Skip to main content
Metrics type: Key MetricsCategory: Ad Platform

At a glance

Return on ad spend for Taboola native, the headline efficiency reading. conversionsValue ÷ spent. Native ROAS reads structurally lower than Search or Social on a last-click basis because Taboola is upper-funnel; benchmark against your own history rather than other channels. A 2× last-click ROAS on Taboola often translates to 4, 5× true contribution once multi-touch and assisted conversions are factored in.

Calculation

Calculated automatically from your Taboola data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.

Worked example

The same UK supplements brand. The 30-day window is 02 Apr 26 to 01 May 26. Shopify-truth context: UTM-tagged Taboola revenue £18,400. Business ROAS on Shopify-truth = 1.23×. Both numbers are real; they answer different questions. What’s interesting:
  1. Account ROAS 2.68× sits in the “warn” band on the gauge thresholds (good ≥4, warn ≥2). That is normal for native; the gauge thresholds were calibrated against ad-platform peers (Search, Social) where last-click ROAS is structurally higher. For Taboola specifically, 2, 3× last-click is an acceptable working figure for an upper-funnel channel; treat <1.5× as the action line, not <2×.
  2. Lookalike Retargeting at 4.88× is the standout campaign. Retargeting on Taboola works disproportionately well because the second touch closes; the first Taboola click did the brand awareness, the second touch (now retargeting) closes the loop. If you can scale this campaign, do; it is the cleanest unit-economics on the account.
  3. Premium Publisher pause was right. 1.33× ROAS sat below the account average; the £2,400 is better redeployed to Lookalike Retargeting where ROAS is 3, 4× higher.
  4. The 30-day prior window had ROAS 3.69×. This 30-day window dropped to 2.68×, a 27% decline. Spend up 21%, revenue down 12%. That is the classic “scaled beyond efficient frontier” shape. Action: cap budget at the prior-period average for two weeks, refresh creative, test new lookalike audiences before re-expanding.
  5. Don’t quote 2.68× to the CFO without context. The Shopify-truth ROAS is 1.23×, which sounds dire. The honest blended figure (40% Taboola-claim + 60% Shopify-truth) is roughly 1.81×, marginal but not unprofitable on 60% gross-margin economics with retargeting lift.
Quick sanity tests:
  • ROAS up + spend up = healthy scaling, expand if pacing allows.
  • ROAS flat + spend down = budget cut without channel deterioration; re-engage when ready.
  • ROAS down + spend up = scaling beyond efficient frontier (most common cause of an alert on Taboola). Cap budget, audit publisher allowlists, refresh creative.
  • ROAS down + spend flat = something changed in attribution, conversions, or audience quality. Check pixel health on Clicks vs Conversions before changing anything.
  • ROAS down + spend down = the algorithm pulled back on its own (budget pacing, low-quality inventory withdrawal). Often self-corrects within 7 days.

Sibling cards merchants should reference together

Reconciling against the vendor’s own dashboard

Where to look in Taboola Backstage: Backstage → Campaigns → Performance → “ROAS” column at account level. The headline figure should match this card to within sub-percent rounding. Other Backstage views that look similar but are not the same number:
  • All Conversions ROAS: includes view-through and offline. This card uses click-attributed primary conversions only.
  • Reports → Site Breakdown: per-publisher ROAS. The weighted average should equal the account ROAS but per-site rows can be noisy on small samples.
Why our number may legitimately differ from Backstage: Cross-connector reconciliation, the important one: This card is Taboola’s view of Taboola-driven efficiency. Multi-touch reality usually compresses ROAS substantially:
Why native ROAS reads so differently from Search and Social Native ROAS sits structurally low on a last-click basis because:
  1. Cold audience. Users on news / blog sites are not actively shopping; they are entertainment-mode. The click value is genuinely lower than a Search click (intent-driven) or a Social click (interest-driven on a feed they curated).
  2. Long conversion lag. Native conversions take an average 3, 7 days vs Search at 1, 2 days; many of those orders are subsequently last-clicked by another channel and lost from Taboola attribution.
  3. Multi-touch hand-off. Native is most often the first paid touch; the user later returns via Search, Email, or Direct. Last-click attribution credits the closer, not the seeder.
For this reason, comparing Taboola ROAS to Google Ads ROAS is misleading. Compare Taboola ROAS to its own historical baseline (this card’s vsP), and judge channel value via incrementality (holdout testing) or multi-touch attribution rather than absolute ROAS.

Known limitations / merchant FAQs

My Taboola ROAS is 2× and my Google Ads ROAS is 6×, should I move budget? Maybe, but not on that comparison alone. The two channels do different jobs. Google Search captures intent (someone searching for your product was buying anyway, often); Taboola seeds intent on cold audiences (you’re paying to introduce the brand). Moving budget from Taboola to Google often shows a Google ROAS lift in the short term and an absolute revenue contraction in the medium term as the funnel-feeder dries up. Run a holdout test before reallocating: pause Taboola for 14 days and watch Direct, Branded Search, and Email volumes. If they hold, move the budget. If they drop 10, 20%, Taboola was working harder than its last-click ROAS suggested. Why is my Taboola ROAS so much lower than Meta? Meta’s ROAS is structurally inflated post-iOS-14.5 (modeled conversions fill in gaps, often over-claiming by 30, 80%). Taboola’s ROAS is structurally deflated by multi-touch hand-off (other channels claim the conversion last). The two reads of the same business often differ by 3, 5×; that does not mean Taboola is bad and Meta is good, it means they measure different attribution shapes. Use multi-touch or holdout testing for the honest read. My ROAS dropped 30% overnight, what should I check first? In order of likelihood: (1) Pixel broke. Check Clicks vs Conversions; if clicks held but conversions cratered, the Taboola pixel stopped firing. Most common cause: site redeploy that lost the pixel snippet. (2) Auto-bid moved into bad inventory. Check Spend by Campaign for a per-campaign shift. (3) Attribution window changed. Someone shortened the window from 30d to 7d; ROAS will read 15, 30% lower from that day forward. (4) Creative fatigued. Check CTR Trend for a rolling decline over 7, 14 days. (5) Competitor entered the auction, pushing CPC up; check CPC Trend. Should I optimise for ROAS or absolute revenue on Taboola? For native, optimise for absolute revenue subject to a ROAS floor. The ROAS-floor approach (e.g. “spend whatever holds ROAS ≥1.8×”) prevents waste while letting auto-bid scale the channel. Pure ROAS-max on native often shrinks the account to a tiny set of premium publisher slots that are not scalable. Pure volume-max overspends into low-quality inventory. My Lookalike Retargeting ROAS is 5× but Article Distribution is 1.6×, can I shut off Article Distribution? You can, but think twice. Article Distribution is the funnel feeder; without it, the Lookalike audience erodes (no new users to seed). The right move is usually: keep Article Distribution at the minimum spend that maintains audience freshness (often 30, 50% of current), and scale Retargeting aggressively. Killing Article Distribution outright collapses the entire flywheel within 30, 60 days. Why does my Backstage ROAS round differently from this card? Backstage rounds at the report layer; this card stores 4 decimal places of precision. The rounded display can differ on small samples by 0.01, 0.05×. For a 30-day window with reasonable spend (>£5k), the gap is invisible. Can I trust the “today” ROAS? Less than the rolling 7-day. Today’s ROAS is built from incomplete data: clicks today often convert tomorrow or later (long-tail conversion delay), and the Taboola pixel takes 1, 4 hours to ingest. The 7-day rolling is the most actionable read for native; daily ROAS swings of 30, 50% are common and are mostly noise. My multi-region setup, how does ROAS work? Taboola accounts are single-currency by design. Multi-region advertisers run separate accounts per region (UK, US, EU). This card is per-account; aggregate ROAS across regions requires a weighted blend at the report layer, which is not currently in scope for the hero card.

Tracked live in Vortex IQ Nerve Centre

ROAS is one of hundreds of KPI pulses Vortex IQ tracks across Taboola and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English. Start for free or book a demo to see this metric running on your own data.