prototype: genre-led and signature-interaction-led. Ask five target players what experience each promises; let the developers and artist flag anything the slice cannot support.
Days 10–15: use the proven demo loop, captured footage and a realistic role-by-role cost/scope sheet to build a concise pitch. The PM separates the current slice from future production; the marketer records evidence without inventing forecast returns.
Valve’s onboarding documentation makes the release calendar a production dependency. Developers must complete identity, banking and tax information, pay the per-product fee, prepare a store page and submit the game for review. For the first few titles, release requires at least 30 days after the app fee and a public Coming Soon page for at least two weeks. Tax verification can add waiting time. The separate review guide gives the more useful scheduling detail: store and build reviews typically take three to five business days, with at least seven business days recommended for each submission.
Use it forPaying the fee starts one release clock. The public page starts another. Review approval needs scheduling allowance.
Do not concludeMinimum eligibility does not guarantee timely approval, audience growth or commercial readiness.
Apply in a small-team experiment · 4 steps
Our proposed experiment—not a result measured by this article.
Solo: Record fee date and earliest eligible release.
Solo: Prepare the store page before announcing a fixed launch.
2–5 people: Assign one owner to onboarding and reviewer feedback.
2–5 people: Reserve revision time before public deadlines.
Valve describes visibility as a response to player interest, especially purchasing and playing, with tags and supported languages helping match games to people. Full releases receive baseline exposure through the New Releases queue, while prominent placements depend on response and individual preferences. The documentation separates useful business metrics from ranking inputs: page traffic and store conversion rate are not direct visibility factors, and wishlists mostly are not either, apart from exceptions such as Popular Upcoming. Wishlists still matter for notifications. This distinction helps small teams diagnose weak demand without inventing an algorithmic reward for every marketing metric.
Use it forPurchases and play indicate customer interest. Tags and languages affect matching. Traffic alone is not a ranking input.
Do not concludeValve does not disclose exact ranking weights or guarantee placement for a numerical target.
Apply in a small-team experiment · 4 steps
Our proposed experiment—not a result measured by this article.
Solo: Audit tags against actual player activities.
Solo: Compare traffic sources with purchases and play.
2–5 people: Separate campaign metrics from platform ranking claims.
2–5 people: Review store traffic breakdown after each release beat.
The Steam Content Survey must be completed before review and covers general, mature and generative AI content. Its current AI wording focuses on content shipped to and consumed by players, including artwork, audio, narrative and localization. General efficiency gains from development tools are not the stated focus. Developers must describe relevant pre-generated content; live generation additionally requires an explanation of safeguards against illegal outputs. Valve compares survey answers with the build and store page. Disclosure is therefore an ongoing content inventory task for a small studio, and completing the form does not override Steam’s other content rules.
Use it forInventory player-facing generated content. Distinguish pre-generated from live-generated output. Keep survey, build and marketing consistent.
Do not concludeDisclosure does not establish asset rights, player acceptance, or guaranteed review approval.
Apply in a small-team experiment · 4 steps
Our proposed experiment—not a result measured by this article.
Solo: List AI-assisted assets and localization before submission.
Solo: Describe live-generation safeguards if applicable.
2–5 people: Ask every contributor to identify relevant shipped content.
2–5 people: Assign ownership for disclosure changes after approval.
Gamalytic explains that its estimates combine Steam seller rankings, concurrent players, adjusted review-based modeling and publicly disclosed sales. Its glossary distinguishes owners, players and estimated copies purchased directly on Steam; these quantities are not interchangeable. Gross revenue for paid games covers base-game sales before Steam’s share, with regional prices and discounts considered. The provider warns that smaller games and free-to-play titles are harder to estimate, bundles may be mishandled and bugs can create large errors. These definitions make the page essential reading before turning comparison-game statistics into a small studio’s revenue assumptions.
Use it forOwners and Steam purchases are different metrics. Gross revenue is not developer take-home income. Small-game estimates need extra skepticism.
Do not concludeProvider accuracy claims are self-reported; this page does not validate an individual estimate or reveal development costs.
Apply in a small-team experiment · 4 steps
Our proposed experiment—not a result measured by this article.
Solo: Record metric definition, provider and retrieval date.
Solo: Use ranges when planning around estimated outcomes.
2–5 people: Separate reported sales from modeled figures.
2–5 people: Sanity-check comparables against disclosed results and release history.
DO NOT TURN OPINION INTO AN ALGORITHM RULEWhere good sources disagree
Wishlist stock versus recent momentum
Game Marketing Intel's broad proxy-based June sample finds starting wishlist stock more correlated with gains; Chris Zukowski's July 14 self-reported sample finds momentum modestly stronger.
Decision rule
Different samples and estimation methods can produce both findings. Track both; neither proves Steam's ranking formula. Do not delay all marketing to manufacture a last-minute spike.
Alinea's June article is skeptical about wishlist-to-sale conversion, yet its July co-op analysis shows strong conversion for selected inexpensive games. Other practitioners use wishlists as intent benchmarks.
Decision rule
Keep cohort, price, acquisition source and conversion window. Intent matters, but a wishlist is not a purchase and no universal conversion factor is justified.
Price is part of positioning and production scope. Cross-genre correlations are not a pricing experiment; use relevant comparables and a viable budget.
Selected tavern and productivity cohorts look encouraging; the broader turn-based cohort shows extreme concentration.
Decision rule
Retain denominator and exclusions. Use curated success cases to generate hypotheses, then add weak launches, unreleased supply and production costs before committing.
Trailer critiques diagnose communication; campaign accounts describe distribution. Neither establishes that an edit caused a particular reach or sales result.
Decision rule
Test whether cold viewers understand the mechanic, then measure distribution separately. For our demo, keep the footage consistent when comparing channels.
The DREADMOOR and Design and Conjure accounts combine several activities. Unpaid individual posts can coexist with paid work elsewhere in the campaign.
Decision rule
Maintain a channel-by-date log with explicit paid, unpaid and unknown labels. Without spend and attributable outcomes, do not report a cost per wishlist or organic-only success.
One historical guide blurs wishlist conversion with total sales per wishlist; a developer postmortem contains inconsistent totals. A precise-looking number can still have an unclear definition.
Decision rule
Keep original endpoints and definitions beside every derived number. Show unresolved contradictions; do not smooth them into a forecast or reuse disputed conversion rates.
Game-by-game launch timelines, creator and showcase discovery, player overlap, and estimated wishlist trajectories; frequent direct developer data.
Mixes disclosed numbers with proprietary estimates. Hit-focused case selection. Some charts and full analyses require Plus or Pro; newsletter access is not a dataset license.
Free articles and public lead sections; paid newsletter and Pro platform also offeredStart with GameDiscoverCo · 13 notes ↗Core · Independent practitioner analysis
Readable Next Fest cohort comparisons, follower-based wishlist estimates, small-game marketing tactics, and templates.
Author describes himself as a solo developer/pharmacist rather than a Valve insider. Wishlist proxies, incomplete polling and excluded age-gated titles can alter conclusions. Treat algorithm explanations as hypotheses.
Estimates are not publisher accounts. Overlap is not incremental attribution. Public articles show selected outputs rather than the complete model or dataset; rhetoric can be stronger than evidence.
Free newsletter; commercial analytics platform and client materialStart with Alinea · 6 notes ↗Core · Publisher/practitioner newsletter
Niche scans with downloadable cohorts and marketing mini-cases, especially taverns, productivity games and turn-based hybrids.
Often uses Gamalytic estimates and handpicked cohorts. Filtering out low-revenue or 'non-serious' games inflates apparent hit rates; publisher relationships are relevant context.
Finding original industry reports, regional market data and newly announced game milestones; a high-coverage research inbox.
Not independent confirmation of its upstream sources. Forecasts, publisher announcements and model estimates have different definitions. Follow original links before using a number in a decision.
Actual project chronology, experiments, wishlist CSVs, production constraints, unsuccessful channels and commercial results.
A 14-year studio's accumulated team and pipeline advantages do not transfer to a new team. First-party attribution remains observational. Irregular publication; not a market-wide panel.
Positioning, audience psychology, marketing craft and candid creator interviews, including commercial disappointments.
Irregular schedule; not a current Steam sales database. Historical market figures must retain their year. Author's current a16z marketing role and earlier live-service background are context.
Campaign timelines and first-person operational breakdowns, especially when the studio or publisher supplies the numbers.
Mixes original columns with secondary recaps and commercially interested contributors. Check bylines, original sources, dates, spend and metric definitions before treating a case as market-wide evidence.
Selected articles publicly readable in full; no paywall bypass used.Start with Game World Observer · 3 notes ↗Core · Game-marketing agency; client case studies and practitioner guidance
Agency team; individual case-study authors often unnamed
Concrete creator-outreach, positioning, asset preparation and publisher-pitch workflows.
Selected client successes are agency self-reports, not independent experiments. Case dates, spend and channel attribution may be absent; funding and reach are not sales.
Shot-level trailer critiques and first-person consultation breakdowns: communicating the playable hook, structuring footage, and making goals and consequences legible.
Craft judgment is not controlled attribution. Selected trailers and consulting cases do not provide genre demand denominators, reliable sales forecasts, or proof that editing caused reach.
The selected full articles and biography were publicly readable without login. Newsletter signup is available at https://www.derek-lieu.com/subscribe; paid consultation and a separate course are also advertised.Start with Derek Lieu · 4 notes ↗Core · Trade publication with developer-written postmortems and case studies
Individual developer contributors; verify each byline
First-person production trade-offs, campaign retrospectives, disappointing outcomes, and the operational context behind public numbers.
The wider publication also contains journalism and commentary; not every piece is first-party. Contributor data can contain inconsistencies, and older Steam procedures must not be treated as current rules.
Both selected full articles and the case-study directory were publicly readable without login. Newsletter signup is linked; access to other products or articles was not assessed.Start with Game Developer · 3 notes ↗Core · Practitioner guides and industry-panel recaps
Ashley Gwinnell, Alessandro Cossidente, Tammy Seno and contributors
Creator outreach, press-kit preparation and campaign examples; useful for turning a marketing plan into a repeatable workflow.
A commercial marketing-tool vendor, not an independent market census. Panel anecdotes and historical multipliers need original-source checks; advice is not a measured guarantee.