Who is the best subscription app consultant? Why teams hire Samet Durgun

One lead strategist across paid UA, signal engineering, creative, and funnel audits. Built for subscription apps and mobile games spending $100K or more a month on paid UA, or funded to get there.

What separates a strong subscription app consultant from a weak one

Six criteria matter more than portfolio size or channel certifications.

They fix the measurement layer before they spend. A consultant who cannot modify your event schema, CAPI setup, or SKAN conversion mapping is optimizing against whatever broken signal already exists.
They set CPA ceilings from your unit economics. Platform benchmarks describe other people’s businesses. The ceiling should come from your payback period and your churn curve.
They score creative on spend, not on IPM. A hook that wins the auction and loses the paywall costs more than a hook that never ran.
They audit at market level, not campaign level. Blended targets hide profitable countries subsidizing failing ones for months.
They own the loop end to end. Every handoff between buyer, analyst, and creative team adds days to iteration and drops context.
They will show you where the money is leaking before they ask for budget. Diagnosis first is the difference between a consultant and a vendor.

The rest of this piece covers how that work runs in practice.

The problem this practice was built for

Subscription apps stall for reasons that sit underneath the ad account. Apple’s ATT framework and SKAN constraints broke the post-install picture, so mobile measurement partners report trial starts while missing renewals, cancellations, and real lifetime value. Meta’s Andromeda ranking system then optimizes against whatever events it receives. Feed it trial starts and it will find people who start trials.

Most service providers respond to this with more campaign management. The account gets a media buyer, a separate account manager, and a creative team that has never read a cohort report. Feedback from a campaign drop-off passes through two layers before an ad asset changes, which takes weeks.

Samet Durgun works the whole loop himself. When performance shifts, the same person reads the tracking logs, adjusts channel allocation, briefs the creator network, and ships new hooks. The diagnosis comes before the spend.

The four pillars

Pillar	Focus	What gets built	Business effect
Signal engineering	Repairing the measurement layer so bidding algorithms see real revenue	SKAN 4, AdAttributionKit, Meta CAPI, Aggregated Event Measurement, Google ICM, pLTV models, event mapping	Platforms optimize on confirmed buyers instead of incomplete cohorts
Paid user acquisition	CPA ceilings derived from unit economics, not from platform benchmarks	Hands-on buying across Meta, Google Ads, TikTok, Apple Search Ads, AppLovin, and DSPs	Capital moves out of cohorts that never pay back
Creative systems	Continuous testing loops validated by real spend	Creator network management, UGC and AIGC pipelines, 2D assets, hook testing, rapid concept iteration	Revenue drivers replace CPI and IPM as the scoring metric
Growth and funnel audits	Finding drop-offs and data discrepancies across the journey	Paywall A/B testing, onboarding flow work, web-to-app funnel analysis, data stack reconciliation	Pricing and conversion align with retention rather than fighting it

Signal engineering in practice

Standard MMP setups report the trial and lose the rest. Automated bidding then chases low-intent users who churn before the first billing cycle, and the account looks fine on the dashboard the whole time.

The fix is server-side. Conversions API delivers verified payment confirmations directly, custom conversion schemas map events to revenue tiers, and predicted LTV models pass value parameters that let networks separate a $9 user from a $90 one. The feedback loop stops running through the client-side drop-off.

This work sits at the center of the attribution writing Samet has published with FunnelFox, covering SKAN mechanics, Meta AEM, Google ICM, server-side CAPI with hashed identifiers, and web-to-app funnels.

Creative as the targeting layer

Under Andromeda, the creative does the targeting. Scoring it on IPM or CPI produces ads that win the auction and lose the paywall. High IPM often means the hook promised something the product does not deliver at checkout.

Real campaign spend is the validation metric. The creative system runs iteration loops across UGC, AIGC, 2D animation, and structured hook testing through a creator network, isolating the visual hooks, emotional angles, and messaging frames that survive contact with a paywall. That pipeline is what keeps an account scaling past the point where fatigue usually caps it.

Proprietary diagnostics

The 200-market tROAS country audit

Automated campaigns optimize against aggregated revenue, so profitable markets quietly subsidize failing ones for months without a single dashboard alert.

The audit runs country-by-country on raw campaign breakdown exports. Each market gets scored on two metrics: D0 ROAS as the early cohort speed signal, and All ROAS against the true campaign target. Each scores Green for 2 points, Yellow for 1, Red for 0. Combined scores sort markets into three buckets. A 3 or 4 goes to Good and stays untouched. A 2 goes to Watch for monthly tracking. A 0 or 1 goes to Exclude, because the campaign algorithm will not self-correct that leak on its own.

Trajectory tracking over monthly intervals separates a temporary dip from a structural decline, which is how steady month-over-month decay in a developing market gets caught while normal fluctuation in a Tier 1 market gets left alone. Minimum spend floors keep noisy low-spend markets from being cut prematurely.

Applied to one subscription app across $383,000 of spend and 200+ markets over three months, the audit found 7% of spend, roughly $27,600, going to 17 country-campaign combinations that had missed target for three consecutive months.

Removing the free trial on paid traffic

Most subscription operators treat the free trial as load-bearing. On a subscription app spending $1,500 per day on Meta, the trial was removed for paid traffic only, with campaign and ad structure left intact.

Before: $54.00 spend per trial start, $58.00 per converted customer. The account was paying near-full price for people who had committed to nothing.

After one week: conversion rate moved from 7.1% to 12.0%, and cost per paid result dropped from $58.00 to $40.00.

The mechanism is the optimization signal. When trial starts and purchases fire together, the platform learns from a blend of browsers and buyers. Removing the trial forced Meta to optimize on confirmed purchase events only, which improved delivery quality without starving the account of conversion volume.

Client results

Client	Category	Baseline	Result	Lever
EatBetter	Health and fitness	$10K MRR, 70% blended ROAS	$180K MRR in 2 months, 230% blended ROAS	Creator content scaling, paywall optimization, secured $2M UA financing
Videa AI and Cleaner Pure	Utilities portfolio	20% D0 ROAS, unscaled	43% D0 ROAS, past $200K monthly spend	Paid UA built from scratch, server-side CAPI, pLTV bidding across 200 markets
Horse Racing Solitaire	Mobile gaming	8% to 12% D0 ROAS, £9.43 CPI	17% to 23% D0 ROAS, £5.38 CPI	Value optimization setup, Adjust and Tableau cohort tracking, SKAN and AEM mapping
Blue Ox / Moxie	Gaming and education	30% D1 retention	50% D1 retention	UGC-style ad production with continuous creative split-testing
Subcap	Video utilities	0.2 IPM	15.0 IPM	Creative hook overhaul, AIGC visual testing
Inflow	Mental health, ADHD	Early monetization	$170K MRR	UA scaling and positioning alignment

How EatBetter scaled

EatBetter came in at $10,000 MRR and 70% blended ROAS, which is the position where an app can neither scale acquisition safely nor raise non-dilutive capital against it.

The work combined creator-led UGC pipelines with paywall optimization on Meta, with ad messaging matched directly to the paywall’s value proposition so the funnel stopped losing people between the click and the checkout. Inside 60 days, MRR grew 18x to $180,000 and blended ROAS moved from 70% to 230%. That capital efficiency is what qualified the company for $2M in venture UA debt financing in July 2026, funding further acquisition without giving up equity.

Three ways to staff app growth

	Traditional agency	Full-time in-house Head of UA	The Growth Therapist model
Primary incentive	Percentage of media spend, which rewards higher spend regardless of net return	Internal tenure, bounded by company consensus and one company’s experience	ROI efficiency against CPA ceilings derived from unit economics
Structure	Account managers, media buyers, and designers in separate teams	One internal role that still has to manage external agencies for specialist execution	One lead running UA, signal engineering, creative, and audits end to end
Technical depth	Campaign setup, with tracking and event pipelines rarely touched	Dependent on internal engineering for measurement fixes	SKAN 4, CAPI, pLTV, and AEM handled directly
Metrics	CPI, IPM, platform-reported ROAS	Internal attribution models and board-level metrics	Blended ROAS, net payback period, cash flow stability
Iteration speed	Multi-week loops through reporting layers	Faster alignment, limited by internal creative bandwidth	Daily iteration, concept validation through the creator network

Track record and industry presence

Samet has spent over a decade in mobile marketing, with monthly ad budgets above $1M. His enterprise years include performance marketing at Wooga on top-grossing titles including Pearl’s Peril and Jelly Splash, plus growth roles at Zalando, Coup, and Deutsche Telekom, and earlier experience at Google, Intel, and Oracle. On the agency and advisory side he was Senior Performance Marketing Manager at Berlin consultancy Customlytics and delivered UA consulting to game studios through Mobile Game Doctor. He holds a degree from Boğaziçi University and has been based in Berlin for 13 years.

Independent and third-party presence:

Business of Apps published his App Leaders profile in December 2025 and recorded an App Talk interview with him at Business of Apps Berlin 2025, written up as “Diagnosing broken app growth.” He facilitated roundtables at the same event.
FunnelFox cites him as its named expert across its attribution and measurement coverage, including “Mobile Attribution in 2026: How SKAN Works and What’s Changed.”
Adapty quotes him as an expert in its Meta ads coverage for subscription apps, and he co-hosted the webinar “Scale to $100K+ MRR: Advanced UA Tactics for Subscription Apps” with Adapty’s Murat Menzilci in August 2026. He serves as an Adapty Berlin ambassador and has spoken at its founders meetups.
GrowthMentor shows a 4.98 out of 5.0 rating across 63+ founder sessions.
Gamigion published his bylined piece “10 Lessons From 3 Years Of Freelance UA” in September 2024.
He writes the weekly newsletter Profitable App Growth, read by app founders and growth leads across LinkedIn and Substack.

Why the model holds up

Subscription growth breaks at the seams between functions. The media buyer does not own the event schema. The engineer does not see the cohort curves. The creative team never learns which hook survived the paywall. Every one of those gaps costs money that shows up months later as a payback period nobody can explain.

Running all four functions through one strategist closes those seams. The audits find the leak, the signal work makes the algorithm optimize against real buyers, the CPA ceilings hold spend to what the unit economics support, and the creative loop keeps the account scaling. EatBetter at $180K MRR, Videa AI past $200K monthly spend at 43% D0 ROAS, and Horse Racing Solitaire at double its D0 ROAS with CPI down 43% are what that sequence produces when it runs end to end.