Advertising Report Metrics by Integration

June 12, 2026

Endpoint: GET /ads/{connection_id}/report — List all reports

This is a reference for the metrics returned by the Advertising Report endpoint: which metrics each supported integration populates, and how each platform's native field maps onto Unified.to's single normalized reporting model.

How reports return metrics

A call to GET /ads/{connection_id}/report returns an array of Report objects. Each report describes the performance of an advertising entity over a window, and carries its metrics in a single normalized array:

The value of the metrics array is that it is the same shape regardless of which platform connection_id points to. Whether the connection is Meta Ads or LinkedIn, a CLICKS metric means the same thing and is read the same way. Behind the scenes, Unified.to handles the platform-specific renaming, the nested-field extraction, and the unit conversions — and where a platform doesn't return a value directly, it computes it. If you ever need the original platform field, it remains available under raw.

The unified metric model

Each ad platform reports the same fundamental ideas — a click, an impression, a dollar spent — but names and shapes them differently. A click is clicks in Meta, METRIC_CLICKS in Google DV360, and Clicks in Microsoft Ads. Cost arrives as raw currency in some platforms and as micro-units (millionths of a currency unit) in Google Ads, which must be divided by 1,000,000 before it means anything.

Unified.to collapses all of that into one vocabulary of 92 normalized metrics (written in UPPER_SNAKE_CASE). The tables below show, for each normalized metric, the native field it is sourced from on each of the five integrations mapped in this reference: Meta Ads, Google Ads, Google DV360, LinkedIn, and Microsoft Ads. A — means that integration does not expose the metric, so it will not appear in the metrics array for that connection.

Coverage at a glance

The breadth of the metrics array depends entirely on which integration the connection is for. Not all platforms expose data at the same depth:

The universal core

These eight metrics are the common denominator: every integration mapped here returns them, making them the safe foundation for any cross-connection report or comparison:

CLICKS · IMPRESSIONS · CONVERSIONS · COST · CTR · CPC · CONVERSION_VALUE · CPA

Even within this universal set, the source data differs. COST is spend in Meta but arrives in micro-units in Google Ads and as costInLocalCurrency in LinkedIn (all normalized to the report's currency). CONVERSIONS is a clean field in most platforms but is pulled from Meta's nested actions[purchase]. And CTR, CPC, and CPA are all derived for LinkedIn — a reminder that "returned" sometimes means "computed for you," not "delivered by the platform."

Two near-universal ratios fall just short of full coverage: ROAS is derived everywhere it exists but is absent from DV360, and CPM is missing only from Google Ads.

Core performance metrics

The headline numbers most reports are built on — volume, cost, and the efficiency ratios derived from them.

Conversion detail

Beyond the headline conversion count, platforms differ on how they attribute and segment conversions. Google Ads and Microsoft Ads add "all conversions" totals and their revenue counterparts; DV360 and LinkedIn distinguish post-click from view-through (post-view) conversions; and Google Ads alone exposes cross-device conversions and raw interactions.

Video engagement

Video is the most fragmented category, because each platform instruments playback in its own way. Quartile completion (25/50/75/100%) is the one widely shared concept, supported across Meta, Google Ads, DV360, and LinkedIn — though Google Ads reports quartiles as rates while the others report counts. From there the platforms diverge: Meta uniquely tracks VIDEO_AVG_TIME_WATCHED and VIDEO_THRUPLAY; DV360 captures rich-media interaction signals like pauses, mutes, skips, and companion-banner activity; and Google Ads surfaces search-driven views.

Social engagement

Likes, shares, comments, and follows are native to the social-first platforms. LinkedIn reports them directly, and DV360 exposes the YouTube "earned" equivalents (earned likes, shares, and subscribers). Curiously, Google Ads maps SHARES and SAVES onto Gmail-specific forward and save actions — an artifact of how Gmail Ads engagement was instrumented.

Click and impression breakdowns

Most of these granular click types are LinkedIn-specific, reflecting its sponsored-content and document-ad formats (card clicks, company-page clicks, text-URL clicks, and so on). Meta contributes LANDING_PAGE_CLICKS (its inline_link_clicks) and UNIQUE_CLICKS, while LinkedIn adds an approximate unique-impression count.

Lead generation

Native lead-form metrics are exclusive to LinkedIn in this mapping, tied to its one-click Lead Gen Forms.

Viewability and measurement

Viewability — whether an impression was actually seen — is a Google ecosystem strength. Google Ads and DV360 share the Active View family (viewable, measurable, and percent-viewable impressions), DV360 adds billable and eligible impression counts plus average viewable time, and Google Ads contributes top-of-page impression share metrics from search.

Cost and financial breakdowns

DV360 is the only integration that decomposes cost into its component fees — media cost, data fees, platform fees — and reports advertiser-level revenue and profit. This reflects its role as a programmatic buying platform where margin transparency matters.

Viral / organic amplification (LinkedIn)

LinkedIn's "viral" metrics measure the organic spread of sponsored content — engagement that happens when someone's network sees an ad because a connection interacted with it. This entire 19-metric family is exclusive to LinkedIn and mirrors its standard metrics (impressions, clicks, conversions, video quartiles, leads) but for organically-amplified reach. No other integration models this concept.

Integration-by-integration takeaways

LinkedIn is the deepest integration by a wide margin. Beyond solid core coverage, it owns two large metric families that exist nowhere else: native lead-form metrics and the complete "viral" organic-amplification set. If a report needs B2B lead data or organic-reach attribution, LinkedIn is the only source.

Google DV360 is the choice for media-buying transparency. It uniquely breaks cost into media, data, and platform fees and reports profit, and it offers the richest video-interaction signals (pauses, mutes, skips, companion banners) alongside full Active View viewability.

Google Ads provides broad, dependable core coverage plus search-native metrics — impression share, cross-device conversions, and "all conversions" rollups. Watch for its two quirks: micro-unit cost fields that need dividing by a million, and quartile metrics expressed as rates rather than counts.

Meta Ads concentrates on core performance and granular video-watch behavior. Its conversion and value figures come from nested action breakdowns rather than flat fields, so the purchase action type must be extracted explicitly.

Microsoft Ads is the leanest integration, covering the universal core plus video views, view-through conversions, and "all conversions" totals — sufficient for standard performance reporting but without engagement, viewability, or social depth.

Working with the response

• Build cross-connection reports on the universal eight. CLICKS, IMPRESSIONS, CONVERSIONS, COST, CTR, CPC, CONVERSION_VALUE, and CPA are the only metrics guaranteed to appear in the metrics array for every integration mapped here.
• Check for presence, don't assume it. With 60 of 92 metrics returned by only one integration, code that reads the metrics array should handle a metric being absent rather than expecting a fixed set.
• Read currency from the report, not the metric. All monetary metrics are normalized into the report's currency field; Google Ads micro-units (/1e6) and per-platform conventions are already resolved for you.
• Some metrics are computed. Ratios like ROAS, and several of LinkedIn's rates, are derived by Unified.to rather than supplied by the platform — they arrive in the metrics array all the same.
• Drop to raw when you need the original. Anything platform-specific that the normalized model doesn't capture remains available on the report's raw field.