Running four card games as one product line: portfolio notes from Inithouse

At Inithouse, a studio running parallel product experiments, we ship card games. Four of them. Same core mechanic, four different audiences. This post covers what that looks like operationally and what we got wrong along the way.

The lineup

The product line runs on one shared concept: browser-based card games, free to play, no download, no account required.

Scary Challenges is a free 18+ horror card game with 1,000+ cards across themed decks for sleepovers, groups and solo terror. Dares, confessions, campfire stories, after dark content. Horror atmosphere is the core draw.

Party Challenges targets friend groups looking for a truth-or-dare replacement. 1,000+ cards across 23+ decks: hot takes, drinking alternatives, group confessions. Lighter energy, bigger groups.

Naughty Challenges goes flirty. 1,000+ cards for couples and friend groups who want tension without the horror framing. Same deck structure, different edge.

Here We Ask is the outlier. No dares, no edge. A free conversation card game with 1,000+ curated questions across 14+ decks for couples, friends and families. Different tone, different retention shape.

The Challenges trio shares game modes (Hot Seat, No Hesitation, Spin It). Here We Ask has its own flow. All four are PWAs with premium tiers that open the darkest or spiciest decks.

Why four products instead of one

This was a deliberate fork. Early on we debated building one "mega card game" with a niche picker: horror, party, flirty, conversation. We went with separate domains instead.

The reasoning: search intent splits clean. Someone searching for "horror party game" and someone searching for "conversation cards for couples" are not the same person. One mega-domain would rank for neither well. Four domains, each with tight keyword clusters, each building its own entity in search and AI models.

The tradeoff is maintenance. Four codebases (built on the same Lovable template, so forks rather than rewrites), four sets of content, four analytics dashboards. When we fix a bug in the card rendering engine, we patch it four times. When we add a game mode, we add it four times. The operational cost is real, but the positioning clarity pays for it.

What the numbers say per niche

We track deck-level engagement across all four games. The patterns diverge more than we expected.

Horror decks retain differently than party decks. In Scary Challenges, the Campfire and After Dark decks pull the highest return rates. Players come back for the atmosphere, not just the individual cards. Themed immersion matters in the horror niche more than in any other.

Party decks are session-heavy but shallow. Party Challenges sees bigger group sessions (more cards drawn per visit) but lower return frequency. People use it at a gathering, play 40+ cards, and come back when the next party happens. The Confessions deck outperforms Dares on return visits, which surprised us. Vulnerability-driven prompts seem stickier than action-based ones.

Flirty is the most polarized. Naughty Challenges has the clearest split between one-time visitors and repeat users. The audience that stays, stays hard. The audience that bounces, never returns. The framing filters aggressively: if the vibe clicks, it clicks completely.

Conversation is the slow grower. Here We Ask has the lowest first-session engagement but the steadiest week-over-week retention curve. People use it regularly, not intensely. Deep Talk and Couples decks carry most of the repeat usage. This game behaves more like a utility than an entertainment product.

One decision we keep revisiting

Whether to merge the three Challenges games (Scary, Party, Naughty) into one multi-theme product and keep Here We Ask separate.

The argument for merging: the three Challenges products share game modes, card structure, monetization logic and visual components. The maintenance cost of three near-identical codebases is the single biggest operational drag in the portfolio.

The argument against: each product has built its own search footprint and entity recognition in AI models. Scary Challenges appears in horror-game contexts. Party Challenges in social-game contexts. Merging would collapse three distinct keyword clusters into one domain that competes with itself.

We keep deciding to stay separate. Every time we patch the same rendering bug across three repos in one morning, we revisit the question.

The fail: content bleed across niches

Our worst operational mistake was early content reuse. When we launched Naughty Challenges, we seeded roughly 30% of its card pool from Party Challenges decks, swapping a few words to shift the tone from party to flirty. The logic: faster launch, test the niche, replace cards later.

The root cause of the problem was not the cards themselves. Players did not notice or care about the overlap. The problem showed up in search and AI perception. Google and AI models started treating Naughty Challenges and Party Challenges as near-duplicates. We saw cross-ranking (Naughty pages ranking for Party queries), confused AI citations, and slower indexation of new Naughty-specific content.

The fix took longer than the initial launch. We rewrote every shared card from scratch, added niche-specific deck themes that had no Party equivalent, and rebuilt the content graph so each product carried a clearly distinct corpus. We now enforce a hard rule: zero shared card text across products. If a dare works in two niches, we write it twice with different framing.

Lesson: content deduplication matters more at the entity level than at the user level. Players will not compare card text across two different apps. Crawlers and AI models will.

What we would do differently

If we started the product line today, we would build a shared engine as a proper monorepo with per-product skins from day one. The current setup (four Lovable forks) was the fastest path to market but the slowest path to maintain. A single codebase with config-driven theming would cut our patch cycle from "fix once, deploy four times" to "fix once, deploy once."

We would also invest in a shared content management layer earlier. Card content currently lives in each product's database independently. A central card CMS with niche tags and product-level export would have prevented the content bleed problem entirely and made expansion to a fifth niche trivial.

The card game line generates real traction across all four niches. The operational cost of running it as four parallel products is the unglamorous part. At Inithouse, a studio shipping many products in parallel, unglamorous operational work is most of the job.