Singular this month published its annual ROI Index 2026, a performance benchmark built on trillions of impressions, billions of clicks, and billions of installs drawn from thousands of ad networks and millions of campaigns worldwide. The report, authored by Saadi Muslu with contributions from Business Intelligence Lead Gaston J. Laterza and Data Engineer Pablo Agustin Navarro, marks the first edition to include dedicated multi-touch attribution (MTA) leaderboards, adding a new analytical layer to what has become one of the most closely watched benchmarking reports in mobile advertising.

The headline finding is stark. According to the report, every platform appearing repeatedly across the 2026 leaderboards is investing heavily in automated bidding, creative-level optimisation, predictive modelling under sparse signals, and cross-channel signal aggregation. "AI is not the edge. It is the baseline," the document states. The differentiator, according to Singular, is how effectively each platform applies those capabilities at scale.

Scale concentrated among a core group

A core group of media channels appears repeatedly across all-platform global lists, gaming and non-gaming splits, and iOS and Android scaled leaderboards. According to the report, that group includes Apple Ads, AppLovin, Google Ads, Liftoff, Meta, Mintegral, Moloco, TikTok For Business, Unity Ads, Snapchat Ads, and X Ads. These platforms are described not merely as large but as "versatile and performing across operating systems and verticals."

The platform density section of the report - which tallies how many leaderboard positions each network occupies across all segments - shows the top four networks each reaching 31 rankings. That density spans OS splits, vertical categories, and the new MTA views. According to Singular, this cross-category durability is meaningful signal for marketers building diversified media portfolios.

The report separates qualifying networks into two tiers: Scaled Ad Partners, which meet defined minimum spend and volume thresholds within each category, and Growth Ad Partners, which demonstrate strong ROI performance below those thresholds. Leaderboards are presented alphabetically within each qualifying tier. The distinction is designed to prevent scale from distorting efficiency comparisons - a structural choice that allows smaller or more specialised networks to surface without competing head-to-head against the largest global platforms.

Apple Ads occupies an unusually prominent position throughout the 2026 report. According to Singular's data, it appears on the global scaled leaderboard across all verticals, and SplitMetrics - whose contribution appears in a dedicated section of the document - provides benchmark data specific to the platform from its Apple Ads Search Results Benchmarks Report 2026, based on aggregated data from apps and mobile games that optimised Apple Ads search results campaigns between January and December 2025.

According to SplitMetrics head of customer success Mike Talashko, campaigns on Apple Ads search results show an average tap-through rate of 9.7% and a conversion rate of 66.2%, though with significant disparities across app categories. The average cost per tap across categories was $2.25 in 2025, while the overall average cost per acquisition stood at $3.76. Again, category-level variation is substantial. In Finance and Sports - particularly sports betting - high average CPA figures reflect correspondingly high average lifetime value, according to the SplitMetrics data.

The SplitMetrics contribution also highlights category-specific behaviour. Finance apps use additional placements to reach users who deliver strong post-install performance. Entertainment apps lean on visual storytelling and Today Tab placement for what Talashko describes as impulse discovery. Sports apps contend with pronounced seasonality, with single-day events capable of producing noticeable engagement spikes. Food and Drink and Health and Fitness apps may benefit from strategic daily budget allocation tied to moments of high intent - a food order or a morning workout.

Custom product pages are identified as a significant lever. According to SplitMetrics data cited in the report, top categories such as Sports and Finance show a custom product page adoption rate of 65-80%, with a double-digit conversion rate uplift of 20-50%. That combination implies direct impact on cost per tap and cost per acquisition. The report also notes that as of March 2026, Apple began rolling out an additional placement in search results - a development covered by PPC Land at the time - which the SplitMetrics section describes as a new opportunity to generate more impressions.

The Apple Ads context in the 2026 ROI Index sits against a backdrop of significant platform change. PPC Land has tracked the platform's structural expansion closely: Apple rebranded from Apple Search Ads to Apple Ads in April 2025, expanded to 21 new countries in October 2024, and began rolling out multiple ad positions in App Store search results starting March 3, 2026 in the UK and Japan. More recently, Apple launched Apple Business on April 14 with Maps advertising expected in the United States and Canada this summer.

Multi-touch attribution added for the first time

The most structurally new element of the 2026 ROI Index is the introduction of MTA leaderboards. According to Singular, last-touch attribution remains the industry default, but it captures only the final interaction before conversion. In reality, most user journeys involve multiple touchpoints across discovery, consideration, and conversion. Channels that influence early engagement or mid-funnel consideration often receive no credit under last-touch models - an omission that can materially distort budget allocation.

Three distinct MTA metrics appear in the report. Exclusive Reach measures installs where a network was the only engagement in the path to attribution, highlighting platforms delivering truly unique audiences. Assist Power captures installs where a network appeared in the journey but did not receive last-touch credit, surfacing channels that influence discovery and consideration without closing the conversion. MTA Uplift measures the increase in a network's measured contribution under multi-touch attribution compared with last-touch attribution, revealing platforms whose influence extends beyond final-click reporting.

According to Singular's analysis, networks appearing across multiple MTA metrics are influencing more than just the final step in the conversion journey. Networks that dominate Exclusive Reach serve audiences that would otherwise go unreached. Networks with high Assist Power shape intent before the closing interaction arrives. The Exclusive Reach MTA leaderboard includes Adjoe, Apple Search Ads, AppLovin, Google Ads, Meta Ads, Mintegral, Reddit, Snapchat, TikTok, and Unity Ads. The combination of names on that list is notable: several platforms that appear here do not necessarily dominate last-touch rankings, which is precisely the point Singular is making.

This matters for the marketing community because, as PPC Land has reported on the limitations of last-click modelspreviously, budget allocation decisions based solely on last-touch data can structurally undervalue channels that initiate or sustain user journeys. The ROI Index is now quantifying that gap with specific network-level data.

Privacy-first measurement described as mature infrastructure

The 2026 report draws a clear line under what Singular calls the "ATT shock" phase. According to the document, 2025 accelerated infrastructure stability beyond SKAdNetwork, with Meta's Aggregated Event Measurement, TikTok's advanced real-time reporting, and Google Ads' expanded on-device modelling all contributing to a more stable iOS measurement environment. The consistency of scaled iOS leaderboards, according to Singular, reflects that maturity. iOS is described as "no longer unstable" - simply more model-dependent, meaning sophistication in measurement and modelling now determines performance outcomes.

Regional data in the index remains largely Android-weighted. According to Singular, this reflects privacy-safe limitations on granular iOS geo reporting under SKAdNetwork (SKAN). Under SKAN, geographical data is limited for most iOS installs, and the data that is available is embedded in each advertiser's own customised SKAN conversion model. The regional leaderboards covering APAC, EMEA, North America, and South America therefore reflect a predominantly Android-based dataset, and Singular notes this explicitly as a methodological constraint rather than a data gap.

Emerging markets are described as stabilising. According to the report, Android-heavy regions show more consistent ROI patterns than in previous volatile cycles, with infrastructure investment and platform maturity reducing unpredictability. The practical implication flagged in the document is that geo expansion strategies that were put on hold during periods of higher volatility may warrant re-evaluation.

The ATT opt-in dynamics documented in the report align with trajectory data from other measurement providers. Adjust's Mobile App Trends 2026 report, published on February 18, 2026 and covered by PPC Land, found that the industry-average ATT opt-in rate rose from 35% in Q1 2025 to 38% in Q1 2026 - a directional signal consistent with Singular's characterisation of a stabilised iOS environment.

Rewarded and incentive-driven inventory now classified as core

One shift in framing from previous editions of the index concerns rewarded and incentive-driven inventory. According to Singular, these environments "continue to appear prominently across gaming and growth leaderboards" and are described as "no longer experimental budget lines." The report attributes this change in status to stronger retention filtering, better fraud controls, and more disciplined optimisation. Marketers who previously deprioritised rewarded inventory on quality grounds are directed to re-test under what the report calls "modern guardrails."

Gaming and non-gaming no longer clearly separated

In previous years, according to Singular, it was possible to bucket networks by vertical strength - gaming or non-gaming. The 2026 data blurs that line. Many scaled partners now appear across both gaming and non-gaming leaderboards, on both operating systems. Optimisation systems are described as adapting more fluidly across signal types, including subscription events, purchase flows, and ad monetisation. The practical consequence is that assumptions about which channels work within which verticals may have degraded in accuracy over the past two years.

Consolidation and M&A amplifying density

The report notes that M&A activity and ecosystem expansion are increasing cross-leaderboard density for certain platforms. Broader infrastructure, according to Singular, translates into stronger modelling, more signal access, and cross-portfolio durability. The document does not name specific transactions, but the patterns are visible in the platform density table, where several platforms with recent acquisition histories appear at the top of the ranking count.

Growth-tier networks are described as still carving out efficient pockets. The recommended posture is to balance core scaled allocation with disciplined testing in the growth tier - a framing that acknowledges concentration at the top without treating it as a reason to abandon diversification.

Methodology and scope

According to the full methodology section, the index is based on aggregated, anonymised performance data across trillions of impressions, billions of clicks, billions of installs, thousands of ad networks, and millions of campaigns. On Android, Singular prioritises networks with scale in both spend and number of customers, analysing retention, click-to-install counts, and relative fraud levels alongside return on investment. On iOS, similar scale parameters apply, with Singular Unified Measurement serving as the primary data source. That includes data from SKAdNetwork, where Singular analyses raw postbacks and the information embedded in conversion values and ad network parameters before translating them into higher-ROI conversion signals. Apple Search Ads uses a separate API and is handled accordingly.

The report includes an explicit caveat. According to Singular, the findings are drawn from the company's own client base, and "while the findings are based on billions of dollars of spend and installs, and trillions of ad impressions, making it highly indicative of what's actually going on, marketers should consider the ROI Index as part of a broader context and complement them with additional data sources and strategic judgment."

The Singular ROI Index is published annually, with occasional refreshes. Singular also publishes a separate Quarterly Trends Report covering CPI rates, ad network spend share, ATT opt-in rates, paid versus organic split, and regional data points. The ROI Index 2026 is available via the Singular website.

Timeline

Summary

Who: Singular, a mobile attribution and analytics company, authored the ROI Index 2026 with contributions from SplitMetrics for the Apple Ads benchmarking section. The report was written by Saadi Muslu, with Business Intelligence Lead Gaston J. Laterza and Data Engineer Pablo Agustin Navarro.

What: The Singular ROI Index 2026 is an annual mobile advertising performance benchmark built on trillions of impressions, billions of clicks, and billions of installs across thousands of ad networks and millions of campaigns. The 2026 edition introduces dedicated multi-touch attribution leaderboards for the first time, using three metrics - Exclusive Reach, Assist Power, and MTA Uplift - to quantify network contribution beyond last-touch reporting. Apple Ads, AppLovin, Google Ads, Liftoff, Meta, Mintegral, Moloco, TikTok For Business, Unity Ads, Snapchat Ads, and X Ads appear across the core scaled leaderboards.

When: The ROI Index 2026 was published on April 19, 2026. The Apple Ads benchmarking section draws on SplitMetrics data from January through December 2025, released in 2026.

Where: The report covers global mobile advertising performance across all major markets, with regional leaderboards for APAC, EMEA, North America, and South America. Regional data is largely Android-weighted due to privacy-safe limitations on granular iOS geo reporting under SKAdNetwork.

Why: The index matters for the marketing community because it provides a structured, data-driven view of which platforms deliver consistent ROI at scale across operating systems and verticals. The introduction of MTA leaderboards is particularly significant, as it allows budget allocation decisions to account for assist-heavy contributors that last-touch reporting systematically undervalues. For mobile app marketers specifically, the Apple Ads benchmarking data from SplitMetrics provides cost and conversion benchmarks - average cost per tap of $2.25, average cost per acquisition of $3.76, tap-through rate of 9.7%, and conversion rate of 66.2% - that are otherwise difficult to obtain at category level in a single public source.

Share this article
The link has been copied!