Build something worth returning to
Retention is evidence of usefulness and an input to the economics. The game should stand on its own.
An exploration in progress
What becomes possible when a small team can build useful websites, bring people to them, and connect advertising services with far less overhead?
Start with one real experimentAdvertising has mature buying, delivery, measurement, and payment systems. The idea is to wire those systems together effectively, with agents reducing the effort of building and running the service. The outcome should be an easier way for useful websites to earn from the audiences they serve.
Cribbage is the first destination: a game worth playing and returning to. Its value to the player is the foundation. Paid distribution can introduce people to it; advertising may eventually help fund that experience.
Retention is evidence of usefulness and an input to the economics. The game should stand on its own.
Connect established services. Choose the smallest integration that answers the next important question.
Real people, accurately represented placements, a clear publisher share, and an experience that respects attention.
PHASE ONE · BOOTSTRAP
Start with Cribbage. Purchase a bounded campaign, observe real play, and learn the cost of acquiring and retaining players. Use existing infrastructure where needed.
The first budget buys evidence. Immediate advertising payback is an open question.
PHASE TWO · THE EXISTING WEB
Help independent publishers access advertiser demand and manage their monetization. Their audiences may already arrive through search, repeat visits, or their own distribution.
The platform earns a fee or share. It does not need to purchase every visit.
Agent-centric could mean an owner’s authorized agent connects a site, sets preferences, and manages earnings through an API. Ads inside AI products are a separate possible channel. Neither requires a new advertising exchange.
One paid introduction can lead to one session or many. A publisher’s ad receipts and a platform’s retained fee are different quantities. The models below keep them separate.
WordStream’s 2026 report lists traffic-campaign CPCs of $0.34 for Arts & Entertainment and $0.60 overall. These are broad campaign benchmarks, not prices we have achieved for a browser game. Read the source ↗
MODEL A · OWNED SITE
Revenue per 1,000 sessions is after upstream ad fees, before our operating costs. No comparable Cribbage yield has been measured. Each acquired visitor’s average includes people who never return.
At these assumptions, repeat play has not yet repaid acquisition.
| Revenue / 1,000 sessions | Per session | Sessions to recover $0.10 | Sessions to recover $0.34 |
|---|---|---|---|
| $2 | $0.002 | 50 | 170 |
| $10 | $0.01 | 10 | 34 |
| $25 | $0.025 | 4 | 14, rounded up |
MODEL B · PUBLISHER SERVICE
Publishers bring their audiences. This model applies our share to receipts after upstream fees. The publisher payout is not our revenue. The operating-cost input is illustrative and should include support, payments, infrastructure, sales, and other service costs.
A 10% take rate is not a 10% profit margin. Founder compensation, taxes, timing of payments, and unmodeled costs still matter.
It can lower development and operating effort, and make more experiments affordable. Acquisition prices, advertiser demand, retention, partner access, and support workload still need their own evidence. Two visible ad slots do not establish the number of billable impressions in a long game.
PROPOSED · NO SPEND AUTHORIZED
Begin with an advertisement for a useful game. Observe what happens after the click. A full publisher platform and advertising inside the game can follow the evidence.
The earlier $100 campaign and $0.10 click were examples. They are neither approved budgets nor achieved results. No paid campaign, demand agreement, or public launch of Cribbage is authorized by this research page.
These sources establish possible mechanisms and constraints. They do not establish access, negotiated terms, or this project’s performance. Checked 4 October 2026.
Channels open to an individual, click ids and server-side conversions, which ad platforms an agent can drive, demand partners with their minimums, the agentic standards that are live versus announced, and payout rails. Sourced.
→ ACQUISITION2026 Meta campaign benchmarks. Broad US sample; directional context with methodology editing inconsistencies, not game-specific estimates.
↗ ENTRY REQUIREMENTSOriginal, useful content and policy compliance. No numerical traffic minimum appears in this checklist; approval remains a separate step.
↗ PAID ACQUISITIONGoogle allows compliant paid promotion and places responsibility for traffic quality on the publisher.
↗ VALUE TO THE VISITORUseful original destinations matter. Advertising-heavy pages and replicated content without added value can be disapproved.
↗ MONEY FLOWPublisher earnings follow upstream fees. Keep advertiser spending, publisher receipts, and a subsequent platform share separate.
↗ MULTIPLE PUBLISHERSAccount or inventory delegation and payment arrangements for publishers. An Ad Manager 360 feature; access is not established here.
↗ PLATFORM INTEGRATIONEnterprise onboarding, reporting, and revenue sharing for suitable platforms and site builders. A research lead, not a selected provider.
↗ EXISTING INFRASTRUCTURETools for connecting demand bids to ad opportunities. The software and the commercial demand relationships are separate pieces.
↗ DEMAND CONNECTIONSProgrammatic interfaces for demand partners. Its setup requires a commercial demand agreement before the integration.
↗ PARTNER CONSTRAINTSPublished website minimum: 500,000 monthly pageviews. Net-60 payment terms. Its layout requirements need comparison with a restrained game experience.
↗ AGENTSAgentic advertising standards and buyer/seller reference implementations. A foundation to examine as the agent’s role becomes clearer.
↗