Google Unifies Vertical Video Buying in DV360

Google announced Vertical Video Unification with Display & Video 360 on September 28, 2026. The short version, from Google's own post: an advertiser uploads one 9:16 vertical video asset, and DV360 buys and measures it programmatically across vertical video inventory — YouTube Shorts, streaming apps, premium publishers and the open web — as a single workstream. Gemini handles buying decisions, delivery and full-funnel measurement, including frequency capping at the individual person level rather than per placement.

During initial testing, Unilever's Hellmann's and WPP Media reported a 24% increase in unique reach while reducing cost per unique user by 25% — Google's number, from one brand and its agency, stated as an initial test rather than a benchmark. Agency breakdowns of the launch repeat the same two figures and the same one-asset framing.

The headline is not "vertical video is big". That has been true for years. The headline is that the buying layer stopped caring which surface the vertical asset lands on — which moves the constraint in the campaign one step to the left, onto creative supply.

What changed, precisely

Before this, vertical video buys were fragmented by surface. You built a campaign on YouTube Shorts, another for streaming apps, another for publisher inventory — three targeting setups, three bidding strategies, three dashboards. The agency write-ups describe the three costs of that: the same person seeing the same ad across platforms, paying separately to reach audiences that overlap, and no unified read on which touchpoint drove the conversion.

Unification collapses that into one workstream with one measurement view, and Gemini reallocates spend across inventory sources as performance signals arrive instead of against manual per-channel splits. If you have ever been handed a report where three platforms each claimed the same conversion, you understand why that is the interesting part.

	Siloed buying	Unified vertical buying
Assets	Separate cut per surface	One 9:16 master
Frequency control	Per placement — one person can be hit on five surfaces	Per person across channels
Measurement	Platform-by-platform, overlapping attribution	One workstream-level view
Optimization	Manual budget splits	Automated reallocation by Gemini

The part that lands on your creative team

Here is the consequence nobody puts in the launch post. When buying is unified and frequency is capped per person, the same asset reaches more unique people and repeats to each of them less often. Reach goes up; per-person exposure goes down.

That is good for media efficiency and bad for anyone whose creative plan was "one strong ad, let it run". Fewer repetitions per person means creative fatigue arrives sooner, not later, because the system deliberately spreads your asset thinner across a wider audience. The unit of competition shifts from which surface to how many distinct, on-brand vertical variants you can actually supply.

Three practical implications for an ecommerce advertiser:

1. Ship variants, not re-cuts. Trimming the same 30-second spot to 9:16 gives the algorithm one idea in one costume. Different hooks, different openings, different offers are what a person-level frequency cap can actually distribute.
2. Give the measurement something to grip. Gemini connects vertical video impressions to downstream conversions, but creative-level attribution is still your naming discipline — hook, format and offer in the asset name, so the unified report can tell you which variant earned the conversion instead of just that a conversion happened.
3. Treat the 9:16 master as the primary asset. If your pipeline starts from a horizontal TV cut and reformats down, every variant inherits the horizontal framing decisions. Vertical-first is cheaper to produce and looks correct on the surfaces DV360 now buys in one workstream.

Where Prizmad fits

This is the exact problem Prizmad was built around: not "make one beautiful video", but "supply enough vertical ads that a test stays alive".

You paste a product URL and get a finished 9:16 UGC-style ad in roughly five minutes — 50+ avatars, AI Studio, full commercial rights — then change the hook, the avatar or the offer and generate the next variant from the same product page. Token math, so you can price variants before you commit: a finished AI UGC ad is ≈1,000 tokens, and plans run $39 for 3,000 tokens, $99 for 8,000, $249 for 35,000 (≈$12 per finished ad on Starter). Ten variants stop being a production project and become a line item.

That matters more, not less, on the back of this announcement. A unified buying layer rewards a queue of distinct vertical assets; it cannot manufacture them.

The honest caveats

• DV360 is enterprise programmatic. It needs a seat, a licensing conversation and a media team that can operate a DSP. The Hellmann's/WPP test is a large-advertiser test — if you buy on Meta and TikTok yourself, this launch changes nothing in your account today.
• Google's numbers are Google's, from one brand and one agency, presented as initial testing. Treat 24% and 25% as directional, not as a forecast for your account.
• Unification is about buying, not creative. Nothing in the launch fixes a weak hook, an unverifiable claim or a product page that does not convert.
• Timing pressure is real but generic. Q4 vertical inventory gets more expensive for everyone, and any automated buying system needs a few weeks of data to learn. That argues for starting earlier, not for rushing a creatives backlog you cannot review.

Read the primary announcement rather than the summaries, especially if you are planning media for a brand that actually has a DV360 seat. And read it as what it is: a buying-layer upgrade that raises the value of vertical creative supply — the half of the equation you still own.

If the bottleneck in your account is variants rather than budget, see how a product URL becomes a finished vertical ad, or work through the cost maths in how much AI video ads actually cost. This is the fourth Google-ads change we have covered this quarter, after Asset Studio video dubbing — the pattern is consistent: the platform automates the buying and the assembly, and leaves the offer and the hook to you.