Hellmann's DV360 vertical video test: what it means

A reported 24% gain in unique reach for Hellmann's, from a vertical video test in Google's Display & Video 360, is the kind of number that ends up in a planning meeting by Friday. PPC Land carried the story, and The Tech Buzz has separately covered Google's vertical video unification for DV360. The useful question isn't whether vertical works. It's whether it adds reach for your brand, at a price you'd pay, with creative you can actually produce.

Key takeaways

• PPC Land reports Hellmann's gained 24% unique reach in a Google DV360 vertical video test. Treat it as one advertiser's result, not a benchmark for your account.
• Vertical video only adds reach if you have vertical creative built for the format, not a cropped landscape cut.
• Run your own test with a clean holdout and a unique reach metric agreed before launch.
• Judge the result on cost per incremental reached person and on brand or attention outcomes, not on reach alone.
• Brief a partner for the hard parts (measurement design, creative operations, QA) and keep setup and trafficking in-house if your team already runs DV360 well.

What does the Hellmann's vertical video result actually tell you?

PPC Land reports that Hellmann's saw a 24% gain in unique reach in a vertical video test run in Google's DV360. That is the claim, and it's worth being exact about its shape. It's a lift in the number of different people reached, not in impressions, and not in sales.

A unique reach lift is a plausible outcome for a reason that has nothing to do with magic. Many people already see your landscape video ads often. Add a format that fits how a large share of people hold their phones, and you open up placements and moments where your ad wasn't competing before. The same budget then touches more distinct people and fewer of the same ones again and again.

That logic carries a condition. The gain is relative to a baseline, and baselines differ wildly. A brand that was already capping frequency tightly and buying wide has less room to improve than one that was hammering a narrow audience. A brand with a strong vertical asset will do better than one that stretched a landscape file. So the 24% is a reason to run a test, not a number to put in your forecast.

What the figure doesn't tell you is just as important. It doesn't say what the cost per reached person was. It doesn't say whether attention, recall or conversions moved with it. If you want those details, go to the PPC Land piece and, for how reach is defined and counted, to Google's own DV360 documentation.

Why does vertical video unification matter for DV360 buying?

The Tech Buzz describes the broader move as Google unifying vertical video in Display & Video 360. In plain terms, the direction is toward managing vertical video as part of your normal DV360 video buying rather than as a separate side project. Check Google's documentation for exactly what is available in your account, which inventory it covers and whether any of it is still rolling out, because availability can vary and can change.

For a buyer, the practical point is about workflow. When a format is awkward to buy, teams avoid it, and the learning never happens. When it sits in the same line items, the same reporting and the same frequency controls as everything else, the cost of testing falls. The barrier moves from platform setup to creative and measurement, which is where it should have been all along.

There's a second effect worth planning for. If vertical becomes easy to buy, many advertisers will test it in the same quarter. Auction pressure on that inventory can rise, and early efficiency can fade. That's a reason to test sooner rather than later, and also a reason not to bank on early cost levels staying put.

Who should act now, and who can wait?

Not every DV360 team needs to move this month. A rough sort helps.

The fourth and fifth rows matter most. A team with no clean way to compare test and control will produce a number nobody trusts, and then argue about it. Sorting out measurement is dull, but it's the difference between learning something and having a story.

Agency-side, this is also a conversation to have with clients who run reach campaigns. If you manage several DV360 accounts, a shared test design saves time, and a shared definition of unique reach saves arguments.

How should you set up a vertical video test in DV360?

Keep the design boring. The goal is one clear comparison, not a clever experiment.

1. Pick one campaign objective. Reach or awareness is the natural choice. Don't mix goals in one test.
2. Make real vertical creative. Design the first seconds, the on-screen text, the logo placement and the captions for a tall frame. If you only have landscape assets, a cropped version tests your editing, not the format.
3. Split the audience cleanly. Use a control group that keeps your current video setup and a test group that adds vertical. Use the same audience definitions, geography and flight dates.
4. Fix the budget split in advance. Put enough into the test arm to produce readable numbers. If it's too small, you'll see noise and call it a result.
5. Set frequency controls on both arms. Otherwise a reach difference can come from frequency differences, which isn't what you're trying to measure.
6. Decide your success rule before launch. Write down the metric, the time window and the threshold that would make you scale, hold or stop.
7. Run it long enough. Reach builds over time, so a few days can mislead. Let the flight run through its normal pacing.

A worked example. A food brand runs a six-week reach flight across a national audience. It keeps its existing landscape video line items as the control and adds a test arm using a vertical cut made from the same shoot. Both arms share frequency caps and audience lists. At the end, the team compares unique reach per unit of spend, not total impressions. If vertical reached more distinct people at a similar cost, they scale it. If it reached more people at double the cost, they've learned something different and just as useful.

How do you measure unique reach without fooling yourself?

Start by agreeing the definition. Unique reach in DV360 depends on how the platform counts people across devices and environments, and different tools count differently. Read Google's documentation for how it works in your account, and make sure everyone on the team is quoting the same figure from the same report.

Then choose the numbers that go together. Reach on its own is easy to inflate. Pair it with:

• Cost per incremental reached person. The amount you paid for each additional distinct person the vertical arm reached over the control.
• Average frequency. If reach rose because frequency fell, say so. That's a fine outcome, but it's a different story.
• View quality signals. Completion and viewability measures tell you whether the extra reach was watched or merely served.
• A brand or outcome measure. Brand lift, search interest, site visits or conversions, whichever your business already trusts. Reach that doesn't move anything is worth less than it looks.

Be careful with overlap. If your vertical test runs alongside social campaigns that also use the same creative, the same people may be counted in several places. A clean read needs the DV360 arms to be compared with each other, not with a blended total across channels.

One more discipline: write the result down in the same format you wrote the plan. If the outcome is better than the threshold, scale in steps. If it isn't, don't redefine success afterwards.

What does a sensible rollout look like?

A four-stage rollout keeps risk low and gives you decision points. Owners are suggestions; adjust to your team.

A few notes on the middle stages. QA matters more than usual because new format and placement combinations can behave differently from what you're used to. Check brand safety settings, exclusions and creative rendering on actual devices before spend ramps up. Resist the urge to optimise the test arm halfway through. Every tweak muddies the comparison.

If the result is good, scale in steps rather than flipping the whole budget. Move a further slice, watch whether cost and quality hold, then move again. Early results from a new format often soften as more money enters the same inventory.

What should you not do with this news?

A short list of mistakes that are easy to make.

• Don't copy the 24% into a forecast. It's one reported test from one advertiser. Your own number may differ in either direction.
• Don't crop landscape and call it vertical. You'll test your editing shortcut and conclude the format doesn't work.
• Don't judge on reach alone. A cheap reach gain with weak attention or no brand effect isn't a win.
• Don't change the test mid-flight. Adding audiences, swapping creative or moving budgets turns a comparison into a guess.
• Don't skip the control. Without it you can't tell format effect from seasonality, a promotion or a pacing quirk.
• Don't ignore your other billing and reporting changes. PPC Land has separately reported that Google will drop unrendered ads from DV360 and CM360 billing from February 2027. If that applies to your account, it can shift cost and delivery numbers, so keep an eye on when you compare periods and check Google's documentation for the details.

That last point is a practical one. If you run a test across a period where billing rules change, cost comparisons can drift for reasons unrelated to vertical. Note any such changes in your test log.

Where does an implementation partner help, and where can you go it alone?

A team that already runs DV360 well can do most of this itself. Creating line items, applying frequency caps, setting up a control and reading a reach report are everyday tasks. If you have an analyst who can write a clear test plan, you probably don't need outside help for the first round.

A partner earns their fee in three places. The first is measurement design: choosing a clean control, deciding how to deal with overlap across channels and building a read that finance will accept. The second is creative operations: if you need many vertical cuts, versions and sizes at speed, a production pipeline matters more than platform skill. The third is QA and governance across several accounts or markets, where a missed setting can quietly waste spend.

If you're deciding whether to bring someone in, start by listing what you can't do in-house, not what sounds impressive. A short written brief gets better proposals than a vague request. The DV360 RFP template gives you a structure for scope, success measures and deliverables. To see who works on DV360, browse the DV360 partners and use compare to line up shortlisted firms side by side.

Whatever route you take, ask any partner how they would design the control, which metric they'd report and what result would make them recommend against scaling. A partner who can answer that plainly is thinking about your outcome. One who only talks about format potential is selling.

The position worth taking is a modest one. Vertical video is probably worth a controlled test in your next reach flight, and the Hellmann's figure reported by PPC Land is a fair reason to schedule it. It isn't a reason to rebuild your plan around it before you've seen your own numbers.

Sources

• PPC Land — Report on Hellmann's 24% unique reach gain in a Google DV360 vertical video test
• The Tech Buzz — Coverage of vertical video unification for Display & Video 360