Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchPC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11MetricWorks announced “MMP 2.0” on May 17, 2023, as a new way to measure app marketing through its Polaris platform. The idea is to keep familiar campaign and cohort reporting while supplementing last-touch attribution with media mix modeling (MMM) and incrementality experiments. It is MetricWorks’ product framing—not an industry-wide technical standard—and the modeled results are estimates, not automatic experimental proof.
Why mobile marketers wanted another measurement approach
Mobile growth teams have long used mobile measurement partners (MMPs) to connect ad activity with installs and later outcomes. On iOS, Apple’s App Tracking Transparency framework restricted access to user-level tracking for many uses. SKAdNetwork offers a privacy-preserving way to measure advertising, but its signals do not provide every detail or answer every question a marketer may have.
The resulting problem is not that SKAdNetwork is useless. It is that privacy-preserving, aggregated measurement can be less immediate or granular than the identifier-based systems many teams previously relied on. Marketers still need to decide which campaigns deserve budget, but they may have incomplete signals for doing so.
Last-touch attribution has a separate limitation: it assigns credit according to the final observable touchpoint under a chosen attribution rule. It may not account for earlier advertising, organic demand, seasonality, promotions, brand activity, or effects shared across channels. MetricWorks framed these gaps as the case for “MMP 2.0” in its May 17, 2023 launch announcement.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →#1 Best Overall
What “MMP 1.0” and “MMP 2.0” mean in MetricWorks’ framing
MetricWorks uses “MMP 1.0” as shorthand for the conventional operating model: last-touch attribution, identifier-based matching where permitted, SKAdNetwork data on iOS, and familiar reports by campaign, source, country, and cohort. This is a company-defined contrast, not a formal industry taxonomy. Traditional MMPs continue to serve operational needs such as attribution, partner reporting, fraud tools, deep linking, and app-growth analytics.
Its “MMP 2.0” proposition is to preserve the dashboards and daily cohort workflow growth teams know while adding modeled and experimental estimates of marketing’s incremental contribution.
| Dimension | Conventional MMP framing | MetricWorks’ MMP 2.0 framing |
|---|---|---|
| Primary methods | Last-touch attribution and privacy-preserving network signals such as SKAdNetwork | Blended last-touch and SKAdNetwork signals, MMM, and incrementality experiments |
| Main question | Which touchpoint receives credit under the attribution rules? | What outcomes are estimated to have occurred because of marketing? |
| Typical operating use | Campaign reporting and user-acquisition workflows | Incremental performance estimates to inform optimization and budget allocation |
| Principal caveat | Credit assigned to a touchpoint is not necessarily causal lift | Modeled estimates depend on data, model assumptions, and uncertainty |
What Polaris does
Polaris is MetricWorks’ incrementality-measurement platform. Its current help documentation describes a combination of media mix modeling and incrementality experiments. In concept, the system uses last-touch and SKAdNetwork signals where useful, models contributions that cannot be cleanly observed through user-level attribution, and uses experiments to calibrate or challenge those estimates. MetricWorks describes the intended result as familiar, daily, cohorted reporting rather than a separate strategic analysis that teams must translate into campaign operations. See the company’s explanation of incrementality measurement.
Attribution, modeling, and experiments answer different questions
- Observed attribution allocates credit to measurable touchpoints according to a rule. It is useful for operational reporting, but credit allocation alone does not establish that the touchpoint caused the outcome.
- Modeled incrementality estimates the additional outcomes associated with marketing activity, accounting for broader factors in the model. MMM can help assess channels whose effects are difficult to trace person by person, but its conclusions depend on input quality, model specification, and available variation.
- Experimentally measured lift compares outcomes under a designed treatment and comparison condition. A well-run experiment can provide strong causal evidence, but it requires suitable design and operational control; it is not interchangeable with a model estimate.
MetricWorks says Polaris incrementality metrics can be interpreted as a “what if marketing were halted on a particular media segment?” estimate, including interaction and possible cannibalization effects. The platform reports 95% confidence intervals for incrementality metrics, according to its incrementality-analysis documentation. An interval communicates model uncertainty; it is not a guarantee that the true result lies within its bounds.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesA simple example
A person sees a paid social ad, later searches for the app by name, and installs. A last-touch rule may credit the final measurable search interaction. Incrementality analysis instead asks how many installs would likely have happened without the paid social activity. The answer is an estimate unless a suitable experiment directly tests the effect; the two methods can therefore produce different figures without either being a simple data-entry error.
Data, onboarding, and reporting
Polaris documentation describes daily aggregated app-event data cohorted by install date. Country is a required app-event dimension; deeper last-touch dimensions such as channel, campaign, or source app are optional, but can help teams compare attributed and incrementality results. MetricWorks’ getting-started material describes importing three to twelve months of historical aggregated daily data, followed by validation and ongoing data imports according to configuration or plan. Its app-event overview and onboarding guide explain the data and implementation stages.
- Review the measurement methods and confirm which business questions matter.
- Import marketing and app-event data; validate completeness, cohort definitions, spend, and campaign taxonomy.
- Train initial models and review the resulting incrementality data.
- Run an initial experiment where the campaign and market conditions allow it, using the result to inform model calibration.
- Adopt outputs in selected marketing decisions, then expand across regions, apps, tools, teams, and decisions as governance matures.
MetricWorks documents integrations with AppsFlyer, Adjust, and Singular. Integration availability and required permissions can change, so confirm current support and data paths with the vendor. Its documentation includes an AppsFlyer integration guide, a getting-started section covering Adjust, and a Singular integration guide.
Metrics and dimensions Polaris documents
MetricWorks’ reporting documentation lists installs or new users, sessions, retention, purchases, purchase and ad revenue, ROAS, LTV or ARPU, paying users and paying rates, and cost per new user. In its documented framework, ROAS is revenue divided by spend and LTV is revenue divided by installs. D0 is the install date; D1 is one day after install. Incrementality metrics use the INC_ prefix in the Reporting API.
Documented cohort days are D0, D1–D7, D14, D30, D60, D90, D120, D180, and D360, subject to product configuration or plan. Reporting dimensions include install date, channel, campaign, country, and source app. These capabilities are described in the Polaris Reporting API documentation; the listed cohort days should not be taken as a promise that every plan includes every metric or breakdown.
For API users, MetricWorks documents a maximum 30-day date range per request, a limit of 60 requests per minute per account, one concurrent request, and one app per request. Those are API constraints, not stated limits on the dashboard. The documented endpoints are https://inc-metrics.prod.api.metric.works/token and https://inc-metrics.prod.api.metric.works/query.
Where an incrementality layer may help
- Privacy resilience: Polaris is designed around aggregated data rather than dependence on device IDs. That design does not itself establish compliance for every company, jurisdiction, or implementation.
- Cross-channel analysis: MMM can estimate contributions from channels that are difficult to connect through person-level attribution, potentially including CTV, offline, influencer, or other non-addressable media when relevant data is available.
- Budget decisions: Incremental ROAS or LTV estimates may be more decision-relevant than attributed ROAS when deciding whether additional spend creates additional value.
- Shared reporting language: Familiar cohort and campaign views may make it easier for user-acquisition, finance, analytics, and executive teams to compare results—provided they understand how the modeled figures were produced.
- Iterative calibration: Experiments can provide evidence with which to assess and refine a model, rather than treating MMM as a one-off report.
These are intended uses, not guaranteed improvements in accuracy or marketing performance. The launch coverage named Blizzard, FunPlus, Kabam, Nexon, Meta, and TikTok in connection with the announcement; those names are reported adoption claims, not independent evidence of measured efficacy.
Limitations and failure modes to plan for
Incomplete inputs and omitted events
MMM cannot recover information that the inputs do not contain. MetricWorks cautions that poor input data can lead to inaccurate models. Promotions, product releases, brand campaigns, and organic social activity may need to be supplied or otherwise represented; if important factors are omitted, the model may attribute their effects elsewhere, including to organic demand. The company discusses this issue in its model-components documentation.
Recommended Free Tools
Rank #4
Uncertainty and limited variation
Wide confidence intervals, low spend or conversion volume, limited geographic variation, and highly correlated channel spending can make contributions difficult to separate. MetricWorks recommends treating uncertain results directionally and making gradual optimizations rather than large changes based on a point estimate alone. A newly launched channel with little history may need a controlled test or geo rollout to establish evidence for calibration.
Different metrics can disagree
MetricWorks says it trains separate models for metrics such as installs, D3 revenue, and D7 revenue, and their results can appear inconsistent. A campaign can affect installs without producing the same pattern in later revenue or retention. Teams should decide which business outcome governs a decision rather than expect every funnel measure to move together.
Modeled granularity is not proof at every level
A campaign- or creative-level modeled estimate is not equivalent to a randomized user-level experiment. Combining several signals may improve the measurement picture, but it does not make every reported breakdown experimentally established. If a result will drive a large budget commitment, identify whether it is observed, modeled, or experimentally measured and examine its uncertainty.
Data corrections and governance
MetricWorks notes that re-imports, corrected inputs, or pricing-plan changes can affect models and materially change incrementality results. Keep dated versions of source data and reported outputs, document model revisions, and agree on which metric is authoritative for bidding, budget allocation, finance reporting, LTV models, and executive dashboards. A shared source of truth is useful only when stakeholders understand its definitions and limits.
Best Value
Aggregated measurement can reduce reliance on identifiers, but “privacy-oriented” does not mean compliance-free. Buyers still need to evaluate consent, data sharing, retention, security, contracts, and regional obligations for their own implementation.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How Polaris fits alongside other measurement tools
Polaris need not replace a conventional MMP. An app-growth stack can retain an MMP for operational attribution, fraud and deep-linking needs, then evaluate an incrementality layer for cross-channel causal analysis, with experiments providing calibration and a warehouse or BI layer supporting governance.
| Option | Typical fit | Relationship to Polaris |
|---|---|---|
| AppsFlyer | Mobile attribution, fraud protection, deep linking, partner integrations, and UA operations | MetricWorks documents a data integration; Polaris is positioned around modeled incrementality |
| Adjust | Mobile measurement, attribution, fraud prevention, and campaign reporting | Can remain the operational MMP while Polaris is assessed for incremental budget analysis |
| Singular | Marketing-data aggregation, attribution, and cross-channel reporting | Its marketing-intelligence role differs from Polaris’ emphasis on incrementality |
| Branch | Deep linking, attribution infrastructure, and user journeys | Not primarily positioned as a full MMM-plus-experiment incrementality workflow |
| In-house MMM and experimentation | Teams needing control and customization and able to sustain data engineering, statistical expertise, and experiment operations | Offers an alternative to buying a packaged incrementality platform, with greater internal build and maintenance burden |
| Platform-native measurement | Channel-specific aggregated reporting, modeled conversions, or lift studies | May aid optimization within a platform but may not provide a neutral cross-channel view |
Official product sites: AppsFlyer, Adjust, Singular, and Branch.
What to verify before evaluating or buying
MetricWorks’ May 2023 launch coverage described a free Polaris tier for one title on either iOS or Android, one cohorted incrementality metric, coverage across countries and campaign-related dimensions, and up to 12 months of historical daily visibility. These are launch-era terms, not confirmed current 2026 availability or plan limits. The available official documentation does not establish current pricing or whether that free tier remains available. Confirm current terms directly with MetricWorks through its official site; do not treat third-party historical price signals as a verified quote.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Quick Recap
- What minimum history, data volume, and spend are required for useful results?
- Which MMPs, APIs, metrics, cohort days, and channel inputs are supported under the current plan?
- Can results and model versions be exported to the company’s warehouse or BI tools?
- How are experiments designed, run, and priced, and what happens when intervals are wide?
- How are promotions, product changes, brand activity, and organic factors represented?
- What security, retention, privacy, and data-processing commitments apply?
- Which team owns the definitions of ROAS and LTV, and how will the modeled metrics affect budgets and finance reporting?
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
