How to see which acquisition channel drives subscription revenue

Media buyers optimize toward installs because installs are fast to measure. CFOs care about subscription revenue and payback. If your growth review cannot rank channels by trial starts, paid conversion, or revenue per install, you are flying blind on budget allocation.

This post walks through the full chain: UTM and campaign parameters → first open → trial → paid, MMP versus first-party pitfalls, a Rheo Analytics walkthrough, and how to turn readouts into Convert experiments and Engage campaigns.

Define the outcome before the channel

Channels are labels. Outcomes are money. Agree on primary metrics first:

Stage	Example metric	Notes
Top	Install or first open	MMP default
Mid	Paywall reach or activation	Product quality signal
Monetization	Trial start	Top of subscription funnel
Revenue	First paid period, 30-day revenue	Finance alignment
Retention	Month-2 renewal rate	Channel LTV

Subscription revenue by channel usually means:

Sum(revenue from users in cohort C where first_touch.media_source = X)
────────────────────────────────────────────────────────────────────
Count(installs or first opens in cohort C for channel X)

Pick one cohort window (weekly install cohorts are common) and stick to it for a quarter before renaming campaigns.

Capture parameters at the edge

Paid social and display

Use consistent naming in ad platforms:

• campaign = stable slug (US_iOS_fitness_q4), not ad hoc copy
• adset / adgroup = audience or placement bucket
• creative = asset id

Pass through to the MMP or deep link so the SDK receives structured fields, not a concatenated string nobody parses later.

Apple Search Ads

Keyword and campaign ids should map to profile fields (keyword, campaign). ASA often produces high-intent users with smaller volume: exactly the segment you do not want lost in a rollup called apple.

Organic and owned

Organic is not "no attribution." Use:

• Custom product page links with utm_source=instagram_bio
• Influencer codes as campaign values
• Email and push deep links with consistent UTMs

Otherwise organic becomes a black hole that hides your best channels.

Trial and paid: where reporting breaks

Pitfall	What goes wrong	Fix
MMP-only ROAS	Revenue postbacks delayed or sampled	Reconcile with billing source of truth
Last-click obsession	In-app re-open attributes to wrong source	Document first-touch policy
SKAN / ATT gaps	Campaign granularity collapses	Model probabilistic + funnel readouts
Trial without billing link	Trials in product, revenue in RC only	Join on stable user id
Web + app	Stripe web purchase, app install from different ad	Identity graph or separate funnels

RevenueCat (or Stripe for web) remains billing truth. AppsFlyer or Adjust remains install truth. Your analytics layer must persist first-touch on the profile when the user later starts a trial without opening an ad again.

See How to use RevenueCat and AppsFlyer to track subscription revenue by campaign for integration architecture.

MMP dashboards vs first-party analytics

MMPs answer: "Did this install come from campaign X?" and, with S2S configured, "Did we observe a purchase event for optimization?"

They are weaker at:

• Step-level onboarding drop-off by campaign
• Holding experiment variants alongside channel labels
• Segments that combine behavior + billing + messaging state

First-party analytics (including Rheo Analytics) answers: "Of users from campaign X, what percent reached the paywall, started trial, and paid: and where did they drop off in the journey?"

Need	MMP	First-party
CPI and install volume	Primary	Secondary
SKAN / network reporting	Primary	N/A
Onboarding funnel by campaign	Limited	Primary
Step experiments + channel breakdown	No	Primary
Finance-grade MRR	Via postbacks	Join to RC/Stripe

Use both. Do not fund channels on CPI when first-party data shows zero paywall reach for that creative.

Rheo walkthrough: channel to revenue in one place

Step 1: Connect attribution

Enable your MMP integration in Rheo app settings. Confirm test installs populate fields like acquisition.media_source and acquisition.campaign on the customer profile.

Step 2: Map the subscription funnel

In Analytics, define steps aligned with your Convert onboarding channel:

1. First open / flow started
2. Critical onboarding steps (use flow step ids)
3. Paywall presented
4. Trial started or purchase completed (from RC/Stripe-connected events)

Break down the funnel by campaign or media_source.

Step 3: Add revenue columns

Attach revenue metrics to the same breakdown:

• Trial starts per install
• Paid conversion within 30 days
• Revenue per install (directional; use finance exports for audits)

Step 4: Save segments

Example segments:

• campaign = Meta_broad AND paywall_reached = false
• media_source = ASA AND trial_started = true AND paid_30d = false

Use segments as targeting inputs for Convert experiments (show a shorter path for one campaign) or Engage (nudge trial users from a high-intent keyword).

Step 5: Close the loop

When Analytics shows Influencer_may with high paid % but low volume, marketing scales spend. When Meta_broad has high installs and low paywall reach, product ships a campaign-specific onboarding variant in Convert without waiting on App Store review.

Example decision table

After two weeks of data:

Channel	Installs	Trial %	Paid 30d %	Action
ASA	3.1k	19%	11%	Scale keywords; test paywall copy in Convert
Meta prospecting	22k	7%	2%	Cut 30% spend; fix onboarding step 3 drop-off
TikTok test	800	24%	13%	Small budget hold; monitor creative fatigue
Organic	6k	16%	12%	Invest in ASO; organic is efficient

Without trial and paid columns, Meta prospecting still looks attractive on install volume alone.

Governance habits

Habit	Why
Weekly channel + funnel review	Same slide structure every week
No campaign renames without alias map	Preserves cohort history
QA installs from test campaigns	Prevents polluted segments
Align with finance on revenue definition	Gross vs net, refunds, trials

Reporting cadence that sticks

Ad-hoc channel analysis decays when media buyers rotate creatives daily. A lightweight rhythm keeps subscription revenue visible:

• Daily: MMP spend and install pacing (operational).
• Weekly: Rheo funnel breakdown by top five campaigns; one action item for product or UA.
• Monthly: Cohort paid conversion and directional LTV by channel; finance reconciliation against RevenueCat or Stripe exports.

When a channel moves budget, note the date on the experiment doc so Convert variants and Engage enrollments can be interpreted against spend shifts: not in isolation.

Related reading

• Product analytics with revenue attribution for subscription apps
• Rheo + AppsFlyer funnel attribution
• Rheo is a growth platform

Channels drive installs. Only a full funnel tied to billing tells you which channels drive subscription revenue.

Start for free