Stillwater Media creative macro photo of a brass loupe magnifying a watch movement on emerald velvet, illustrating are view-through conversions real

Are View-Through Conversions Real? A CMO's Guide

Stillwater Media•2026-11-18•10 min read

Look closer before you count it: are view-through conversions real? Only a holdout test can tell influence from coincidence.

Are view-through conversions real? It is the question every CMO eventually asks when a display or streaming report shows hundreds of conversions from people who never clicked anything. The honest answer is: some are, many are not, and the share that is real varies enormously by channel, audience and window. For a brand selling private aviation cards, luxury residences or wealth management relationships, the difference matters, because view-through credit often decides whether upper-funnel media keeps its budget.

This guide explains what a view-through conversion actually measures, why the number is usually inflated, where it understates real impact, and how to test the truth for your own accounts.

What is a view-through conversion?

A view-through conversion (VTC) is a conversion credited to an ad that a person saw, or was served, but did not click, within a defined lookback window. If a household is served a streaming spot on Monday and someone in it submits an inquiry form on Thursday, a platform with a 7-day view window will report one view-through conversion.

The mechanism is simple. The platform logs an impression against an identifier (a cookie, mobile ad ID, IP-based household graph or hashed email), the advertiser's pixel or conversion import logs a conversion against a matching identifier, and the platform joins the two if the conversion falls inside the window. Nothing in that join proves the ad caused the conversion. It proves only that the two events happened to the same identifier in sequence.

That is the crux of the debate. A view-through conversion is a correlation report, not a causal one.

Are view-through conversions real, or just correlation?

Think of view-through conversions as three groups mixed together:

  • Caused conversions. The person saw the ad, it changed their behavior, and they converted. These are real.
  • Coincidental conversions. The person was going to convert anyway and happened to be served an ad first. Retargeting is the classic example: the people most likely to buy are exactly the people the algorithm serves most.
  • Phantom conversions. The impression was never meaningfully seen. It loaded below the fold, played on a television that was off, or was delivered to a bot or a made-for-advertising site. The conversion would have been credited regardless.

Platforms report all three together. The job of measurement is to estimate what share is the first group. In our planning experience, that share runs anywhere from under 10 percent for broad retargeting to well over half for premium CTV reaching new households. The number on the dashboard tells you none of this.

Why are view-through conversions usually inflated?

Several structural forces push reported VTCs above true impact:

  • Targeting selects converters. Optimized campaigns find people already in market. A wealth management prospect who has visited your site four times will be served more ads and will also convert more often. The ad gets credit for intent it did not create.
  • Every platform claims the same sale. Meta, Google, your DSP and your CTV partner can each report the same inquiry as their own view-through conversion. Summed platform reports routinely exceed actual sales, a problem we cover in our guide to incrementality vs attribution.
  • Long windows sweep in coincidence. A 30-day view window on a high-frequency display buy will touch most of your converting audience at some point. Our analysis of attribution window length for high-consideration purchases explains why view windows should stay much shorter than click windows.
  • Viewability is not the same as viewing. An impression counts as viewable under the industry standard if half its pixels are in view for one second (two seconds for video). That is a low bar for a $40,000 decision.
  • Low-quality inventory is cheap and prolific. Made-for-advertising sites and cluttered pages generate huge impression volume per dollar, which multiplies the chances of a coincidental match. The ANA's programmatic transparency research found that a material share of open-web spend landed on such sites.

When do view-through conversions understate real impact?

The skeptics are not always right. In some situations view-through reporting misses real effect:

  • Premium CTV on the big screen. A streaming ad on a living-room television cannot be clicked. Its effect shows up later as branded search, direct site visits or a phone call, often from a different device. Household-level matching catches part of this; short windows and device-level matching miss much of it.
  • Long sales cycles. If your buyer takes 60 to 180 days to commit, a 1-day or 7-day view window will capture only the tail of the effect.
  • Offline conversions. Showroom visits, advisor meetings and signed contracts often never reach the pixel at all unless you import them. Our guide to CTV CRM matchback covers how to close that gap.

So the right posture is neither belief nor dismissal. It is testing.

How do platforms count view-through conversions in 2026?

Settings differ by platform, and defaults change. Here is the current landscape as we understand it; verify in each account before relying on it.

PlatformDefault view-through windowWhat counts as a "view"Common distortion
Google Ads (Display, YouTube, Demand Gen)1 day for most conversion actions; adjustableImpression served, with engaged-view rules for YouTubeBranded-search buyers credited to YouTube views
Meta1-day view in the default 7-day click / 1-day view settingImpression delivered in feed, Stories or ReelsHeavy overlap with organic and email converters
Programmatic DSPs (e.g. The Trade Desk, DV360)Advertiser-set, often 7 to 30 daysImpression served, sometimes filtered for viewabilityLong windows on high-frequency display
CTV platforms and CTV-specialist DSPsAdvertiser-set, often 7 to 30 days, household IP matchAd served to a household deviceShared IPs and always-on retargeting inflate matches

The pattern is consistent: the advertiser usually controls the window, and the platform's incentive is to make the number look healthy. Google documents its settings in its conversion window help article.

Are view-through conversions real in your account? How to test it

The only reliable way to answer "are view-through conversions real?" for your account is a controlled experiment that compares exposed and unexposed groups. Four methods, in rough order of rigor:

1. Ghost-ad or ghost-bid holdouts

The platform withholds ads from a randomly assigned control group but logs the auction it would have won. You then compare conversion rates between the exposed group and the would-have-been-exposed control. This is the method described in the academic work of Johnson, Lewis and Nubbemeyer on ghost ads, and several DSPs, including The Trade Desk's conversion lift product, now offer versions of it without paying for public service ads.

2. PSA holdouts

The control group is served a charity or blank ad instead of yours. It works, but it costs media and can be biased if the platform optimizes delivery differently for the PSA creative.

3. Geographic holdouts

Turn the channel off in matched markets and compare outcomes. This is the best option for CTV and for brands whose conversions happen offline. Our guide to geo experiment design covers market matching and test length.

4. Synthetic control and media mix modeling

When clean holdouts are impossible, statistical methods can estimate the counterfactual. They are weaker than randomized tests but stronger than raw attribution. See our explainer on holdout testing in advertising for where each fits.

What does a realistic view-through incrementality look like?

Every account differs, but these Stillwater planning ranges help set expectations before a test. They describe the share of reported view-through conversions that typically prove incremental in a well-run holdout:

Channel and tacticTypical reported VTC share of conversionsShare likely incremental (Stillwater planning range)
Site retargeting display, 30-day view windowHigh5% to 20%
Open-web prospecting displayModerate10% to 30%
Paid social prospecting, 1-day viewModerate20% to 45%
YouTube prospecting with engaged-view rulesModerate25% to 50%
Premium CTV to net-new affluent householdsHigh, often most of the channel's credit35% to 70%

Treat these as starting hypotheses, not facts. The point is that one blanket rule, such as counting view-throughs at 50 percent, is almost always wrong for at least one channel in the mix.

Which signals make a view-through conversion more credible?

Before a holdout is possible, a few quality signals help sort stronger view-through credit from weaker:

  • Time to convert. VTCs clustered within hours of an impression on a high-frequency retargeting line are weaker evidence than conversions from net-new households that arrive days later via branded search.
  • New versus known users. A conversion from someone with no prior site visits, email engagement or CRM record is far more likely to reflect real influence than one from a returning visitor.
  • Placement quality. Full-screen, non-skippable CTV on premium apps and verified publisher placements deserve more weight than below-the-fold display.
  • Frequency. If most VTCs come from households served 15 or more times, the campaign is probably following converters rather than creating them.
  • Downstream quality. View-through inquiries that become qualified opportunities at the same rate as click-through inquiries suggest real intent rather than noise.

None of these replaces a test, but together they tell you where to run one first.

What are the most common mistakes with view-through conversions?

  • Summing platform reports. If four platforms each claim the same VTC, the business did not get four sales. Reconcile to a single source of truth, usually the CRM.
  • Using one window for everything. A 30-day view window can be defensible for premium CTV and indefensible for retargeting display.
  • Optimizing bids to view-through conversions. The algorithm will learn to find people who were going to convert anyway. Optimize to incremental outcomes, qualified inquiries or modeled value instead.
  • Ignoring inventory quality. VTCs from made-for-advertising sites, unknown apps or cheap CTV resellers deserve the least trust. Insist on a placement-level report.
  • Applying a fixed discount without testing. Haircuts like "count VTCs at 25 percent" feel prudent but are arbitrary. Run a holdout and calibrate.
  • Throwing out VTCs entirely. Dropping all view-through credit systematically defunds CTV, video and audio, which are the channels most likely to build future demand among affluent buyers.

How should luxury brands report view-through conversions to finance?

Finance teams distrust view-through numbers for good reason. A reporting structure that earns credibility looks like this:

  • Separate columns. Report click-through and view-through conversions separately, never blended into one ROAS figure.
  • Calibrated multipliers. Apply a channel-specific incrementality factor from your most recent holdout, and show the date and design of that test.
  • CRM reconciliation. Tie total credited conversions back to actual opportunities and revenue in the CRM each month, and show the gap.
  • A refresh schedule. Re-run holdouts at least twice a year or after any major change in targeting, creative or inventory.

For wealth managers in particular, the evidence trail also supports compliance reviews; see our wealth management advertising approach.

A worked example

Consider a hypothetical luxury residential developer running premium CTV, prospecting display and retargeting. Over a quarter, platform reports show 420 view-through inquiries: 210 from retargeting, 90 from display and 120 from CTV.

A geo holdout on CTV and a ghost-bid holdout on display and retargeting produce these illustrative incrementality factors: retargeting 12 percent, display 22 percent and CTV 55 percent. Applied to the reported numbers, true incremental view-through inquiries are roughly 25 from retargeting, 20 from display and 66 from CTV.

The reported ranking put retargeting first. The tested ranking puts CTV first by a wide margin. That reversal is common, and it is the main reason to test rather than argue.

So, are view-through conversions real?

Yes, partly. View-through conversions contain real influence, coincidence and noise in proportions that differ by channel. Platforms will not separate them for you. The brands that allocate well are the ones that treat every view-through figure as a hypothesis, test it with holdouts, and report calibrated numbers that finance can trust.

If you want a measurement plan that tells you which of your view-through conversions are real before you move another dollar, apply to work with Stillwater Media. We take on a limited number of engagements each quarter.

About the author: Stillwater Media is a selective performance media agency for luxury and high-consideration brands, specializing in premium CTV, programmatic, affluent audience engineering and incrementality testing. Signal. Strategy. Scale.

Frequently Asked Questions

Are view-through conversions real?

Partly. A view-through conversion records that an identifier saw an ad and later converted, which mixes real influence with coincidence and unseen impressions. Only a holdout test, such as a ghost-bid or geo test, shows what share is incremental.

Why are view-through conversions often inflated?

Optimized campaigns serve the most ads to people already likely to convert, and every platform can claim the same sale. Long view windows and low-quality inventory add coincidental matches on top.

Should view-through conversions count for CTV?

Yes, with calibration. CTV ads cannot be clicked, so view-through and household matchback are the main ways to see their effect, and holdout tests often show CTV is more incremental than retargeting.

What is a good view-through attribution window?

Keep view windows short: usually 1 day for social and display and up to 7 days for premium CTV. Longer windows capture more real effect but much more coincidence, so calibrate with tests.

How do you test view-through conversions?

Use a randomized holdout: ghost bids or PSA ads for digital channels, or matched geographic markets for CTV and offline conversions. Compare conversion rates between exposed and control groups to find the incremental share.

Ready to discuss your strategy?

Discover how our approach can transform your brand's media performance.

Related insights