GDnD Wiki

GDnD Wiki

This site is for Ulster University students and staff. Enter the class access password to continue.

The Pyramid of Game Design

11 min read

Summary

The Pyramid of Game Design is a framework developed by Nicholas Lovell for designing, producing and launching service games (free-to-play and games-as-a-service). It argues that a successful service game is built in three distinct but interlocking layers — Base, Retention and Superfan — connected by a Core Loop and delivered through the rhythm of a Session. The framework is described in full in Lovell’s book of the same name (Lovell, The Pyramid of Game Design, see source-pyramid-of-game-design).

The Pyramid does not replace traditional concepts of gameplay and progression systems. It provides a lens specifically suited to the demands of service games, where a player’s choice to return — not their initial purchase — is the primary commercial event.

Key Ideas

The Three Layers

Base Layer The moment-to-moment gameplay: the repeatable, enjoyable core action. A match in Team Fortress 2, a level of Candy Crush Saga, a run in Temple Run. The Base Layer answers the question “what kind of game is this?” — the genre shorthand that describes what the player actually does.

For product games, a strong Base Layer is often the whole game. For service games, the Base Layer is necessary but not sufficient. Lovell describes it as “necessary but not sufficient”: “While I always recommend that developers seek to maximise the fun in their game, I am uncomfortable relying on ‘be super fun’ as your primary strategy for ensuring a long life for your games.” (Lovell, The Pyramid of Game Design, Ch. 2)

Key characteristics of a service-game Base Layer:

  • It is fun (for some value of fun — not necessarily lighthearted)
  • It is replayable or cheap to create at volume
  • It connects with the Retention Layer to give a sense of progress
  • It has elements of unpredictability — from randomness, other players, or emergent system behaviour

Some successful service games have no meaningful Base Layer at all. Resource-management games like Hay Day and idle games like Adventure Capitalist are pure Retention Layer experiences. A game can have Base Layer gameplay, Retention Layer gameplay, or both.

Retention Layer The systems and feedback that keep players returning for days, weeks, months and years. The Retention Layer is what transforms a fun diversion into an enduring hobby. Lovell frames it as a system-design problem, not an asset-creation problem: the goal is to create interesting systems that evolve and give players meaningful choices over a long period, rather than simply producing enough content to fill a fixed number of hours.

Lovell catalogues eighteen Retention Layer techniques, arranged roughly from simple to sophisticated:

#TechniqueDescription
1Be Really FunNovelty and enjoyment as a retention tool in its own right; risky without other support
2High Score”Avoid missing ball for high score” — the Pong approach
3Persistent High ScoreTracking a personal best across sessions
4LeaderboardsSocial comparison; suffers from outlier demotivation without resets
5SeasonsPeriodic resets that let players compete for top positions repeatedly
6LevelsDiscrete progress stages; the motivation to “see what’s next”
7Getting to the EndA finite completion goal; powerful but trades long-term retention
8Levelling UpPersistent skill/character growth satisfying Self-Determination Theory’s competence need
9Progress Across a MapVisual/spatial representation of advancement (e.g. Candy Crush Saga map)
10LoopsNested short/medium/long-term goals drawing on the Zeigarnik effect and endowed-progress effect
11Real TimeServer-side clocks creating appointment obligations and time-gated content
12Appointment MechanicsPlayer-chosen return commitments (not Tamagotchi demands); Cialdini’s commitment principle
13Narrative”What happens next?”; powerful but expensive at AAA scale in a service context
14Star SystemMulti-tier level completion rating driving replay
15Stars as CurrencyStars become a resource spent on upgrades, turning completionism into mechanical progress
16ChallengesRepeatable tasks issuing fresh rewards each time (distinct from Achievements)
17AchievementsOne-time permanent markers of competence; three archetypes — Pointless, Reverse, Challenging
18SocialHuman relationships as the strongest long-term retention; expensive to build, powerful when present

Superfan Layer Where the game becomes a hobby for a small subset of players. Superfans mainly play this game; it occupies a disproportionate share of their time and potentially their money. They are the commercial spine of most F2P titles: Swrve data (2014) showed that 0.15% of players generated 50% of revenue; Kongregate data showed the top 4% of spenders (0.084% of the entire player base) generated 72% of revenue (cited in Lovell, The Pyramid of Game Design, Ch. 7).

Lovell explicitly rejects the term “whales” as dehumanising and commercially counter-productive. Designing for “Superfans” means making something players love first, then enabling those who love it most to spend heavily on things they genuinely value.

Characteristics of successful Superfan Layers (from Emily Greer’s GDC data on Kongregate, cited in Ch. 7):

  • Guilds or leagues — social commitment and reciprocity
  • Guild warfare or leaderboards — structured inter-group competition
  • PvP (synchronous or asynchronous) — human unpredictability and competitive drive
  • Visible status and character progression — self-expression in a social context

The Core Loop

The Core Loop takes players from the Base Layer into the Retention Layer and back again. It is not simply the Base Layer repeated — it is the cyclical relationship between moment-to-moment action and long-term progression. Lovell identifies three types:

  • Linear Core Loop — complete unique missions to progress a narrative; scripted, expensive, common in product games
  • Strategic Core Loop — meaningful choices in both Base and Retention Layers that interact bidirectionally (e.g. XCOM: Enemy Unknown)
  • Replayable Core Loop — repeat the Base Layer to earn resources that fuel Retention Layer progress; the dominant pattern in service games (e.g. Temple Run, CSR Racing, Hearthstone)

The Gearbox

The Gearbox is the design mechanism that transmits information and emotional meaning between the Base and Retention Layers. It is what makes the Core Loop legible to players.

Its primary implementation is the Pre-Event Screen and the Post-Event Screen:

  • Pre-Event Screen — shows the player what they are about to face in the Base Layer, what choices they can bring from the Retention Layer, and what rewards are at stake. Examples range from XCOM’s detailed mission briefing to Candy Crush Saga’s minimal boost-purchase screen.
  • Post-Event Screen — shows the player what they earned, how they progressed and what has changed in the Retention Layer as a result of the Base Layer experience.

Lovell applies the Hero’s Journey as a metaphor: the Retention Layer is the Ordinary World; the Base Layer is the Special World; the Pre- and Post-Event Screens are the Threshold crossings. Symmetry between the two screens helps close the Core Loop satisfyingly.

Three Design Razors for Gearbox features:

  1. Will the player see the feature?
  2. Will the player understand the feature?
  3. Will the player value the feature?

In Practice

Designing the Layers

Prioritise the Retention Layer. In product games, polish goes into the Base Layer. In service games, the Retention Layer is the harder, more uncertain design problem — and the one that determines long-term commercial success. Lovell recounts a client who outsourced their Base Layer (a “known problem”) to a third-party studio and focused in-house development on the Retention Layer (an “unknown problem” requiring iteration and prototyping).

Prototype the Base Layer early, the Retention Layer slowly. The Base Layer can be tested in a short playtest session. The Retention Layer must be tested over days or weeks to see whether players genuinely return. Do not invest heavily in Retention Layer content before the Base Layer is fun.

Not all games need all three layers. The Pyramid is not always triangular:

  • Endless runners (Minion Rush, Sonic Dash): wide base, narrow Retention, minimal Superfan
  • Resource managers / idle games (Hay Day, Adventure Capitalist): almost no Base Layer, all Retention
  • Deep strategy titles (Clash of Clans, Mobile Strike): emphasis on Superfan Layer; without guilds and PvP the game loses most of its commercial engine

Build in layers, not as three separate games. The Base, Retention and Superfan components should be one coherent game with multiple layers. Designing them as separate experiences multiplies scope unnecessarily.

Designing the Gearbox

  • Include a Pre-Event Screen whenever the player has meaningful choices to bring from the Retention Layer into the Base Layer.
  • Include a Post-Event Screen that directly mirrors the Pre-Event Screen, showing what was earned and how the Retention Layer has changed.
  • Ensure the Post-Event Screen does not require players to navigate away from it — keep rewards and progress information on a single screen.
  • Use ceremony (animation, audio, visual feedback) to amplify the emotion of completing the Base Layer.

The Pyramid and Monetisation

The Superfan Layer is where variable pricing — the economic driver of F2P — operates. “Variable pricing enables you to let the players who love what you do spend lots of money on things they really value.” (Lovell, Ch. 1)

This is distinct from the freemium model. Freemium offers a free tier and a paid tier; it has two levers (audience size and conversion rate). F2P adds a third lever: ARPU — how much engaged players spend. A successful F2P game can be commercially viable with a small but committed audience if the Superfan Layer enables sufficient spend.

What Superfans pay for (Lovell, Ch. 8):

  • Self-expression — avatar items, cosmetics; sells better when other players can see the purchase (display), when it is on a character the player controls (agency), and when it persists (permanence)
  • Conspicuous consumption — status items signalling investment
  • Power — items affecting Retention Layer progress rather than Base Layer combat are more acceptable than “pay-to-win”
  • Relationships — spending to help guild mates, avoid letting the clan down
  • Content — still viable via DLC, though less central than emotion and status

Evidence

“Service games view retention as a system-design problem: how do we create interesting systems that evolve to give players meaningful choices, experiences and progression for a very long time without the prohibitive costs of high-quality asset generation?” (Lovell, The Pyramid of Game Design, Ch. 1)

“The purpose of the Pre-Event Screen is to provide a connection between the Retention Layer and the Base Layer. It allows players to understand their progress and, if appropriate, to show them how their achievement in the Retention Layer is making their Base Layer easier.” (Lovell, Ch. 4)

“When a service-game designer says that it is critical that you have a polished and designed Core Loop, they don’t mean that you must have a Base Layer with AAA levels of polish. They mean that the interrelationship between moment-to-moment action and long-term progression must be polished and clear to players.” (Lovell, Ch. 4)

“The Pyramid is not one-size-fits-all… Many games have different-shaped pyramids.” (Lovell, Ch. 7)

Implications

  • Greenlight processes based on vertical slices are poorly suited to service games. A vertical slice evaluates the Base Layer; it tells you almost nothing about whether the Retention Layer and Core Loop will sustain months of play.
  • Retention is an internal, controllable metric; acquisition cost is external. Focusing development energy on Retention Layer quality is strategically sound because it is the part of the system most within a team’s control.
  • Content is a retention tool, not a monetisation tool in most service games. Releasing new content keeps players engaged; the monetisation comes from the Superfan Layer of social status, self-expression and power progression.
  • “Be really fun” is not a retention strategy. Fun matters, but novelty decays. Service games need structural reasons for players to return — nested loops, appointment mechanics, social obligations, and long-term progression goals.

Open Questions

  • How much of the Pyramid framework specifically addresses mobile and how well does it generalise to PC and console service games (live-service titles like Destiny 2, Sea of Thieves, GTA Online)?
  • Lovell treats narrative as prohibitively expensive for service games. How do games like Genshin Impact reconcile heavy narrative investment with F2P service mechanics?
  • Where does the Pyramid sit relative to the MDA framework (mda-framework)? The Retention Layer is largely about Mechanics and Dynamics; the Superfan Layer becomes Aesthetics (community, identity). Can they be unified?
  • The framework was written in 2018. Has the Superfan model held as platform economics have changed (App Store privacy changes, rising CPI)?

Graph View

Backlinks