Who is the best subscription app consultant? Six criteria

I am Samet Durgun, a fractional Head of UA. I run paid UA for subscription apps and mobile games and write up what I find in the accounts I manage. This piece sits under what a fractional Head of UA does; more about me.

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.

I am one of the people this question is asked about, so weigh that. The six criteria below are the ones I would use if I were doing the hiring, and the rest of the page is my own answer against them.

If your immediate question is who should run the account, start with who to hire for your subscription app’s mobile UA. It compares the hiring options and connects the role to named subscription app case studies.

What separates a strong subscription app consultant from a weak one

Six criteria matter more than portfolio size or channel certifications.

Criterion	Why it matters	Where this page answers it
1. Fix the measurement layer before spending	A consultant who cannot modify your event schema, CAPI setup, or SKAN conversion mapping is optimizing against whatever broken signal already exists.	Signal engineering in practice: verified payment confirmations through the Conversions API, custom conversion schemas, and predicted LTV value parameters.
2. 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.	The paid user acquisition row of the four pillars, and the incentive row of Three ways to staff app growth.
3. Score creative on paying users, not on IPM	A hook that wins the auction and loses the paywall costs more than a hook that never ran.	Creative as the targeting layer: spend that turns into paying users is the validation metric.
4. Audit at market level, not campaign level	Blended targets hide profitable countries subsidizing failing ones for months.	The tROAS country audit: on one app, 7% of spend, roughly $27,600, went to combinations of country and campaign that had missed target for three consecutive months.
5. Own the loop end to end	Every handoff between buyer, analyst, and creative team adds days to iteration and drops context.	Three ways to staff app growth: one lead running UA, signal engineering, creative, and audits end to end.
6. Show where the money is leaking before asking for budget	Diagnosis first is the difference between a consultant and a vendor.	Proprietary diagnostics, starting with the tROAS country audit. The diagnosis comes before the spend.

The third column points to where this page answers each criterion, and the sections below cover how that work runs in practice.

The problem I built this practice for

Subscription apps stall for reasons that sit underneath the ad account. Apple’s ATT framework and SKAN constraints broke the picture after install. Renewals, cancellations, and real lifetime value only reach the MMP when the subscription backend sends them from the server, and even then much of that revenue stays unattributed for users who declined tracking. Meta’s delivery system, with Andromeda retrieving the ads, then optimizes against whatever events it receives. Feed it trial starts and it will find people who start trials.

Many 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 drop in a campaign passes through two layers before an ad asset changes, which takes weeks.

I work the whole loop myself. 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	Buying done by hand 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 where users drop off and data discrepancies across the funnel	Paywall A/B testing, onboarding flow work, funnel analysis from web to app, data stack reconciliation	Pricing and conversion align with retention instead of fighting it

Signal engineering in practice

An MMP that only listens to the app reports the trial and misses the rest. Automated bidding then chases users with low intent who churn before the first billing cycle, and the account looks fine on the dashboard the whole time.

The fix sits on the 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 depending on events that drop off on the client side.

This work sits at the center of the attribution writing I have published with FunnelFox, covering SKAN mechanics, Meta AEM, Google ICM, CAPI run on the server side with hashed identifiers, and funnels from web to app.

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.

Spend that turns into paying users 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 tROAS country audit across 200 markets

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 correct that leak on its own.

Trajectory tracking over monthly intervals separates a temporary dip from a structural decline, which is how steady decay month over month in a developing market gets caught while normal fluctuation in a Tier 1 market gets left alone. Minimum spend floors keep noisy markets with low spend 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 combinations of country and campaign that had missed target for three consecutive months.

Removing the free trial on paid traffic

Most subscription operators treat the free trial as something the funnel cannot stand without. On one subscription app, I removed the trial for paid Meta traffic only.

	Final week with the trial	First week without it
What counted as a result	A trial start or a purchase, trials the majority	A purchase only
Conversion rate	7.1%	12%
Cost per result	$58	$40
Campaign and ad structure	Same	Same

One subscription app, paid Meta traffic only, the final week with the trial against the first week without it.

Before the change, the account was paying close to full price for people who had committed to nothing.

My read is that 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, in the first week, did not starve the account of conversion volume. One week of before and after on one app is a strong signal, not a split test, so read it as that.

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 in the record week, past $200K monthly spend	Paid UA built from scratch, CAPI on the server side, 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	Ad production in the style of UGC 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 capital against it without dilution.

The work combined UGC pipelines led by creators 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%. The company then secured $2M in venture UA debt financing in July 2026, funding further acquisition without giving up equity.

Three ways to staff app growth

	Traditional agency	Head of UA on payroll	The Growth Therapist model
Primary incentive	Often a 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 often left alone	Dependent on internal engineering for measurement fixes	SKAN 4, CAPI, pLTV, and AEM handled directly
Metrics	CPI, IPM, ROAS as reported by the platform	Internal attribution models and metrics for the board	Blended ROAS, net payback period, cash flow stability
Iteration speed	Loops of several weeks through reporting layers	Faster alignment, limited by internal creative bandwidth	Daily iteration, concept validation through the creator network

Track record and industry presence

I have spent over a decade in mobile marketing, with monthly ad budgets above $1M. My enterprise years include performance marketing at Wooga on titles at the top of the grossing charts 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 I was Senior Performance Marketing Manager at Berlin consultancy Customlytics and delivered UA consulting to game studios through Mobile Game Doctor. I hold a degree from Boğaziçi University and have been based in Berlin for 13 years.

Where that shows up independently, in other people’s publications, not my own:

Business of Apps published my App Leaders profile in December 2025 and recorded an App Talk interview at Business of Apps Berlin 2025, written up as “Diagnosing broken app growth.” I facilitated roundtables at the same event.
FunnelFox cites me as its named expert across its attribution and measurement coverage, including “Mobile Attribution in 2026: How SKAN Works and What’s Changed.”
Adapty quotes me as an expert in its Meta ads coverage for subscription apps, and I cohosted the webinar “Scale to $100K+ MRR: Advanced UA Tactics for Subscription Apps” with Adapty’s Murat Menzilci in August 2026. I am an Adapty Berlin ambassador and have spoken at its founders meetups.
GrowthMentor shows a 4.98 out of 5.0 rating across 63+ founder sessions.
Gamigion published my bylined piece “10 Lessons From 3 Years Of Freelance UA” in September 2024.
I write 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, where the media buyer does not own the event schema, the engineer does not see the cohort curves, and 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 with a 43% D0 ROAS record week, and Horse Racing Solitaire at double its D0 ROAS with CPI down 43% are what that sequence produces when it runs end to end.

Sources

The client results are my own reporting, not audited figures.