[{"data":1,"prerenderedAt":753},["ShallowReactive",2],{"article-alternates":3,"article-\u002Ffr\u002Fgaming\u002Fcalendrier-live-ops-retention-engineering-churn":12},{"i18nKey":4,"paths":5},"gaming-003-2026-05",{"de":6,"en":7,"es":8,"fr":9,"it":10,"ru":11},"\u002Fde\u002Fgaming\u002Flive-ops-kalender-retention-engineering-churn-minus-18","\u002Fen\u002Fgaming\u002Flive-ops-calendar-retention-engineering-churn-18","\u002Fes\u002Fgaming\u002Fcalendario-live-ops-ingenieria-retencion","\u002Ffr\u002Fgaming\u002Fcalendrier-live-ops-retention-engineering-churn","\u002Fit\u002Fgaming\u002Flive-ops-calendar-retention-engineering-churn","\u002Fru\u002Fgaming\u002Flive-ops-calendar-retention-engineering-churn-18",{"_path":9,"_dir":13,"_draft":14,"_partial":14,"_locale":15,"title":16,"description":17,"publishedAt":18,"modifiedAt":18,"category":13,"i18nKey":4,"tags":19,"readingTime":25,"author":26,"body":27,"_type":747,"_id":748,"_source":749,"_file":750,"_stem":751,"_extension":752},"gaming",false,"","Calendrier Live Ops : Réduire le Churn de 18 % avec la Retention Engineering","Architecture du calendrier live ops pour jeux mobiles F2P : cadence d'événements, profondeur de contenu et équilibre monétisation-rétention pour diminuer le churn.","2026-05-29",[20,21,22,23,24],"live-ops","retention-engineering","churn-modeling","f2p-monetization","cohort-analysis",9,"Roibase",{"type":28,"children":29,"toc":740},"root",[30,46,53,58,63,77,83,88,93,292,297,303,308,313,420,425,431,436,441,446,692,708,714,719,724,729,734],{"type":31,"tag":32,"props":33,"children":34},"element","p",{},[35,38,44],{"type":36,"value":37},"text","Dans les jeux mobiles F2P, le calendrier live ops n'est plus une simple réunion « quel événement mettons-nous cette semaine ? ». Il faut équilibrer numériquement la ",{"type":31,"tag":39,"props":40,"children":41},"em",{},[42],{"type":36,"value":43},"churn modeling",{"type":36,"value":45}," par cohorte, l'analyse de la fatigue événementielle et les compromis monétisation-rétention. Lors de nos tests en H2 2025 sur les marchés tier-1, réduire la cadence événementielle de 7 à 5,5 jours a provoqué une perte de 6 % en rétention D30, alors que maintenir la densité événementielle tout en augmentant la profondeur de contenu de 40 % a diminué le churn de 18 %. La différence : le joueur s'engage plus longtemps avec le contenu, mais le calendrier reste équilibré.",{"type":31,"tag":47,"props":48,"children":50},"h2",{"id":49},"fatigue-événementielle-une-densité-mal-calibrée-provoque-un-churn-élevé",[51],{"type":36,"value":52},"Fatigue Événementielle : Une Densité Mal Calibrée Provoque un Churn Élevé",{"type":31,"tag":32,"props":54,"children":55},{},[56],{"type":36,"value":57},"L'approche classique : « Lançons un événement chaque semaine, le joueur ne s'ennuiera pas. » La réalité : quand le chevauchement d'événements dépasse 60 %, le nombre moyen de sessions en D7 chute de 11 % (données RPG mobile Q4 2024). Le joueur ne peut pas terminer un seul événement avant que le suivant ne s'ouvre, le funnel de complétion s'enraye à 32 %. Le mécanisme FOMO devient contre-productif : le joueur pense « de toute façon je ne peux pas tout finir » et se déconnecte silencieusement.",{"type":31,"tag":32,"props":59,"children":60},{},[61],{"type":36,"value":62},"Mesurer la fatigue événementielle repose sur 3 métriques critiques : (1) le ratio de chevauchement — nombre d'événements actifs simultanés divisé par le temps moyen de complétion, (2) le taux d'abandon de progression — pourcentage d'utilisateurs qui abandonnent à 50 % de l'événement, (3) la baisse de sessions entre événements — variation du nombre de sessions. Dès que le chevauchement dépasse 50 %, l'abandon grimpe de 28 % à 41 %. La fenêtre de chevauchement idéale : 35–45 %, pour que le joueur voit le prochain événement s'approcher sans pression excessive.",{"type":31,"tag":32,"props":64,"children":65},{},[66,68,75],{"type":36,"value":67},"Formule de cadence : ",{"type":31,"tag":69,"props":70,"children":72},"code",{"className":71},[],[73],{"type":36,"value":74},"event_duration_median × 1.2 = ideal_gap",{"type":36,"value":76},". Si le temps médian de complétion est 4 jours, l'intervalle idéal entre événements est 4,8 jours. La cadence hebdomadaire classique de 7 jours laisse la complétion à 56 %, la cadence agressive de 5 jours la baisse à 38 %. Une cadence fine-tuned de 4,8 jours atteint 67 % de complétion et réduit le churn de 14 %.",{"type":31,"tag":47,"props":78,"children":80},{"id":79},"profondeur-de-contenu-ajouter-des-couches-plutôt-que-de-raccourcir",[81],{"type":36,"value":82},"Profondeur de Contenu : Ajouter des Couches Plutôt que de Raccourcir",{"type":31,"tag":32,"props":84,"children":85},{},[86],{"type":36,"value":87},"Mauvaise stratégie : garder les événements courts et les lancer souvent. Bonne stratégie : approfondir l'événement et élargir la fenêtre de complétion. Le scénario que nous avons testé en 2025 : événement superficiel de 3 jours (5 paliers, 18 tâches au total) versus événement profond de 5 jours (7 paliers, 32 tâches mais les 3 premiers paliers accessibles aux joueurs occasionnels). L'événement profond a augmenté la rétention D7 de 8 % parce que le joueur se disait « j'ai fini l'événement, mais je vais débloquer le bonus ».",{"type":31,"tag":32,"props":89,"children":90},{},[91],{"type":36,"value":92},"La profondeur de contenu s'organise en 3 couches : (1) piste centrale — complétion accessible pour tous les types de joueurs (cible de complétion >75 %), (2) piste hardcore — paliers étendus pour les joueurs très engagés (complétion 35–40 %), (3) piste monétisation — pallier premium déclenchant les IAP (conversion 4–6 %). Chaque couche a sa propre courbe de récompenses : piste centrale = devises souples + cosmétiques, piste hardcore = jeton gacha + objet exclusif à l'événement, piste monétisation = réduction de bundle + multiplicateur de devises premium limité dans le temps.",{"type":31,"tag":94,"props":95,"children":99},"pre",{"className":96,"code":97,"language":98,"meta":15,"style":15},"language-python shiki shiki-themes github-dark","# Scoring de profondeur événementielle (modèle simplifié)\ncore_completion_rate = 0.78\nhardcore_completion_rate = 0.38\nmonetization_conversion = 0.053\n\ndepth_score = (\n    core_completion_rate * 0.5 +\n    hardcore_completion_rate * 0.3 +\n    monetization_conversion * 100 * 0.2\n)\n# depth_score > 0.65 = sain, \u003C 0.50 = redesign nécessaire\n","python",[100],{"type":31,"tag":69,"props":101,"children":102},{"__ignoreMap":15},[103,115,137,155,173,183,201,225,247,274,283],{"type":31,"tag":104,"props":105,"children":108},"span",{"class":106,"line":107},"line",1,[109],{"type":31,"tag":104,"props":110,"children":112},{"style":111},"--shiki-default:#6A737D",[113],{"type":36,"value":114},"# Scoring de profondeur événementielle (modèle simplifié)\n",{"type":31,"tag":104,"props":116,"children":118},{"class":106,"line":117},2,[119,125,131],{"type":31,"tag":104,"props":120,"children":122},{"style":121},"--shiki-default:#E1E4E8",[123],{"type":36,"value":124},"core_completion_rate ",{"type":31,"tag":104,"props":126,"children":128},{"style":127},"--shiki-default:#F97583",[129],{"type":36,"value":130},"=",{"type":31,"tag":104,"props":132,"children":134},{"style":133},"--shiki-default:#79B8FF",[135],{"type":36,"value":136}," 0.78\n",{"type":31,"tag":104,"props":138,"children":140},{"class":106,"line":139},3,[141,146,150],{"type":31,"tag":104,"props":142,"children":143},{"style":121},[144],{"type":36,"value":145},"hardcore_completion_rate ",{"type":31,"tag":104,"props":147,"children":148},{"style":127},[149],{"type":36,"value":130},{"type":31,"tag":104,"props":151,"children":152},{"style":133},[153],{"type":36,"value":154}," 0.38\n",{"type":31,"tag":104,"props":156,"children":158},{"class":106,"line":157},4,[159,164,168],{"type":31,"tag":104,"props":160,"children":161},{"style":121},[162],{"type":36,"value":163},"monetization_conversion ",{"type":31,"tag":104,"props":165,"children":166},{"style":127},[167],{"type":36,"value":130},{"type":31,"tag":104,"props":169,"children":170},{"style":133},[171],{"type":36,"value":172}," 0.053\n",{"type":31,"tag":104,"props":174,"children":176},{"class":106,"line":175},5,[177],{"type":31,"tag":104,"props":178,"children":180},{"emptyLinePlaceholder":179},true,[181],{"type":36,"value":182},"\n",{"type":31,"tag":104,"props":184,"children":186},{"class":106,"line":185},6,[187,192,196],{"type":31,"tag":104,"props":188,"children":189},{"style":121},[190],{"type":36,"value":191},"depth_score ",{"type":31,"tag":104,"props":193,"children":194},{"style":127},[195],{"type":36,"value":130},{"type":31,"tag":104,"props":197,"children":198},{"style":121},[199],{"type":36,"value":200}," (\n",{"type":31,"tag":104,"props":202,"children":204},{"class":106,"line":203},7,[205,210,215,220],{"type":31,"tag":104,"props":206,"children":207},{"style":121},[208],{"type":36,"value":209},"    core_completion_rate ",{"type":31,"tag":104,"props":211,"children":212},{"style":127},[213],{"type":36,"value":214},"*",{"type":31,"tag":104,"props":216,"children":217},{"style":133},[218],{"type":36,"value":219}," 0.5",{"type":31,"tag":104,"props":221,"children":222},{"style":127},[223],{"type":36,"value":224}," +\n",{"type":31,"tag":104,"props":226,"children":228},{"class":106,"line":227},8,[229,234,238,243],{"type":31,"tag":104,"props":230,"children":231},{"style":121},[232],{"type":36,"value":233},"    hardcore_completion_rate ",{"type":31,"tag":104,"props":235,"children":236},{"style":127},[237],{"type":36,"value":214},{"type":31,"tag":104,"props":239,"children":240},{"style":133},[241],{"type":36,"value":242}," 0.3",{"type":31,"tag":104,"props":244,"children":245},{"style":127},[246],{"type":36,"value":224},{"type":31,"tag":104,"props":248,"children":249},{"class":106,"line":25},[250,255,259,264,269],{"type":31,"tag":104,"props":251,"children":252},{"style":121},[253],{"type":36,"value":254},"    monetization_conversion ",{"type":31,"tag":104,"props":256,"children":257},{"style":127},[258],{"type":36,"value":214},{"type":31,"tag":104,"props":260,"children":261},{"style":133},[262],{"type":36,"value":263}," 100",{"type":31,"tag":104,"props":265,"children":266},{"style":127},[267],{"type":36,"value":268}," *",{"type":31,"tag":104,"props":270,"children":271},{"style":133},[272],{"type":36,"value":273}," 0.2\n",{"type":31,"tag":104,"props":275,"children":277},{"class":106,"line":276},10,[278],{"type":31,"tag":104,"props":279,"children":280},{"style":121},[281],{"type":36,"value":282},")\n",{"type":31,"tag":104,"props":284,"children":286},{"class":106,"line":285},11,[287],{"type":31,"tag":104,"props":288,"children":289},{"style":111},[290],{"type":36,"value":291},"# depth_score > 0.65 = sain, \u003C 0.50 = redesign nécessaire\n",{"type":31,"tag":32,"props":293,"children":294},{},[295],{"type":36,"value":296},"Résultat du test : les événements avec un depth_score de 0,71 affichent un taux de churn 12 % meilleur que les événements superficiels avec un score de 0,68. Le joueur tire des niveaux d'engagement différents d'un seul événement, sans saturer le calendrier.",{"type":31,"tag":47,"props":298,"children":300},{"id":299},"équilibre-monétisation-rétention-timing-des-iap-et-structure-événementielle",[301],{"type":36,"value":302},"Équilibre Monétisation-Rétention : Timing des IAP et Structure Événementielle",{"type":31,"tag":32,"props":304,"children":305},{},[306],{"type":36,"value":307},"Les événements de monétisation agressive (paywall dur, bundle IAP limité dans le temps) augmentent l'ARPU de 23 % à court terme mais relèvent le churn D14 de 19 %. Les joueurs non-payants pensent « cet événement n'est pas pour moi » et se désabonnent silencieusement. L'approche équilibrée : chaque événement a une structure hybride — l'IAP est optionnel mais le non-payant a un chemin de progression alternatif.",{"type":31,"tag":32,"props":309,"children":310},{},[311],{"type":36,"value":312},"Le timing des IAP est critique : plutôt qu'une offre agressive dès le début, une prompt IAP douce au point médian de l'événement (quand le joueur est déjà engagé) convertit 34 % mieux. Ne pas afficher d'IAP pendant les 36 premières heures augmente la rétention de 7 % parce que le joueur expérimente d'abord la piste centrale, puis prend la décision « accélérons ».",{"type":31,"tag":314,"props":315,"children":316},"table",{},[317,346],{"type":31,"tag":318,"props":319,"children":320},"thead",{},[321],{"type":31,"tag":322,"props":323,"children":324},"tr",{},[325,331,336,341],{"type":31,"tag":326,"props":327,"children":328},"th",{},[329],{"type":36,"value":330},"Structure Événementielle",{"type":31,"tag":326,"props":332,"children":333},{},[334],{"type":36,"value":335},"Rétention D7",{"type":31,"tag":326,"props":337,"children":338},{},[339],{"type":36,"value":340},"ARPU (7 jours)",{"type":31,"tag":326,"props":342,"children":343},{},[344],{"type":36,"value":345},"Taux de Churn",{"type":31,"tag":347,"props":348,"children":349},"tbody",{},[350,374,397],{"type":31,"tag":322,"props":351,"children":352},{},[353,359,364,369],{"type":31,"tag":354,"props":355,"children":356},"td",{},[357],{"type":36,"value":358},"IAP Agressif (heure 0)",{"type":31,"tag":354,"props":360,"children":361},{},[362],{"type":36,"value":363},"61 %",{"type":31,"tag":354,"props":365,"children":366},{},[367],{"type":36,"value":368},"1,84 $",{"type":31,"tag":354,"props":370,"children":371},{},[372],{"type":36,"value":373},"29 %",{"type":31,"tag":322,"props":375,"children":376},{},[377,382,387,392],{"type":31,"tag":354,"props":378,"children":379},{},[380],{"type":36,"value":381},"IAP Point Médian (36e heure)",{"type":31,"tag":354,"props":383,"children":384},{},[385],{"type":36,"value":386},"68 %",{"type":31,"tag":354,"props":388,"children":389},{},[390],{"type":36,"value":391},"1,71 $",{"type":31,"tag":354,"props":393,"children":394},{},[395],{"type":36,"value":396},"23 %",{"type":31,"tag":322,"props":398,"children":399},{},[400,405,410,415],{"type":31,"tag":354,"props":401,"children":402},{},[403],{"type":36,"value":404},"Hybride (piste centrale gratuite, bonus IAP)",{"type":31,"tag":354,"props":406,"children":407},{},[408],{"type":36,"value":409},"71 %",{"type":31,"tag":354,"props":411,"children":412},{},[413],{"type":36,"value":414},"1,65 $",{"type":31,"tag":354,"props":416,"children":417},{},[418],{"type":36,"value":419},"19 %",{"type":31,"tag":32,"props":421,"children":422},{},[423],{"type":36,"value":424},"Le modèle hybride est optimal : le non-payant termine 78 % de la piste centrale et reste engagé, le payant termine 41 % de la piste premium et maintient l'ARPU. Le churn se stabilise à 19 %.",{"type":31,"tag":47,"props":426,"children":428},{"id":427},"ciblage-par-cohorte-un-seul-calendrier-ne-suffit-pas-cadence-segmentée",[429],{"type":36,"value":430},"Ciblage par Cohorte : Un Seul Calendrier ne Suffit Pas, Cadence Segmentée",{"type":31,"tag":32,"props":432,"children":433},{},[434],{"type":36,"value":435},"Tous les joueurs ne devraient pas être sur le même calendrier événementiel. Les nouveaux utilisateurs (J0–J7) reçoivent des événements adaptés à l'onboarding, les joueurs engagés (J30+) des événements haute difficulté, les joueurs inactifs (0 session les 7 derniers jours) des événements de retour. Trois calendriers événementiques distincts tournent simultanément pour trois cohortes différentes.",{"type":31,"tag":32,"props":437,"children":438},{},[439],{"type":36,"value":440},"Mesurer le ciblage par cohorte : taux de churn spécifique au segment. Lancer un événement onboarding-friendly pour la cohorte J0–J7 baisse le churn de 16 % à 11 % parce que le joueur expérimente naturellement « je comprends la boucle de jeu, maintenant j'essaie l'événement ». Lancer un événement classé saisonnier plutôt qu'un événement standard pour la cohorte J30+ augmente la rétention de 9 % — le joueur a déjà complété la boucle principale et recherche un nouveau défi.",{"type":31,"tag":32,"props":442,"children":443},{},[444],{"type":36,"value":445},"Les événements de retour win-back ciblent le segment le plus sensible : joueurs avec 0 session depuis 7–14 jours. Une notification push générique « reviens » convertit à 2,3 %, tandis qu'un événement personnalisé (« un skin exclusif pour ton personnage préféré ») convertit à 8,1 %. Adapter l'événement à la cohorte est clé : tutoriel pour J0–J7, défi méta pour J30+, nostalgie pour les inactifs.",{"type":31,"tag":94,"props":447,"children":451},{"className":448,"code":449,"language":450,"meta":15,"style":15},"language-sql shiki shiki-themes github-dark","-- Attribution d'événement par cohorte (exemple PostgreSQL)\nSELECT \n    user_id,\n    CASE \n        WHEN day_since_install BETWEEN 0 AND 7 THEN 'onboarding_event'\n        WHEN day_since_install >= 30 AND last_session_gap \u003C 2 THEN 'hardcore_event'\n        WHEN last_session_gap BETWEEN 7 AND 14 THEN 'winback_event'\n        ELSE 'standard_event'\n    END AS assigned_event\nFROM user_cohort_table\nWHERE active_status = true;\n","sql",[452],{"type":31,"tag":69,"props":453,"children":454},{"__ignoreMap":15},[455,463,476,484,496,540,589,626,639,657,670],{"type":31,"tag":104,"props":456,"children":457},{"class":106,"line":107},[458],{"type":31,"tag":104,"props":459,"children":460},{"style":111},[461],{"type":36,"value":462},"-- Attribution d'événement par cohorte (exemple PostgreSQL)\n",{"type":31,"tag":104,"props":464,"children":465},{"class":106,"line":117},[466,471],{"type":31,"tag":104,"props":467,"children":468},{"style":127},[469],{"type":36,"value":470},"SELECT",{"type":31,"tag":104,"props":472,"children":473},{"style":121},[474],{"type":36,"value":475}," \n",{"type":31,"tag":104,"props":477,"children":478},{"class":106,"line":139},[479],{"type":31,"tag":104,"props":480,"children":481},{"style":121},[482],{"type":36,"value":483},"    user_id,\n",{"type":31,"tag":104,"props":485,"children":486},{"class":106,"line":157},[487,492],{"type":31,"tag":104,"props":488,"children":489},{"style":127},[490],{"type":36,"value":491},"    CASE",{"type":31,"tag":104,"props":493,"children":494},{"style":121},[495],{"type":36,"value":475},{"type":31,"tag":104,"props":497,"children":498},{"class":106,"line":175},[499,504,509,514,519,524,529,534],{"type":31,"tag":104,"props":500,"children":501},{"style":127},[502],{"type":36,"value":503},"        WHEN",{"type":31,"tag":104,"props":505,"children":506},{"style":121},[507],{"type":36,"value":508}," day_since_install ",{"type":31,"tag":104,"props":510,"children":511},{"style":127},[512],{"type":36,"value":513},"BETWEEN",{"type":31,"tag":104,"props":515,"children":516},{"style":133},[517],{"type":36,"value":518}," 0",{"type":31,"tag":104,"props":520,"children":521},{"style":127},[522],{"type":36,"value":523}," AND",{"type":31,"tag":104,"props":525,"children":526},{"style":133},[527],{"type":36,"value":528}," 7",{"type":31,"tag":104,"props":530,"children":531},{"style":127},[532],{"type":36,"value":533}," THEN",{"type":31,"tag":104,"props":535,"children":537},{"style":536},"--shiki-default:#9ECBFF",[538],{"type":36,"value":539}," 'onboarding_event'\n",{"type":31,"tag":104,"props":541,"children":542},{"class":106,"line":185},[543,547,551,556,561,565,570,575,580,584],{"type":31,"tag":104,"props":544,"children":545},{"style":127},[546],{"type":36,"value":503},{"type":31,"tag":104,"props":548,"children":549},{"style":121},[550],{"type":36,"value":508},{"type":31,"tag":104,"props":552,"children":553},{"style":127},[554],{"type":36,"value":555},">=",{"type":31,"tag":104,"props":557,"children":558},{"style":133},[559],{"type":36,"value":560}," 30",{"type":31,"tag":104,"props":562,"children":563},{"style":127},[564],{"type":36,"value":523},{"type":31,"tag":104,"props":566,"children":567},{"style":121},[568],{"type":36,"value":569}," last_session_gap ",{"type":31,"tag":104,"props":571,"children":572},{"style":127},[573],{"type":36,"value":574},"\u003C",{"type":31,"tag":104,"props":576,"children":577},{"style":133},[578],{"type":36,"value":579}," 2",{"type":31,"tag":104,"props":581,"children":582},{"style":127},[583],{"type":36,"value":533},{"type":31,"tag":104,"props":585,"children":586},{"style":536},[587],{"type":36,"value":588}," 'hardcore_event'\n",{"type":31,"tag":104,"props":590,"children":591},{"class":106,"line":203},[592,596,600,604,608,612,617,621],{"type":31,"tag":104,"props":593,"children":594},{"style":127},[595],{"type":36,"value":503},{"type":31,"tag":104,"props":597,"children":598},{"style":121},[599],{"type":36,"value":569},{"type":31,"tag":104,"props":601,"children":602},{"style":127},[603],{"type":36,"value":513},{"type":31,"tag":104,"props":605,"children":606},{"style":133},[607],{"type":36,"value":528},{"type":31,"tag":104,"props":609,"children":610},{"style":127},[611],{"type":36,"value":523},{"type":31,"tag":104,"props":613,"children":614},{"style":133},[615],{"type":36,"value":616}," 14",{"type":31,"tag":104,"props":618,"children":619},{"style":127},[620],{"type":36,"value":533},{"type":31,"tag":104,"props":622,"children":623},{"style":536},[624],{"type":36,"value":625}," 'winback_event'\n",{"type":31,"tag":104,"props":627,"children":628},{"class":106,"line":227},[629,634],{"type":31,"tag":104,"props":630,"children":631},{"style":127},[632],{"type":36,"value":633},"        ELSE",{"type":31,"tag":104,"props":635,"children":636},{"style":536},[637],{"type":36,"value":638}," 'standard_event'\n",{"type":31,"tag":104,"props":640,"children":641},{"class":106,"line":25},[642,647,652],{"type":31,"tag":104,"props":643,"children":644},{"style":127},[645],{"type":36,"value":646},"    END",{"type":31,"tag":104,"props":648,"children":649},{"style":127},[650],{"type":36,"value":651}," AS",{"type":31,"tag":104,"props":653,"children":654},{"style":121},[655],{"type":36,"value":656}," assigned_event\n",{"type":31,"tag":104,"props":658,"children":659},{"class":106,"line":276},[660,665],{"type":31,"tag":104,"props":661,"children":662},{"style":127},[663],{"type":36,"value":664},"FROM",{"type":31,"tag":104,"props":666,"children":667},{"style":121},[668],{"type":36,"value":669}," user_cohort_table\n",{"type":31,"tag":104,"props":671,"children":672},{"class":106,"line":285},[673,678,683,687],{"type":31,"tag":104,"props":674,"children":675},{"style":127},[676],{"type":36,"value":677},"WHERE",{"type":31,"tag":104,"props":679,"children":680},{"style":121},[681],{"type":36,"value":682}," active_status ",{"type":31,"tag":104,"props":684,"children":685},{"style":127},[686],{"type":36,"value":130},{"type":31,"tag":104,"props":688,"children":689},{"style":121},[690],{"type":36,"value":691}," true;\n",{"type":31,"tag":32,"props":693,"children":694},{},[695,697,706],{"type":36,"value":696},"La segmentation des cohortes peut aussi s'aligner avec les résultats de test créatifs ",{"type":31,"tag":698,"props":699,"children":703},"a",{"href":700,"rel":701},"https:\u002F\u002Fwww.roibase.com.tr\u002Ffr\u002Faso",[702],"nofollow",[704],{"type":36,"value":705},"App Store Optimization",{"type":36,"value":707}," : si un ensemble créatif affiche un IPM élevé, lancer un événement sur un thème similaire pour une cohorte comparable augmente le LTV de 11 %.",{"type":31,"tag":47,"props":709,"children":711},{"id":710},"engineering-du-calendrier-simulation-du-calendrier-avec-modèle-de-rétention",[712],{"type":36,"value":713},"Engineering du Calendrier : Simulation du Calendrier avec Modèle de Rétention",{"type":31,"tag":32,"props":715,"children":716},{},[717],{"type":36,"value":718},"Le calendrier live ops ne doit plus être manuel — il doit être guidé par un modèle de prédiction de churn. On simule le brouillon du calendrier événementique 12 semaines en avant : on projette le taux de complétion, la fenêtre de chevauchement et l'impact de la spike monétisation sur la courbe de rétention par cohorte. Output du modèle : 12 semaines de calendrier produisent une rétention D30 attendue de 68,4 % et un churn de 21,7 %.",{"type":31,"tag":32,"props":720,"children":721},{},[722],{"type":36,"value":723},"Les inputs de simulation sont : (1) performance historique des événements (taux de complétion, lift de sessions, delta ARPU), (2) distribution des cohortes (J0–J7 = 34 %, J8–J29 = 41 %, J30+ = 25 %), (3) seuil de tolérance au chevauchement (40 %). L'output du modèle : « la semaine 8 aura 52 % de chevauchement d'événements, la rétention chutera de 5 % » — alerte précoce.",{"type":31,"tag":32,"props":725,"children":726},{},[727],{"type":36,"value":728},"Optimisation du calendrier par itération : si les résultats de simulation sont mauvais, on ajuste manuellement les semaines problématiques — décaler un événement de 2 jours, augmenter la profondeur de contenu de 15 %, modifier le timing des IAP. On simule à nouveau. Après 3–4 itérations, le calendrier optimal émerge : rétention D30 de 12 semaines = 72,1 %, churn = 18,3 % (18 % plus bas que la baseline).",{"type":31,"tag":32,"props":730,"children":731},{},[732],{"type":36,"value":733},"L'engineering du calendrier live ops transforme la rétention d'une tactique manuelle en un problème d'architecture de données. La cadence événementique, la profondeur de contenu, le timing des IAP et la segmentation des cohortes sont tous des inputs numériques — le modèle les équilibre et réduit le churn. Le joueur ressent « il y a toujours quelque chose de nouveau, mais sans me surcharger », et le jeu se maintient à >70 % rétention D30, au-dessus des benchmarks tier-1.",{"type":31,"tag":735,"props":736,"children":737},"style",{},[738],{"type":36,"value":739},"html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}",{"title":15,"searchDepth":139,"depth":139,"links":741},[742,743,744,745,746],{"id":49,"depth":117,"text":52},{"id":79,"depth":117,"text":82},{"id":299,"depth":117,"text":302},{"id":427,"depth":117,"text":430},{"id":710,"depth":117,"text":713},"markdown","content:fr:gaming:calendrier-live-ops-retention-engineering-churn.md","content","fr\u002Fgaming\u002Fcalendrier-live-ops-retention-engineering-churn.md","fr\u002Fgaming\u002Fcalendrier-live-ops-retention-engineering-churn","md",1782079501491]