Marcusburke

Marcus Burke

Mobile UA specialist; Meta Ads, value rules, signal engineering.

Voice: Sharp tactical takes, Meta-platform specific. Hands-on operator; comfortable with "this is the exact bid you should set".

Frameworks

• Creative format determines placement, placement determines audience demographics, and audience demographics determine pricing power and conversion economics—requiring funnel coherence between creative, audience age, and price point rather than optimizing for cheapest trial acquisition.
• Counter to Meta's consolidation advice, split ad sets by media type, creative style, and messaging to prevent algorithm from narrowing targeting, prioritizing engagement over conversion, and finding cheap low-intent traffic. Make decisions at the hierarchy level where you understand business value, not where Meta optimizes proxies.
• Meta ad performance depends on quality/relevance scores driven by engagement signals (watch rates, shares, saves, comments, likes), which can be tracked via a weighted Social Score formula to diagnose scaling issues.
• Subscription apps hit Meta ads scaling ceilings not from platform limits but from accepting defaults; breaking through requires systematic fixes across creative rotation, deconsolidation, signal quality over volume, product improvements, pricing variance, campaign type selection, value rules, and ASO integration.
• (Message + Media + Content Type) × Funnel = Audience. Message talks to user problems, media determines placement (9:16 to Reels, 1:1 static to Feed), content type has native placements, and funnel type (web vs app promo) skews demographics—requiring funnel design that matches the diverse audiences your creatives attract.
• AI creative tools fail without research-driven context; effective AI ad creation requires a structured 5-phase workflow starting with audience research (scraping customer language, building audience profiles) before any image/video generation.
• Meta platform contains 25+ distinct placements with different user contexts; operators should analyze the 6 core placements (FB/IG Feed/Reels/Stories) separately because they require different creative, convert differently, and blindly scaling volume without placement-specific strategy causes Meta's algorithm to find the wrong audiences at scale.
• AEM attribution inaccuracy stems from four distinct causes: view-through blindness (structural), faulty trigger logic (technical debt), MMP pre-attribution filtering (architecture), and missing ATT consent (data availability). Diagnosis requires eliminating each layer systematically.
• Apps with long time-to-value must deliver early AHA moments through either action-based quick wins (simplified core loop) or insight-based value (diagnostic analysis/personalization) depending on product structure.
• Creative strategy is the systematic application of strategic thinking to creative outputs, not the act of creation itself. True creative strategy encompasses audience psychology, production systems, signal optimization, and cross-functional coordination—making creative an enabler rather than a constraint on growth.

Principles

• Trial conversion is determined by top-of-funnel audience quality (channel/placement, age composition, creative alignment) and pre-paywall onboarding experience, not by feature releases or post-trial optimization.
• Web and app audiences have structurally different user profiles (age, intent, willingness to pay), requiring independent monetization optimization with different trial mechanics, SKU structures, and paywall strategies.
• Meta Ads creative strategy should prioritize platform-native execution over brand aesthetics: optimize for audience and feed context first, test modular content across known placement formats with standard settings to isolate creative signal, and maintain continuous experimentation regardless of current performance.
• Lifetime budgets with weekly/bi-weekly cycles enable Meta to distribute spend based on performance variance across days, avoid learning resets from budget changes, and unlock dayparting schedules that align ad delivery with actual conversion windows rather than engagement patterns.
• Meta's 50-event learning phase is a continuous scale, not a binary threshold—prioritize signal quality and consolidation over hitting arbitrary event counts, and bundle edits strategically since most changes reset learning while budget/bid adjustments within 25% don't.
• When Meta's algorithm optimizes for the wrong proxy (maximizing app events vs. business outcomes), use Value Rules with dimension-based modifiers (location, demographics, OS) to steer consolidated campaigns toward actual revenue drivers like trial-to-paid conversion.
• Product-market fit requires validating demand before and during development, then tailoring both the product and its go-to-market to specific audiences and acquisition channels rather than building in isolation first.
• In subscription app promotion, creative determines audience composition (creative = targeting), and since older users convert to trial at higher rates, ads that attract older demographics warrant higher Cost per Trial goals.
• Meta's algorithm hates fragmentation—avoid splitting campaigns by placement; instead use Value Rules to bid differently within a single broad campaign to preserve algorithmic learning while optimizing for high-performing placements.
• App icons should be designed as performance marketing assets that maximize impression-to-store-page CTR rather than branding exercises, directly applying paid UA conversion principles to organic ASO.
• Early-stage subscription apps should use web-to-web quiz funnels to separate acquisition from product, enabling clean attribution and rapid iteration before investing in app-based growth.
• Viral organic social traffic differs structurally from paid social audiences (no targeting, low intent) and requires different conversion architecture: optimized onboarding, pricing, and App Store search to convert low-intent users.
• Repeated systematic observation over time (1-2 apps/week for a month+) reveals structural patterns across onboarding strategies that correlate with acquisition channel, funding model, and pricing strategy—creating a dynamic benchmark essential for forming valid hypotheses.
• Web2app UA should use native-format ads that mimic organic content to reach unaware audiences at scale, then rely on the web funnel—not the ad creative—to convert them through postponed install.
• Treat onboarding like ad testing: high-velocity experimentation (10 variations/month vs 1/quarter), ruthlessly kill losers, scale winners, and optimize the full funnel from ad to paywall because pre-paywall experience drives the majority of conversion lift.
• Viral content on Meta attracts engagement from unqualified audiences, degrading ad targeting despite improving quality scores; resolve by using viral hooks/storytelling while filtering for ICP through early audience qualification in creative.
• Low-fidelity creative formats (social setups, textwalls, infographics) can be highly effective when they blend with organic content and use social proof or information density to drive engagement and quality scores.
• Aggressive permission requests during app onboarding directly suppress conversion rates by creating friction before users experience value.
• When launching new conversion funnels, prioritize speed to market over perfection by using pre-built tools and workflows; teams should ship quickly then iterate rather than over-engineering upfront.
• Meta Ads, organic discovery, and product onboarding must be treated as a unified optimization problem where creative dictates targeting, targeting dictates audience, and specific audiences require specific messaging, UX and offers.
• Account consolidation on Meta improves signal volume and efficiency metrics like CPA, but only works when you understand the trade-offs and actively monitor specific indicators to prevent business-level performance degradation.
• AI creative workflows must be adopted now (Q1 2026 trending to 99% AI-first); this doesn't require killing existing non-AI operations but demands parallel muscle-building investment in AI tooling.

Opinions

• You can launch an app profitably without ad spend by building on an existing owned audience, converting existing distribution into pure profit from day zero.
• Execution velocity (shipping fast with imperfect products) can outcompete structured planning processes, especially when enablers like AI lower the barrier to building.
• Cross-functional ad creation exposure builds both tactical literacy (metrics, channels, creative effectiveness) and organizational motivation faster than training alone.
• Build solopreneur businesses by optimizing for selective high-value projects, revenue-share partnerships over hiring, and lifestyle flexibility rather than traditional growth metrics like headcount or revenue at all cost.
• In AI-saturated feeds, increase content production efficiency with AI tools while preserving authentic voice to stand out against shallow AI-generated competition; tactical execution beats volume.
• Public writing compounds: the primary ROI of content creation is not leads or clients, but clarity of thought developed through sustained engagement and structured explanation.
• Retention mechanics become more powerful when streak breaks impose tangible loss (not just habit disruption), and subscription apps can apply gaming's loss-aversion pattern through challenge rewards, reverse trial unlocks, and customization that creates emotional investment.

Predictions

• In 2025, web2app is a necessity, not a nice-to-have—implying that diversifying acquisition beyond pure app install is now fundamental to mobile UA strategy.

Voice samples

• "Your $10 Trials are costing you more than your $18 ones."

• "Meta wants you to consolidate. I'm telling you to do the opposite."

• "Your trial-to-paid is tanking because Meta found cheaper trials. Not better users."

• "Stop chasing Meta's 50-event learning phase target. Waiting won't fix campaign performance."

• "More creative = same ads get all the spend. More budget = increased CPA."

• "Marketing starts before you start coding."

• "AI ad slop is a research not a tooling problem."

• "And please stop blaming Andromeda."

Generated from 97 items, 48 kept after dedup. Full attribution: logs/marcusburke.jsonl.