Skip to main content
Metrics type: Key Metrics • Category: Ad Platform

​
At a glance

Return on ad spend on TTD. conversionValue ÷ spend with click-through, view-through, and cross-device (UID 2.0) revenue all included. TTD ROAS reads materially higher than peer DSPs because of the cross-device attribution capture; on the same campaigns, expect TTD ROAS to be 30, 60% higher than StackAdapt or MediaMath. The trade-off: harder to validate against Shopify-truth.

​
Calculation

Calculated automatically from your The Trade Desk 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 enterprise homeware brand. Window 02 Apr 26 to 01 May 26. Shopify-truth context: UTM-tagged TTD revenue $94,000; business ROAS = 0.73×. Blended honest figure (40/60 weighted) = ~1.33×. What’s interesting:
  1. Account ROAS 2.21× sits in the warn band (good ≥4, warn ≥2). For TTD this is on the lower side but consistent with the heavy CTV mix (48% of spend at 1.56× combined).
  2. Retargeting at 4.63× via UID 2.0 is the platform’s signature win. Cross-device retargeting reaches the same user across web, mobile, CTV without third-party cookies; conversion rate is materially higher than cookie-only retargeting.
  3. CTV Premium PMP at 1.56× ROAS reads low but cross-device contributes 57% of CTV revenue (38,200of38,200 of 66,600). On a click-only basis, CTV Premium would be effectively 0.66× ROAS; the cross-device dimension makes the channel viable.
  4. DOOH at 2.17× ROAS is unique; no peer DSP would compute this number reliably. The ROAS is 100% cross-device-derived; validate via geo-holdout testing if you’re scaling.
  5. Prior 30D ROAS was 2.41×. Down to 2.21%, an 8% decline. Spend up 19%, revenue up 15%. Mild scaling pressure; acceptable.
  6. Compared to MediaMath (during migration evaluation): TTD’s 2.21× vs MediaMath’s 2.28× on similar windows. Within noise. The reason advertisers migrate isn’t the headline ROAS, it’s reliability, inventory access, and the cross-device measurement edge.
Quick sanity tests:
  • ROAS up + spend up = healthy.
  • ROAS down + spend up = scaling beyond efficient frontier.
  • ROAS down + cross-device share collapsed = UID 2.0 implementation issue.
  • ROAS materially higher than peer DSPs = expected; cross-device dimension explains most of it.
  • Retargeting ROAS lower than 3× = UID 2.0 audience match degraded; check.

​
Sibling cards merchants should reference together

​
Reconciling against the vendor’s own dashboard

Where to look in The Trade Desk: TTD Platform → Reports → Performance → ROAS column at advertiser level. Should match this card to within sub-percent rounding. Why our number may legitimately differ from TTD UI: Cross-connector reconciliation: Why TTD ROAS reads higher than peer DSPs The cross-device attribution via UID 2.0 captures revenue paths that other DSPs simply cannot measure:
  1. CTV-to-mobile conversion: household sees CTV ad on smart TV; converts on mobile 3 days later. UID 2.0 matches the household identity; peer DSPs miss the link entirely.
  2. DOOH-to-device: billboard impression matched to device-level conversion via UID 2.0 device-graph.
  3. Cross-app retargeting: same user in browser and native mobile app, recognised as one identity.
The cross-device share typically adds 25, 40% to TTD’s reported revenue vs peer DSPs on equivalent campaigns. The honest interpretation: TTD’s ROAS is more complete than peers, but also more vulnerable to over-claim if UID 2.0 implementation isn’t tight. Validate via incrementality testing.

​
Known limitations / merchant FAQs

My TTD ROAS is 3× and Google Ads ROAS is 6×, should I move budget? The cross-platform question. Search captures intent; TTD seeds and follows it. Run a holdout test before reallocating. TTD’s cross-device attribution captures more of the funnel than Search measures; the ROAS gap is real but the channel substitution math doesn’t equate. Most enterprise advertisers run both and balance based on absolute revenue contribution, not relative ROAS. Why is CTV ROAS so much lower than retargeting ROAS? CTV is cold-audience top-of-funnel; retargeting is warm-audience close-of-funnel. Same dynamic on every DSP. TTD-specific advantage: cross-device attribution lifts CTV ROAS 30, 60% above what peer DSPs would credit, but retargeting still wins on absolute efficiency. My TTD ROAS dropped 25% after I disabled cross-device attribution, is the cross-device portion fake? No. The cross-device portion is the genuine measurement of household-level conversion paths. Disabling it is like switching off your bedroom light to see how dark the room is, you’re choosing not to measure something real. Cross-device matters for understanding TTD’s contribution; disable it only if your strategic goal is comparing to a DSP-without-cross-device baseline (e.g. for like-for-like comparison with MediaMath ROAS). Should I use TTD’s ROAS for board reporting? With caveats. Recommended approach: report TTD ROAS in the methodology footnote as “cross-device attributed”, and the blended figure (40% TTD + 60% Shopify-UTM) as the operating ROAS. The two together give the board the optimistic and conservative bounds. Why is my TTD ROAS so much higher than the same campaigns running on MediaMath? Cross-device is the biggest wedge. UID 2.0 captures attribution that MediaMath cannot. Also: TTD’s pixel infrastructure is more reliable, so fewer false-zero days. Don’t compare TTD ROAS directly to MediaMath ROAS without normalising for cross-device. My retargeting ROAS dropped from 5× to 3×, what happened? First check UID 2.0 health. The most common cause of a sharp retargeting ROAS drop on TTD is a degraded cross-device match rate (often from a UID 2.0 email-hash exchange issue or a CRM-pipeline change). Restoring UID 2.0 health typically recovers the ROAS within 7, 14 days. Should I optimise Koa for ROAS or CPA? For high-AOV (>200): ROAS. For low-AOV (<80): CPA. Mid-range: try both for 14 days, pick the winner. Kokai (TTD’s next-gen platform) generally optimises ROAS more reliably than legacy Koa for AOV >$150. Can I trust the “today” ROAS? More than peer DSPs, less than 7-day rolling. TTD’s faster ingest helps but the same long-tail conversion lag (especially CTV) applies. Why does TTD ROAS sometimes spike unexpectedly? Common causes: (1) UID 2.0 cross-device match rate improvement (e.g. after a CRM data refresh); (2) PMP deal kicking in with strong inventory; (3) Koa graduation moving to a more efficient bidding pattern; (4) seasonal lift on CTV (sports finals, awards) where cross-device match for engaged audiences is especially good. My multi-region account, how does ROAS work? TTD organisations are typically multi-currency-capable through child advertisers. Each advertiser is single-currency; aggregate ROAS across regions requires weighted blend at report layer.

​
Tracked live in Vortex IQ Nerve Centre

ROAS is one of hundreds of KPI pulses Vortex IQ tracks across The Trade Desk 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.