DV360 Low-Converting Ads: Audit Where Spend Goes

PPC Land reports that a study found roughly half of Kroger's combined Meta and DV360 spend went to ads that converted poorly. The headline number belongs to one very large advertiser, but the pattern behind it is common: budget piles up behind a handful of creatives and audiences, and nobody checks whether those are the ones producing results. If you run Display & Video 360, the useful move is to look at your own creative-level spend before someone else asks you to.

Key takeaways

• PPC Land reports a study finding about half of Kroger's Meta and DV360 spend went to low-converting ads. Treat it as a prompt to audit, not a benchmark to copy.
• Campaign-level ROAS hides weak creatives. Audit at the creative, line item and audience level before changing bids.
• Reallocate in stages. A spend-versus-outcome view per creative tells you where to start.
• Judge creatives on conversion quality and incrementality, not clicks or last-touch credit alone.
• More formats, including the vertical video unification The Tech Buzz covers, mean more variants to govern.
• Reporting and test design are in-house work for most teams. Conversion plumbing and cross-platform measurement are where a partner helps.

What does the PPC Land report on Kroger and DV360 actually tell you?

The headline is simple. According to PPC Land, a study found that half of Kroger's Meta and DV360 spend went to low-converting ads. That is a statement about how one advertiser's money was distributed across creatives, not a claim that DV360 or Meta underperforms. Read PPC Land's original report for how the study defined low-converting and what period it covered, because those definitions decide how much of it transfers to your account.

What does transfer is the shape of the problem. Programmatic and social buying systems optimise toward whatever the account tells them to optimise toward. If a creative gets early delivery, it collects impressions, and a large share of spend can end up behind it long after better options exist. Nothing in the platform is broken. The account is just doing what it was set up to do.

So the question for your team isn't whether you are Kroger. It's whether you could answer, today and in under an hour, what share of last month's DV360 spend sat behind creatives that missed your own cost-per-outcome target. Most teams can't. That gap is the finding worth acting on.

Why does spend end up behind weak ads in Display & Video 360?

Several ordinary habits cause it, and none needs a villain.

The first is reporting at the wrong altitude. A campaign can hit its target while a third of its creatives drag the average down, and a dashboard that stops at campaign level will never show it. Strong performers mask weak ones.

The second is early-delivery bias. Creatives that win the first few days get more impressions, which gives them more data, which keeps them winning. If the early winner was lucky on audience or placement, the account keeps funding it.

The third is optimising to a shallow event. If the goal is a page view, an add-to-cart or a click, a creative can look excellent while producing little that matters downstream. A grocery or retail advertiser, for instance, may see plenty of cheap engagement that never turns into purchase.

The fourth is inertia. Creative refreshes are work, and retiring something that has always run feels riskier than leaving it. So the old banner set keeps its budget share for another quarter.

The fifth is multiplying variants without a review rhythm. Google's reported vertical video unification for DV360, covered by The Tech Buzz, is a reminder that formats keep coming. Check Google's own DV360 documentation for what is available to your account and how it behaves. Whatever the specifics, more formats mean more variants, and more variants mean more places for spend to hide.

Who should act on this, and who can ignore it?

Not every team needs a full audit. Use the table below to decide how urgent this is for you.

If you own media performance, creative operations or analytics, this lands on you. If you are the person who signs off budgets but never sees a creative-level report, ask for one. And if your account is small and tightly managed, don't manufacture a project. Read the PPC Land piece, nod, and carry on.

How do you audit creative-level spend in DV360?

Do this in a spreadsheet or a saved report before you touch any setting. The aim is a single view that puts spend next to outcomes for every creative.

1. Pick a window. Use enough days that each creative has had a fair run, and avoid periods distorted by a sale or a tracking outage.
2. Pull spend and outcomes by creative. Break out line item and audience as well, so you can see where the same creative performed differently in different pairings.
3. Choose one outcome measure. Cost per qualified conversion beats cost per click. Agree it with stakeholders before you look at the numbers, so the target isn't picked to flatter the result.
4. Rank by spend. The top ten creatives by spend usually tell you most of the story. Mark each as above or below your target.
5. Check the paired variants. For each weak creative, find a sibling running in the same line item. If the sibling does much better, the creative is the issue. If both are weak, look at audience, placement or frequency.
6. Confirm the conversion tags. Make sure the same events fire for every variant. A creative can look weak just because its landing page tags differ.
7. Write down the share. The number you want is simple: what portion of spend sat behind creatives below target. That's your version of the PPC Land headline.

If that share is small, you're in good shape and can stop. If it's large, you've found your budget lever.

How should you test and roll out changes without breaking delivery?

Resist the urge to cut every weak creative on Monday. Stage the change so you can tell what caused what.

Two cautions. First, give new variants a fair start. If a fresh creative enters against a mature one with a long data history, it can look worse for reasons unrelated to quality. Second, keep one variable moving at a time. If you change creative, audience and bid strategy together, you won't know which did the work.

Where you can, hold back a small comparison group, such as a geography or audience slice that keeps the old mix. It gives you something to measure against that doesn't depend on platform-reported numbers.

How do you measure whether the reallocation worked?

Platform-reported conversions are a starting point, not a verdict. DV360 and Meta each count conversions under their own attribution rules and windows, so their totals rarely agree and often overlap. If you compare them directly, you'll draw the wrong conclusion about one of them.

Use a shared source instead. Your analytics platform, your first-party sales data or a simple holdout test gives you a common yardstick. For a retailer, that could mean matching ad exposure to purchases in a loyalty or CRM dataset. For a lead-gen business, it might mean tracking leads through to qualified status.

Track these over the test window:

• Cost per qualified outcome for the reallocated budget versus the held-back group
• Conversion quality, such as order value, qualification rate or repeat purchase, not just count
• Reach and frequency, to make sure you haven't squeezed the same people harder
• Share of spend behind below-target creatives, which should fall

And be honest about lag. If your buying cycle runs weeks, a one-week test measures almost nothing. Match the window to the decision.

What should you avoid when cleaning up weak ads?

A few mistakes show up repeatedly.

• Don't judge on clicks. A creative with a strong click rate can still be the one that converts worst. Pick the outcome first.
• Don't kill creatives on tiny samples. Low volume looks like low performance. Wait for enough data to be meaningful for your account.
• Don't reset everything at once. Mass changes can disrupt learning and hide the cause of any improvement or drop.
• Don't copy someone else's benchmark. The Kroger figure is one advertiser's result in one study. Your own account data is the only benchmark that matters for your decisions.
• Don't assume the platform is the problem. Often the issue is how the account was structured or what it was told to optimise for.
• Don't skip the tagging check. Mismatched conversion events make an audit meaningless.

Where does an implementation partner help, and where can you do it yourself?

Most of this audit is in-house work. Building a creative-level report, ranking spend, pairing variants and running a staged budget shift needs a careful analyst and a decisive media lead, not a consultancy. If your problem is that nobody has looked, look.

A partner becomes worth the fee when the problem is structural. Think inconsistent conversion tagging across properties, offline or in-store outcomes that never reach the platform, cross-platform measurement that needs a shared model, or an account so large and loosely named that nobody can tell what's running. Those jobs need DV360 configuration knowledge, data engineering and the patience to document what was done.

If you do go looking, write the brief around the outcome, not the tool. Say what you want to know, such as the share of spend below target by creative, and what data you can provide. The DV360 RFP template is a reasonable starting point for that brief. You can browse DV360 partners to see who covers measurement and creative operations, or use find partners if you want to filter by region and capability. For more write-ups like this one, see the MarTech Partners blog.

The sensible order is this: run your own audit first, learn what your share of weak spend is, and then decide whether the remaining problem is one you can fix yourself. A partner who arrives after you've done that spends their time on the hard part.

Sources

• PPC Land — Report that a study found half of Kroger's Meta and DV360 spend went to low-converting ads.
• The Tech Buzz — Coverage of Google's vertical video unification for Display & Video 360.