Refacto

Podcast episode

Walmart Connect's Ryan Mayward on Bringing Vizio Data to Magnite and the Yahoo DSP

·Marketecture

ctv dsp measurement retail-media ssp

TL;DR

Ryan Mayward, SVP/GM of Walmart Connect, explains a new four-party deal integrating Walmart's shopper-audience data and closed-loop measurement into Magnite (a supply-side platform, or SSP, that helps publishers sell ad inventory) so that Yahoo DSP (software advertisers use to buy digital ads) buyers can target and measure against Walmart signals on Vizio CTV inventory. More DSPs will follow. The episode is a focused, substantive retail-media/CTV infrastructure briefing — worth 16 minutes for anyone in programmatic, retail media, or CTV.

What was covered

  • The Walmart-Yahoo-Magnite-Vizio deal: Walmart Connect embedded its audience segments and closed-loop measurement directly into Magnite's SSP on the sell side. Yahoo DSP is the initial exclusive buying path, letting Yahoo DSP buyers activate campaigns against Walmart audiences on Vizio smart-TV inventory, with Walmart controlling the data throughout.
  • Why sell-side integration: Vizio already runs its ad business primarily on Magnite (plus Index Exchange), so plugging in at the SSP layer was the lowest-friction entry point. Mayward calls it "flexible and lightweight" and notes it keeps Walmart in control of audience and measurement regardless of which DSP is transacting.
  • More DSPs coming: Walmart plans to expand the arrangement to additional DSPs and a direct Magnite buying path over time; Yahoo is explicitly described as first, not exclusive long-term.
  • Vizio OS scale: As of March 2025, Vizio's operating system was the number-one selling smart-TV OS in the U.S. — roughly one in five TVs shipped carry Vizio OS. Walmart is also applying the Vizio OS to its private-label TV brand, ONN, to expand household penetration.
  • Walmart DSP and existing CTV partnerships: Walmart built its DSP in partnership with The Trade Desk and has used it for years to run Walmart-audience-targeted CTV campaigns on Paramount+, NBCUniversal, and Disney inventory. The new Magnite integration is additive, not a replacement.
  • Outcomes and iROAS (incremental return on ad spend): Mayward says standard ROAS (return on ad spend, a basic revenue-per-dollar-spent metric) has "diminished in importance" over the past 12–15 months; brands now demand iROAS — proof that the media drove sales that wouldn't have happened without the ad. Walmart uses holdout groups, ghost ads, and synthetic control groups to deliver this across online and in-store purchases.
  • In-store media expansion: Walmart is testing large digital screens at end-of-aisles in grocery, beauty, and other store sections, with broader rollout planned later in 2025. Details are being held for Cannes. Mayward notes the format lends itself naturally to incrementality measurement via per-store holdout designs.

Notable claims & predictions

  • Mayward on Vizio OS reach: "As of March, the Vizio operating system was the number-one selling operating system in the U.S. — about one in five TVs shipped had Vizio OS powering them." — Signals Walmart's CTV supply position is growing faster than widely appreciated.
  • Mayward on ROAS obsolescence: "For the last 12–15 years, return on ad spend got retail media going, but has really diminished in importance. Brands want iROAS at this point in time." — A direct statement that the dominant retail-media metric is shifting, with implications for how measurement vendors and agencies pitch.
  • Mayward on data openness post-Vizio acquisition: "We want to be where advertisers are spending their time and applying their budgets. And that's in multiple DSPs." — Directly addresses fears that the Vizio acquisition would wall off the inventory.
  • Mayward on consistent measurement across DSPs: "The objective for us is to make sure that brands have consistent measurement, regardless of the ad format or channel that they're accessing through Walmart Connect." — Walmart is positioning its attribution layer, not its DSP, as the durable competitive moat.
  • Mayward on in-store screens: "We do plan to test new large screens at the end of aisles throughout grocery and beauty and other parts of Walmart stores…we'll be expanding on that test later this year." — Confirms physical retail media hardware investment is accelerating.

Why this matters for ad-tech operators

  • SSP-side data integration as the new curation model: Walmart choosing to embed its first-party data into Magnite at the sell side — rather than requiring buys through its own DSP — is a meaningful structural choice. It signals that large data owners may increasingly use SSP-layer curation (packaging audience + inventory together on the supply side) as the distribution mechanism, bypassing the need to own or mandate a specific demand path. Other retail media networks and data owners will watch this closely.
  • iROAS as the emerging standard KPI for retail media: Mayward's explicit statement that ROAS is declining in importance and iROAS is the new demand from brands is a buying-behavior signal. Agencies and measurement vendors (VideoAmp, iSpot, Mediaocean, etc.) that haven't built incrementality methodologies into their retail media offerings face a product gap. Publishers and SSPs offering retail media data partnerships will need to speak iROAS fluently in pitches.
  • Vizio OS scale reframes Walmart's CTV supply position: One-in-five U.S. TV shipments running Vizio OS — plus ONN private-label TVs now adopting the OS — means Walmart is accumulating a CTV supply footprint that rivals or exceeds some pure-play streaming publishers. Buyers allocating CTV budgets need to re-score Walmart/Vizio as a supply-side force, not just a data-and-measurement overlay.
  • In-store digital screens as measurable incremental inventory: Mayward's hint at per-store geographic holdout designs for new end-of-aisle screens suggests Walmart will bring the same rigor it applies to digital to physical retail media. For brands and agencies running omnichannel programs, this could make in-store screens more defensible as a line item — but also raises the bar for measurement that other in-store media vendors (digital-out-of-home networks, competitors like Kroger, Target) will need to match.

Full analysis

Step 1 — Frame

Walmart Connect is embedding its shopper data and closed-loop measurement directly into Magnite's SSP (the technology publishers use to sell ad space), so buyers using the Yahoo DSP (the software advertisers use to buy ads) can target and measure Walmart audiences on Vizio smart-TV inventory. Yahoo is first, not exclusive. More DSPs and a direct Magnite path follow.

What's actually being decided (by Walmart, but the implication is industry-wide): whether a giant data owner distributes its audiences through the sell side — packaging data with inventory on the SSP — rather than forcing buyers through its own DSP. That's the structural news. Two secondary signals: iROAS (incremental return on ad spend — proof the ad caused a sale that wouldn't have happened otherwise) is displacing plain ROAS, and Vizio's OS reach makes Walmart a real CTV supply owner, not just a data overlay.

Reversibility: Type 2 for Walmart (sell-side integrations are reversible, additive to the Trade Desk path). But the market expectation it sets is stickier.

Timeline: Live now; more DSPs over 2026; in-store screens and Cannes reveals mid-2025 (already past as framed).

This is a genuinely substantive structural story, not a press release. Proceeding.

Step 2 — The Council

The Market Analyst The headline most people will miss: Walmart is telling the market its measurement layer is the moat, not its DSP. That's a quiet repudiation of the "every retailer needs its own DSP" thesis the last three years sold. In plain terms: Walmart would rather be the referee everyone trusts than own one of the teams. Winners — Magnite, which becomes the canonical pipe for retail data into CTV, and Yahoo, which gets a scarce head start. The Trade Desk loses a sliver of exclusivity it never advertised it had; the Publicis–TTD thaw in your reading suggests TTD is already managing relationship friction on multiple fronts. Other RMNs (Kroger, Target, Albertsons) now face a template they'll be asked to match.

For a non-specialist: Walmart decided its real product is trustworthy proof that ads work, not the cash register that sells them.

The Skeptic The load-bearing assumption is that "Walmart controls the data throughout" is durable and that buyers want a sell-side curation model. Both are shakier than the episode lets on. Sell-side curation has been pitched for years; adoption has been patchy because buyers distrust data they can't inspect and SSPs they can't fully audit. And "Yahoo first, more DSPs coming" is exactly what every exclusive-that-isn't says before the roadmap slips. On iROAS: Walmart grades its own homework. Holdout and synthetic-control methods are real, but when the seller owns the measurement, "incrementality" is marketing until a neutral third party validates it. Don't confuse a clean narrative with a settled outcome.

For a non-specialist: The company selling the ads is also the company proving the ads worked — be a little skeptical.

The Operator Tuesday-morning reality: a Yahoo DSP trader can now buy Walmart audiences on Vizio. Great — but the second DSP integration is where this gets hard. Each DSP wants the data modeled its way, line-item and creative mapping breaks (your own Index Exchange standup notes flag exactly this "line item and creative usage complexity"), and "consistent measurement across DSPs" is a promise that dies in reconciliation. At 90 days the failure mode isn't strategy, it's mismatched numbers: Yahoo's iROAS report won't tie to the next DSP's, and the brand's analyst escalates. Whoever owns the measurement layer also owns every "why don't these match?" support ticket.

For a non-specialist: Connecting one buying tool is a demo; connecting five and making the reports agree is the actual job.

The Customer / End User (the brand / agency buyer) Two customers here. The brand wins: Walmart audiences on premium CTV, measured to in-store sales, is genuinely useful, and iROAS is what CFOs at brands actually want — proof the spend added sales. But the agency has mixed feelings. Consistent cross-DSP measurement controlled by Walmart erodes the agency's own measurement and arbitrage value. And buyers should ask whether "iROAS" is a standard or a Walmart dialect — if Kroger and Target each define incrementality differently, the buyer is back to comparing apples to oranges, just with fancier math.

For a non-specialist: Advertisers get better proof their ads worked, but every retailer may define "worked" its own way.

The CFO The real cost isn't the integration; it's the measurement infrastructure and the people to run holdouts, ghost ads, and synthetic controls at scale across online and in-store. That's expensive, recurring, and hard to staff. Walmart can amortize it across enormous spend; a mid-size publisher or RMN copying this will bleed on it. iROAS also tends to shrink the credited number versus ROAS — incrementality is always smaller than total attributed sales — so anyone whose business was sold on inflated ROAS faces an uncomfortable repricing conversation with clients.

For a non-specialist: Honest measurement usually makes the results look smaller, and that changes what you can charge.

Step 3 — The Tensions

  1. Moat: measurement vs. demand path. The Analyst says owning measurement is smarter than owning a DSP. The Skeptic counters that self-graded measurement isn't a moat — it's a liability waiting for a neutral validator (VideoAmp, iSpot, an MRC accreditation) to either bless or undercut it.

  2. iROAS as progress vs. iROAS as fragmentation. The Customer wants the better metric; the CFO and Skeptic note that if every retail media network defines incrementality its own way, "iROAS everywhere" just relocates the comparability problem rather than solving it.

  3. Strategy vs. plumbing. The Analyst sees an elegant sell-side distribution model; the Operator sees five DSP integrations whose reports won't reconcile, which is where market narratives quietly die.

Step 4 — Synthesis

What this hinges on:

  • Is sell-side curation finally the distribution model for big data owners, or another well-pitched idea that stalls on buyer trust? Walmart has the leverage to make it work where others couldn't — it owns the data, the measurement, and now the supply (Vizio OS at one-in-five TVs). That combination is rare. Most RMNs that copy this lack at least one leg.
  • Does iROAS become a real industry standard or a set of proprietary dialects? This is the bigger ecosystem question. If neutral measurement vendors converge on a shared incrementality methodology, retail media matures. If each retailer keeps its own, buyers stay stuck comparing incompatible numbers.

The council leans: structurally real, narratively oversold. The Vizio supply scale and the measurement-as-moat positioning are the durable takeaways — buyers genuinely need to re-score Walmart as a CTV supply owner, not just a data overlay. The "more DSPs coming" and "consistent measurement everywhere" claims deserve a wait-and-see.

What to verify before treating this as a template: (1) the gap between the first DSP integration and the second — that's the real proof; (2) whether any neutral third party validates Walmart's iROAS; (3) whether competing RMNs adopt compatible incrementality definitions or fork into dialects.

For the broad operator audience: measurement vendors without an incrementality story now have a visible product gap. SSPs should study Magnite's position as the data-distribution pipe. And anyone selling on plain ROAS should expect the conversation to shift under them.

Step 5 — The Prediction

Prediction: By 2026-12-31, Walmart Connect's Magnite/Vizio shopper-data integration will be live on at most one additional DSP beyond Yahoo — not the broad multi-DSP rollout implied — and no neutral third party will have publicly validated Walmart's iROAS methodology.

Revisit by 2026-12-31: We're right if Walmart has added zero or one DSP to this specific Magnite-Vizio sell-side path and iROAS remains Walmart-graded. We're wrong if two or more additional DSPs are live on the integration, or an accredited independent measurement body (MRC, or a neutral vendor like VideoAmp/iSpot under audit) publicly certifies the incrementality methodology.

Sell-side data integrations are reversible and easy to announce but hard to replicate per-DSP — the Operator's reconciliation problem and the line-item complexity flagged in current integration work both slow the second and third connections. And data owners rarely hand their measurement to a neutral referee voluntarily, because self-grading is precisely where the leverage lives.

Listen to the episode →