The advertising dashboard reports conversions. The app analytics dashboard reports trial starts. The payment system reports new paying customers. The totals do not match, and the team wants one number that proves which campaign worked.
Without a complete attribution setup, that number may not exist. You can still make useful decisions if you separate what each system observes from the conclusions it can support.
Marketing measurement without an MMP starts with reliable business totals, clear funnel definitions and explicitly limited channel claims. An MMP—a mobile measurement partner—can support app attribution and related measurement workflows. Until your required setup is implemented and validated, product events and campaign reports should not be treated as interchangeable records of acquisition.
Brano’s portfolio includes product-oriented businesses such as Heli Technology and Zoviz. For businesses working across websites, apps and paid channels, measurement quality is part of the growth system rather than an extra chart added at the end.
Separate the three reporting layers
Business outcomes
Begin with the system that records the outcome itself: a successful payment, an active subscription, a qualified lead or another agreed business event. Define how refunds, renewals, retries, duplicate customers and cancellations are handled.
This gives you a dependable total for the defined outcome. It does not, by itself, explain which channel caused it.
Product behaviour
Product analytics can describe onboarding, activation, repeated use and trial behaviour when the instrumentation is reliable. These events help identify where users experience value or friction.
Keep user identity and event definitions consistent. PostHog’s documentation explains how events and identification support product analysis. A product event can be accurate while its acquisition source remains unknown.
Channel reporting
Advertising platforms describe delivery and the outcomes they attribute under their own available signals and settings. These reports are useful for operating campaigns, but their conversion totals should not simply be added together as if they were unique business customers.
Record what each platform calls a conversion, which time window applies and whether the number represents a direct observation or a modelled or attributed outcome. Keep that context beside the result.
Use one event dictionary
Before trying to reconcile systems, define the events in plain language:
| Event | A useful operational definition | Common source of confusion |
|---|---|---|
| First open | The tracked initial launch for the app instance or identity used | Treating it as identical to every store download |
| Signup completed | A successful account-creation event | Counting attempts or duplicate submissions |
| Activation | A defined first-value event within a stated window | Substituting any app opening |
| Trial started | A unique successful start of the defined trial | Counting repeated paywall views |
| First paid customer | A unique customer’s first qualifying payment | Mixing renewals and new customers |
| Renewal | A later qualifying payment for an existing relationship | Reporting it as another newly acquired customer |
The exact implementation depends on your product. The goal is that two people reading the same metric mean the same thing. For activation specifically, use the SaaS activation metrics guide.
Build a reconciliation table instead of forcing agreement
Create one table with the reporting period, timezone, source system, event definition, total and known limitations. Differences can come from time windows, identity rules, delayed reporting, refunds or the distinction between attribution and actual transactions.
Investigate those causes before changing the numbers. A report becomes less useful when an inconvenient total is silently replaced to make two dashboards agree.
If all trial starts are being used in a paid-media review, label them as total product trial starts, including paid, organic and unattributed users. They can provide business context, but they should not be presented as paid-attributed trials.
Distinguish operating ratios from acquisition cost
Consider an illustrative month with $1,000 in media spend and 20 first-time paying customers across all acquisition sources.
Dividing $1,000 by 20 produces $50 in media spend per new paying customer across all sources. That can be a useful blended operating ratio. It is not proof that paid media acquired those 20 customers, and it is not fully loaded CAC because it excludes other acquisition costs.
Paid-attributed customer-acquisition cost requires a defensible count of new customers attributed to the paid activity, under stated attribution rules. Fully loaded CAC also requires an agreed cost scope. Even with attribution, it is useful to distinguish attributed outcomes from outcomes that would not have happened without the marketing.
If you cannot justify the denominator, name the ratio descriptively. A precise label is more valuable than a familiar acronym attached to incompatible data.
Use campaign parameters where they can travel reliably
For website campaigns, a consistent UTM naming scheme helps organise source, medium and campaign information. Google’s manual campaign-dimension guidance explains how campaign parameters are represented in Analytics.
An illustrative convention could use lowercase names for utm_source, utm_medium and utm_campaign, with a documented utm_content value for creative variations. Apply the convention consistently and keep a registry so two teams do not use different spellings for the same campaign.
Test the actual path through redirects, landing pages, app stores and app openings. Adding parameters to a URL does not guarantee those values will be available inside an installed app after every route. Store original source information only where the implementation and user permissions support it.
Make channel decisions at the level the evidence supports
You can still operate campaigns using delivery, relevant traffic, measured early events and qualitative market feedback. Explain that those are the signals informing the decision.
For example, you might pause a clearly irrelevant keyword, repair a broken landing route or reduce exposure to a market the product cannot serve reliably. Those actions do not require inventing a complete paid-customer attribution model.
More consequential decisions—such as claiming one channel has the best lifetime value—need stronger evidence. Separate what you know, what you suspect and what the next measurement improvement would resolve.
When evaluating growth with limited resources, the small-budget growth strategy helps choose a narrow learning question rather than demanding certainty from every metric.
Prioritise the measurement implementation
Start with the events needed for the business decisions you actually make. Agree on identity, successful-event triggers, payment reconciliation and the acquisition questions the new setup must answer.
Then test end to end: campaign click, landing experience, installation where relevant, first use, trial and qualifying payment. Include duplicate, retry and cancellation cases. Confirm that the people reading reports understand the implemented windows and definitions.
Platform-specific systems can also form part of the picture. Apple’s AdServices documentation describes attribution functionality for Apple advertising; it should be evaluated for that scope rather than treated as a universal answer for every channel.
An MMP selection should follow the required use cases, not simply a vendor list. Consider the platforms, channels, deep-link journeys, reporting needs and implementation resources involved. Validate the resulting setup before promoting a provisional metric into an executive KPI.
Frequently asked questions
Can we report growth while attribution is incomplete?
Yes. Report reliable product and business totals, show channel operating data separately and state which outcomes include organic or unattributed users. Useful reporting depends on clarity about scope.
Can UTMs replace an MMP?
UTMs can help describe website campaign traffic. They do not automatically reproduce every app attribution, install or deferred-linking workflow. Test the specific journey and select measurement tools for the capabilities it requires.
Should all purchases be counted as conversions from paid media?
No. A total purchase count is a business outcome measure. A paid-attributed count needs a supported attribution rule and the relevant data. Keep both if useful, but label them differently.
Explore Brano’s growth marketing service for a measurement plan tied to the decisions your team needs to make.