Game Discoverability Problem Now a Structural Threat, Industry Workgroup Warns

A workgroup at the DevGAMM Madeira Games Summit was asked in May what the games industry should fix while it still has the chance, and returned a unanimous answer: discoverability. The group, led by Modma Studios senior business development manager Michal Bujko, concluded that the industry already makes great games and that almost nobody can find them.

Its findings were published on 4 August by GamesIndustry.biz, three months after the two-day summit closed in Funchal. The group’s central claim is that the problem has stopped being a marketing inconvenience and has become a structural threat to the long-term health and cultural credibility of the medium.

What the Workgroup Found at the Madeira Games Summit

Bujko framed the session by asking attendees to look back at 2026 from five years out and name what should have been fixed. The room agreed on discoverability without dissent, and agreed that years of collective awareness had produced almost no change.

His argument is that content-surfacing algorithms have come to favour output volume, resemblance to existing hits, and immediate engagement, while the taste-led judgement that once gave unusual games a fighting chance has been displaced by metrics. Genuinely distinctive games get buried because the system was never built to uncover them.

Underneath that sits a harsher second finding. The group did not stop at asking how players find great games, and moved to asking whether the industry is still making games worth finding. Chasing proven formulas, replicating last year’s hit and optimising for day-one numbers all add to the noise the problem is made of.

Why the Discoverability Problem Has Not Been Fixed

The workgroup’s answer is that fixing it would be expensive, risky, and against the short-term interests of nearly everyone in the chain. Each party is behaving rationally inside its own incentive, and the sum of those rational choices is the problem.

Nobody, the group found, is financially motivated to repair the system as a whole. If success depends on feeding the algorithm, the safe option wins every time. That finding is the reason the diagnosis has changed nothing on its own: a correct diagnosis has no force when no party stands to gain from acting on it.

Bujko reached for Goodhart’s law to describe the mechanism, the principle that when a measure becomes a target it stops being a good measure. Marketers have chased store ranking, app store optimisation and, more recently, visibility inside AI-powered recommendations and LLM-driven search. Once the metric is gamed, any product can reach the top on sharp marketing alone, and discovery ends up measuring optimisation skill instead of quality.

One platform veteran in the room put the difficulty in blunter terms, arguing that the code may be unbreakable because most of the audience runs on instinct rather than research or reviews. The group treated that as a reason to take the problem seriously rather than a reason to abandon it.

Platforms Hold the Most Power and the Least Accountability

The group identified platforms as the parties with the greatest structural control over discovery and the weakest obligation to explain how it works. Storefronts run at scale as black boxes, and developers and publishers see almost nothing of how their games actually perform inside them. Optimisation carried out without performance data is not optimisation at all, it is trial and error paid for by the developer.

The recommendations aimed at platforms were specific: proper attribution tools, granular analytics, and a real commitment to human curation running alongside algorithmic recommendation. The group also called for better editorial placements and more of them, plus commercial terms that reward a diverse publishing ecosystem rather than concentrating attention on titles that are already visible.

What the Workgroup Told Business Leaders to Stop and Start

The session closed with two lists aimed directly at decision-makers.

What Leaders Should Stop Doing

• Building marketing plans around platform features, since a feature can amplify reach but cannot substitute for a strategy
• Greenlighting on comparables alone, when a game not resembling something that already worked may be the point rather than a warning sign
• Treating curation as somebody else’s job, when editorial instinct and the ability to recognise work worth backing are business competencies
• Doubling down on formats and mechanics that worked five years ago, when the audience and the medium have moved on

What Leaders Should Start Doing

• Owning their marketing by building direct audience relationships through their own channels, starting before the game is finished
• Bringing taste back into the room and reinvesting in editorial judgement, treating metrics as a tool rather than a decision-maker
• Making fewer and more deliberate bets, on games with a clear reason to exist and a clear audience to reach
• Taking genuine creative risks as policy rather than as an exception
• Fixing the information asymmetry, with platforms supplying the data publishers and developers need to understand what is working

Paradox Interactive was the example the room returned to. The Crusader Kings publisher built a self-sustaining community around a specific type of game and turned that community into its own discovery engine. The group’s conclusion was that publishers should commit to a lane and build real expertise inside it, because a broad catalogue with no identity is effectively invisible in discoverability terms.

For developers, the advice was to treat community building as part of the creative process from the moment a game is conceived rather than as a marketing phase that starts at ship. The group also argued that the games that get found are almost always the ones that are genuinely different, which makes unexpected mechanics and unconventional structure a strategic requirement rather than an artistic indulgence. Smaller studios cannot outspend the major players, so outthinking them is the only leverage left.

Investors and service providers, the group noted, are largely absent from this conversation, and that absence is itself part of the problem.

A Second Madeira Workgroup Reached a Parallel Conclusion on Risk

The discoverability session was not the only one at the summit to land on incentives. A separate workgroup hosted by Vlambeer co-founder and Nuclear Throne producer Rami Ismail, titled “We’re moving fast and losing what made us good”, concluded in a report published in July that the industry’s reflexive avoidance of failure has become a greater threat to success than the uncertainty of creative work itself.

Ismail’s group argued that fearful reflexes sand the rough edges off creative work and produce homogenous output that ticks every box while inspiring nobody. Clair Obscur: Expedition 33, Peak and Balatro were named as titles conventional market analysis would never have produced, all of which found audiences because they carried a clear point of view. The group settled on four pillars for creative development: coherence, identifying weaknesses, selling the dream, and connecting the dots.

Speaking to Gamereactor at the summit in May, Ismail made the same case from the funding side, arguing that “all the games that are super successful right now are very stubborn games” and that publishers and investors have become too narrowly focused on the single game rather than the longer arc. Two independent workgroups at the same event arriving at metrics-over-taste from opposite ends suggests the summit’s attendees were describing one problem, not two.

The Counter-Case That Supply, Not Quality, Is the Bottleneck

The workgroup’s suggestion that too little of what the industry makes actually matters sits awkwardly against the available data on release quality. Bloomberg’s analysis of the Steam catalogue, published in September 2025 by Jason Schreier, found that of 1,431 games released the previous year that drew more than 500 user reviews, more than 260 were rated positively by 90 per cent or more of players, and more than 800 scored 80 per cent or better. The piece explicitly rejected any comparison with the 1983 crash, where the flood was of poor product rather than good.

If more than half of the reviewed releases in that sample cleared 80 per cent player approval, the shortage the industry is dealing with is attention rather than craft. That reframes the workgroup’s harder question: the games worth finding may already exist in volume, and the failure sits entirely on the distribution side.

Supply has not slowed since. Steam’s June 2026 Next Fest carried more than 4,300 demos, a record for the event, according to discovery analyst Simon Carless of GameDiscoverCo. The same analysis found that demos with Valve’s AI content disclosure option enabled rose from 750, or 21.2 per cent of the February 2026 festival, to 1,163, or 26.5 per cent in June. Every additional entrant makes the visible slots the workgroup wants platforms to open up more contested, not less.

The Five-Year Window the Group Set Itself

Bujko’s closing position was that the diagnosis is no longer the hard part, and that the window to act is open only while the medium is still growing and the damage is still reversible. The group’s five-year forecast is that the games which deserved to be found and were not will simply be gone, the studios that could not survive the noise will have closed, and the audiences that might have formed around them will never form retroactively.

The deadline the workgroup set is 2031, five years from the summit that produced the finding, and no platform holder named in the report has responded publicly to its recommendations. Bujko’s argument for treating it as this week’s problem rather than this decade’s rests on what he expects the alternative to sound like, warning that “we knew and did nothing” will be the hardest sentence to say out loud.