A source-backed SWOT from real reviews
Point CritMap at any Steam game and it reads the actual player reviews, then sorts what it finds into strengths, weaknesses, opportunities, and threats, grounded in what players wrote rather than the store page. The SWOT is free for any game. On a plan, click any point for Pro Insights: recommendations in priority order and the exact reviews behind them, so you can check the reasoning instead of trusting the summary. Deep Analysis reads a larger sample, and custom scope slices by date, language, playtime, or a single update.
Points with a insights button carry a full breakdown. Click one to open it, exactly as Pro users do:
- The game is praised as a strong and worthy sequel that expands on everything the original did well, with more variety in builds, characters, and strategies.
- Co-op multiplayer mode is a genuine highlight, giving players a fun new way to experience the Spire with friends.
- Replay value is extremely high, with diverse characters and builds keeping players engaged for hundreds of hours.
- Specific boss encounters, particularly Aeonglass, are widely considered unfair, frustrating, and lacking meaningful counter-play.
- Repeated balance patches have nerfed popular builds and made the game progressively harder, alienating casual players.
- RNG changes have made some runs feel maliciously unfair rather than randomly challenging, especially for non-meta players.
- Improving boss design, especially for universally disliked encounters like Aeonglass, could significantly boost player satisfaction.
- A clearer difficulty curve or casual-friendly mode could retain players who are being pushed out by aggressive balance changes.
- Mod support is already seen as a saving grace for some players, and expanding official mod tools could broaden the community.
- Ongoing balance controversies and community backlash against the lead developer risk eroding player goodwill and review scores.
- Casual players are being driven away by increasing difficulty and RNG changes, shrinking the accessible player base over time.
- The Early Access label means players are experiencing an unfinished product, and negative impressions formed now may persist even after full release.
Search any Steam game and read its full SWOT free. More on Pro Insights and Deep Analysis
Edges: what the average hides
Averages smooth over the reviews that disagree with each other, and that is often where your next feature is hiding. Edges pulls out the outliers, quoted in full, plus the themes one group of players loves while another group hates. Both sides come with real counts, so you can see how big each camp actually is before you act on it.
Loud Edges
Substantive minority opinions an averaged summary buries, from 100 reviews in this window. Most-actionable first; treat each as a lead, not a verdict.
accidental save-file deletion wiped 56 hours of progress
→ A recoverable or cloud-backed save system (or a confirmation prompt before overwrite/delete) could prevent permanent data loss that drives players to negative reviews after heavy investment.
I have long been tired of Slay the Spire 2. The checkpoints are brutal, gold in shops is often scarce, cards sometimes get clogged and won't come out, and enemy mechanics are annoying. This morning I saw a save slot on the left side showing my last save as night 9 of day 13, with 56 hours and 3 minutes played. I first assumed it was someone else's file since I often play while watching videos. Without thinking I deleted the entire save, and years of progress vanished. Everything resets to zero. How cruel fate can be.
Show original (schinese)
余厌尖塔贰久矣。关隘峻极，店中金常匮；牌偶壅滞不发，妖怪机谋扰人。今晨见左方存卷，记止十三夜九时，历五十六时三分。初疑为分，以吾戏时恒伴音影闲游。不意误删全档，经年功迹尽消，万物复始，天心何其薄哉。
daily challenge difficulty described as excessive and worst feature
→ If daily challenges are alienating otherwise positive players, a difficulty scaling or opt-out mechanism for that mode could keep the feature from being a net detractor.
Game is an upgrade from the first I like some of the upgrades but enough with the insanity on the daily challenges it is the worst one
poor performance on low-end hardware noted by an otherwise positive player
→ Players on low-end machines are enjoying the game but being let down by performance, a scalability or optimization pass could convert satisfied-but-frustrated players into strong advocates.
I love this game, it just tickles the my brain in the right way. Only certick is that it can run badly on low-end computers (my computer) other than that I give it five spires out of five
Divisive Themes
Aspects players are split on: one camp loves it, another hates it. This is the contested ground a summary flattens into “mixed.”
Counts are reviewers in the analyzed sample (100) who spoke to each theme, not your whole player base.
Developer (Anthony) balance decisions and responsiveness to feedback
3 praised · 6 criticizedPlayers who trust the developers' long-term vision defend the balance changes as part of a healthy early access process, while others feel the lead designer does not play or understand the game and makes changes that hurt specific characters without improvement.
Mostly disliked: the fans are the minority here; weigh changes against losing them.
A lot of people say the developers are telling players how to play the game, but aren't they all still paying for it and playing while complaining?
Show original
好多人说游戏开发商在教人玩游戏，问题是不都在买单边骂边玩么？
WTF stupid Anthony, do you even know how to make games? Have you ever played your own game?
Show original
tmd傻逼安东尼你tm会做游戏吗？是不是🐎死了，你玩过你做的游戏吗？
STS2 difficulty curve compared to the original Slay the Spire
3 praised · 3 criticizedVeteran players find STS2 more polished and less brutally punishing than the first game, while newcomers and casual players find it too difficult and confusing to enjoy.
Genuinely split: changing this risks the camp that loves it.
The skill curve is intense (though thankfully less brutally punishing compared to the first game), but it's one of the most rewarding challenges you can master.
It's too difficult and confusing...
Edges is a tab on every game's analysis page. More about the analysis surface
Review Workbench
A place to actually work through reviews, not just read a chart of them. Go one at a time, keep your place between sessions, and see which complaints your patches may already have covered.
Three negative reviews, three verdicts. Open "Why?" on a card to see the reasoning behind each patch match:
Refunded at first, bought again later. The tutorial dumps you into the storm with no explanation of the wiring system and our whole group bounced off. Heard patch 0.9 reworked the intro so we gave it another shot.
Love the game with friends but our 6 person base turns into a slideshow after a few hours. Storms drop to single digit FPS on a machine that runs much bigger games fine.
The host left and three hours of progress just vanished. Until saves work for everyone in the session this is impossible to recommend to a group.
- Work review by review, with read/unread and workflow state that persists, so a few hundred reviews is something you chip away at instead of restarting each visit.
- Browse by patch window. Attribution badges show whether each review landed before or after a given update.
- CritMap matches old complaints against your changelogs and tags each one likely addressed, possibly addressed, or still open. It reads your notes against the complaint, not your build, so it never claims an outright fix.
- Draft a reply in the reviewer's own language. Nothing posts automatically: you copy it out and send it yourself when you are ready.
- One click opens the review on Steam when you want to respond there.
The weekly digest: the radar comes to you
Opt in per game and get one email a week that tells you what moved. It diffs this week's SWOT against last week's and counts how many negative reviews look winnable, meaning a patch you have already shipped may address them. Skim it in a minute, then click through to the exact analysis behind it.
- Themes are tagged new, returning, trending, or "still on the radar" for the standing issues that will not go away, so you can read the movement at a glance.
- When a complaint goes quiet the digest says "no longer mentioned", not "resolved". It can only see that players stopped bringing it up, not what changed in your build.
- Switch to daily for a launch or patch week, then it expires on its own, so you never end up with a standing daily email you forgot to turn off. If the weekly is on you drop back to weekly; if not, digests simply stop.
- A quiet week gets a short digest, not padded filler. If nothing moved, the email is short and says so.
The digest is part of the Review Workbench. Explore the Workbench
The Action Report: what to fix first
The digest watches. The Action Report digs. You start one on a game you manage and an AI reads its negative reviews one at a time rather than sampling them, then hands back the issues ranked by what they are actually costing you, what is likely behind each one, and what to do first. It is the pass worth running a week after a launch or a big update, when there is finally enough new signal to be worth reading in full.
Dungeon Forge is a genuinely well-liked roguelite carried by one loop players adore, held back by two reliability problems rather than anything about its design. The forge-and-descend core and the combat feel are doing the heavy lifting behind your 74% positive rating, but a bug that wipes forged gear when the game crashes mid-run is now the loudest complaint, and it pushes your most-invested players to refund the moment it costs them a long run.
The encouraging part is that this is a recoverable position, not a structural one. Almost nothing players dislike is about the game you set out to make; it is about reliability at the edges (a crash-time inventory wipe, a GPU-specific hard crash) plus one endgame pacing complaint. Close the inventory-loss bug and you protect the exact crafters who would otherwise be your loudest advocates, with a realistic path back toward 81% positive.
If you only do three things
- 1
Ship the forge-inventory-loss fix as a hotfix, not a content patch
It is your highest-impact issue (impact 0.91), it is rising, and it hits your most-invested players. Every week it stays live it converts a finished run into a refund.
Could contribute to the projected recovery toward 81% positive by protecting the ~145 recoverable reviews already tied to it.
- 2
Scope and fix the GPU-family crash on dungeon generation
It is small in raw volume but heavily concentrated in Simplified Chinese players (about 60% of the affected reviews), so it is quietly capping your rating in one of your largest regions.
May recover a region you are underserving today and unlock word of mouth that the crash currently suppresses.
- 3
Re-frame the store page around the forge-and-descend loop
Your happiest reviewers describe a crafting-first game, while the store leads with combat, and the mismatch generates avoidable early refunds.
Could reduce expectation-driven negatives at the top of the funnel, where they are cheapest to prevent.
Patterns across issues
- The forge-inventory loss and the refund-mention spike are the same story from two angles: the players losing crafted gear are the ones threatening to refund. Fix the former and the latter fades.
- The GPU crash looks niche by raw count, but it lands almost entirely on Simplified Chinese players, so treating it as a small technical footnote understates where it is costing you rating.
- The tier 4 gear grind is a real endgame complaint but a lower-leverage one: it belongs in a balance-tuning content patch, not in the same hotfix as the two reliability wounds.
What NOT to spend the next sprint on
- Difficulty complaints from sub-1-hour, pre-first-forge reviews. The permadeath, forge-to-progress design is the thesis; these reviewers refunded before the forge opened up, and softening the early wall would risk the depth your 20h+ crafters praise.
- Forge UI stutter reports from before Patch 1.3. Almost all of them predate the forge UI performance pass and mentions have gone quiet since; reply to the old reviews rather than reopening that work.
Compiled from the per-issue analyses below.
- Every negative review is read on its own, then grouped into issues and scored by volume, severity, momentum, and how recoverable each one looks. The report states how many reviews it got through, so you can judge the coverage yourself.
- Each top issue comes with a likely root cause, a recommended action, and the player quotes it was built from, so you can check the reasoning before committing a sprint to it.
- A review-score recovery projection: a realistic estimate plus a clearly labelled best-case ceiling, with the method shown. It is an estimate under stated assumptions, not a promise.
- The negative reviews you could win back, split into the ones a shipped update already covers and the ones that would flip if you fix a top issue, each linked into the Workbench reply flow.
- Store-page expectation fixes: concrete copy changes for the places where what players expected and what they got have drifted apart.
- It takes 6–15 minutes and runs in the background. Close the tab and we email you when it is ready, with the full report attached as a PDF and the interactive version waiting in your dashboard.
- Included on every paid plan, every month (Indie 1, Studio 2, Publisher 3). A free trial includes one. Need more in a month? Extra reports are €9 each on a paid plan. VAT included.
It runs on a game you manage, alongside the weekly digest rather than instead of it. Read a full sample report
Research Lab
Zoom out from one game to a whole basket. Research Lab reads reviews across a set of titles at once and synthesizes what they share: the complaints that repeat, the gaps nobody has filled, the needs a genre keeps leaving on the table.
Genre Deep-Dive
What are the must-haves and dealbreakers for a new co-op friendslop game?
Basket: R.E.P.O., PEAK, YAPYAP, Content Warning, RV There Yet? (5 games analyzed)
Loved by Players
- Chaotic, emergent co-op moments with friends are the core emotional product players buy. Every top-rated game in this set is praised primarily for the unscripted, hilarious situations that arise from physics, voice chat, or sandbox interactions rather than designed setpieces.(R.E.P.O., PEAK)
- Accessible, budget-friendly price points drive initial adoption and forgiveness for rough edges. Content Warning, YAPYAP, and RV There Yet? are all praised for delivering strong value at low cost, which lowers the barrier for group purchases.(Content Warning, YAPYAP)
- Physics-based interactions create unpredictable and repeatable comedy that players describe as the reason sessions stay fun. R.E.P.O. highlights goofy physics, RV There Yet? praises winch and RV mechanics, and YAPYAP celebrates floppy physics as core to the experience.(R.E.P.O., RV There Yet?)
- Proximity voice chat and social audio dramatically amplify co-op comedy and immersion. Content Warning is singled out for this as a major strength, and R.E.P.O. players cite social connection as a key pillar of the experience.(Content Warning, R.E.P.O.)
- Randomized or rotating content keeps sessions feeling fresh and gives players a reason to return. R.E.P.O. uses random level and item generation, while PEAK uses daily rotating maps, and both are praised specifically for this feature.(R.E.P.O., PEAK)
Hated by Players
- Solo play is uniformly broken or joyless across the genre. Every game in this set is described as significantly worse or outright unplayable when played alone. This is the single most consistent negative pattern across all five titles.(R.E.P.O., PEAK)
- Thin content that players exhaust in under 3 hours is a repeated complaint that leads to refund requests and negative word of mouth. YAPYAP is the most extreme case at 2-3 hours, but RV There Yet? and PEAK also face drop-off after novelty fades.(YAPYAP, PEAK)
- Persistent bugs that block progress and force restarts are disproportionately damaging to short-session games. PEAK's cannon glitch forces campfire restarts, Content Warning has equipment and ceiling clipping bugs, and R.E.P.O. has launch failures. Players are less forgiving when total playtime is already low.(PEAK, Content Warning)
- Repetitive loops without meaningful progression cause sessions to lose their appeal quickly. RV There Yet? players cite the repair-and-resupply loop as tiring, PEAK players note novelty wears off after a few sessions, and R.E.P.O. players mention repetitive settings and monster variety.(RV There Yet?, PEAK)
- Poor checkpoint and save systems in games with meaningful progression cause outsized frustration. RV There Yet? is the primary case, where players must repeat large sections after failure, but the pattern signals that players expect low-friction recovery from failure in casual co-op.(RV There Yet?)
Divisive Points
- Unique mechanics that differentiate a game are praised when they work consistently but become the top complaint when they are unreliable. YAPYAP's voice-activated spell casting is loved for its comedy and hated for its inconsistent recognition. The mechanic itself is the differentiator, but reliability is non-negotiable.(YAPYAP)
- Modding support dramatically extends game life for engaged players but creates friction for casual audiences. Content Warning players cite mods as a major strength, but also complain that modding is hard to set up and causes hosting issues. YAPYAP and PEAK have no noted mod support.(Content Warning)
- Depth of mechanics creates a split between players who want accessible chaos and players who want genuine challenge. RV There Yet? is praised for its winch and manual transmission depth by some players but labeled as streamer slop by others. R.E.P.O. sits closer to the chaos end and scores higher satisfaction.(RV There Yet?, R.E.P.O.)
Missing Features
Public matchmaking with social filtering so players without friend groups can access the core co-op experience
How to implement:
- • Add lobby tags for playstyle preferences such as casual, competitive, or streamer-friendly so players self-select into compatible groups
- • Implement a simple pre-game voice or text chat lobby so players can gauge group chemistry before committing to a run
- • Consider a region and ping filter as a baseline to reduce the lag complaints seen in R.E.P.O. while building broader matchmaking
⚠️ Risks:
- • Public lobbies in this genre are prone to griefing, which PEAK and R.E.P.O. both document as a trust-breaker
- • Matchmaking infrastructure costs can outpace revenue for games with small or declining concurrent player counts
- • Poorly implemented matchmaking can feel worse than no matchmaking and generate its own negative reviews
✓ Success metrics:
- • Track the ratio of solo versus group session starts and measure whether solo-initiated sessions convert to completed runs after matchmaking is added
- • Monitor review sentiment specifically referencing solo play or matchmaking over the 30 days post-feature launch
- • Measure average session length for matchmade lobbies versus friend-invited lobbies to assess engagement parity
Persistent light progression or meta-layer that rewards returning players between sessions
How to implement:
- • Add a cosmetic unlock system tied to run count or chaos points so players have a visible reason to return even when content is fully seen
- • Implement a challenge or daily mission board that gives each session a distinct goal, similar to how PEAK uses daily rotating maps
- • Consider carrying over a small bonus between runs such as starting items or stat buffs to create a light roguelite sense of momentum without undermining the reset loop
⚠️ Risks:
- • Heavy progression can shift the game's identity from casual party game to grind game, alienating the core audience
- • Cosmetic-only progression may feel hollow if the cosmetics themselves are not interesting or visible enough during play
- • Poorly balanced progression can create skill gaps between veteran and new players that undermine the accessible chaos feel
✓ Success metrics:
- • Measure day 7 and day 30 retention rates before and after the progression system launches
- • Track average number of sessions per player per week to determine if the loop is pulling players back
- • Monitor review language for terms like repetitive or nothing to do as a proxy for whether progression is solving the depth problem
Reliable autosave and low-friction respawn or checkpoint system scaled to session length
How to implement:
- • Implement automatic checkpoint saves at natural pause points such as campfires in PEAK or between floors in R.E.P.O. so failure does not erase more than one segment of progress
- • Add a clear visual indicator of the last save point so players know exactly what they risk before attempting a difficult section
- • For games with short expected session lengths, bias toward more frequent saves since the cost of lost progress is proportionally higher when total play time is already low
⚠️ Risks:
- • Too-frequent saves can trivialize challenge and reduce the comedic tension that comes from genuine risk
- • Save system changes may require significant architecture work in games that were not built with mid-session persistence in mind
- • Players may still complain about checkpoints if they are placed just before unskippable cutscenes or lengthy load sequences
✓ Success metrics:
- • Track early session abandonment rates and compare before and after checkpoint improvements
- • Review Steam reviews for checkpoint or autosave mentions and track sentiment shift
- • Measure completion rates on early-game achievement milestones as a proxy for whether players are progressing rather than quitting in frustration
Underserved Needs
Spectator or clip-sharing tools that let players relive and share their funniest moments without leaving the game
How to implement:
- • Build a short in-game replay buffer similar to Content Warning's footage playback mechanic but extended to all games in the genre, allowing players to save the last 60-90 seconds on demand
- • Add a simple highlight reel screen at the end of each run showing auto-detected high-chaos moments using physics impact thresholds or voice activity spikes
- • Integrate a one-click clip-to-social or clip-to-Steam screenshot feature so players can share moments without breaking session flow
⚠️ Risks:
- • Replay systems are technically complex and can introduce performance overhead that hurts the base experience
- • Players may already use third-party tools like Nvidia ShadowPlay, making an in-game version feel redundant if not done well
✓ Success metrics:
- • Track social media mentions and user-generated content volume after the feature launches
- • Measure how often the replay or clip feature is used per session as a proxy for perceived value
Scalable difficulty or chaos settings that let groups tune the experience to their skill level and desired intensity
How to implement:
- • Add a lobby difficulty slider with named presets such as Chill, Chaotic, and Nightmare that adjust enemy aggression, item spawn rates, and hazard frequency
- • Allow the host to toggle specific modifiers before each run so experienced groups can stack challenge while new players use defaults
- • Surface difficulty options prominently during lobby setup, not buried in settings, to ensure groups discuss expectations before starting
⚠️ Risks:
- • Too many options can overwhelm new players and create analysis paralysis before sessions even begin
- • Difficulty scaling that reduces chaos may inadvertently remove the core comedy that defines the genre
✓ Success metrics:
- • Track what percentage of lobbies use non-default difficulty settings to gauge adoption
- • Monitor session length on higher difficulty settings to assess whether the challenge extends engagement
Market Opportunities
A game in this genre designed for true solo play as a first-class mode, not an afterthought
How to implement:
- • Design AI companions with expressive physics-based behavior so solo players get emergent chaos from bot interactions rather than empty co-op slots
- • Create solo-specific scenarios or narrative objectives that reward playing alone rather than treating solo as a reduced co-op experience
- • Use asynchronous social features such as ghost data or shared leaderboards to give solo players a sense of community without requiring synchronous groups
⚠️ Risks:
- • AI companions risk feeling uncanny or annoying rather than funny, which would directly undermine the game's core identity
- • Building a compelling solo mode requires significant design and engineering investment that may dilute focus from the core co-op product
- • The genre may have a ceiling on solo appeal that no amount of AI investment can overcome
✓ Success metrics:
- • Track the percentage of sessions started in solo mode post-launch and compare to genre peers
- • Measure review sentiment among players who self-identify as solo players to confirm the mode is genuinely satisfying
- Build the basket however you think: search by name, pull in similar games, or drop in your own portfolio.
- Pick what you are asking: a genre deep-dive, a competitive read on specific rivals, or validation for a game you are thinking of making.
- Synthesis runs across every game at once, not one report per title, so you get the cross-game patterns instead of a stack of separate SWOTs to reconcile yourself.
- It runs in the background. Start a lab, close the tab, come back to the writeup.
Run this on your own game
Start with a free SWOT of any title, no account needed. When you want Pro Insights, Edges, the Workbench, and the Action Report pointed at your own game, start a trial. No card up front.
Full plan details on the pricing page.