View-through conversions: when to count them, when to ignore them (saw it is not caused by it)

Chinmay R. · has watched a toggle erase a third of conversions👁️ with a click-vs-view split, not a lecture

Published September 16, 2026

The 60-second version

  • A view is not a cause. Same spend, same week — switching off view-through removes 30–40% of claimed conversions.
  • Views concentrate in retargeting and brand pools, flattering the harvester most while prospecting stays click-led.
  • The fix: set targets on click-only ROAS, keep views as a creative diagnostic, report finance on warehouse MER.
bottom line: clicks decide budgets, views inform creative

Your Meta dashboard claims 520 conversions. You switch off view-through, and it claims 340. Same spend, same week, 35% of conversions vanish with one toggle.

Nothing about your business changed. A view is not a cause — counting every ad seen as an ad that worked manufactures conversions out of coincidence. The toggle did not remove sales; it removed guesses.


What a view-through really is

A click conversion means the customer clicked, then bought. A view-through conversion means the customer saw the ad, did not click, then bought within a short window — typically 1 day on Meta, up to 30 on some display platforms.

Think of the billboard outside a coffee shop. You drive past it every morning. One morning you stop and buy a latte. Did the billboard cause the purchase — or were you already a coffee drinker driving that road anyway?

Three journeys, one credit rule

Data Journey
Stage 1Click
Clicked, then bought

Saw the ad, clicked through, purchased within the click window. The strongest causal claim.

Count it
Stage 2View
Saw, ignored, searched, bought

Scrolled past the ad, later searched the brand and bought. The ad may have nudged — or merely coincided.

Report separately
Stage 3Baseline
Would have bought anyway

Existing customer sees retargeting daily, buys on schedule. The view gets credit for a habit.

Never count

Platform defaults decide how much of your report is journey two and three. Meta ships 7-day click plus 1-day view; Google Display and YouTube push longer view windows; your warehouse counts none of it unless you tell it to. Same behaviour, three numbers. saw it is not caused by it

View-through is modelled furthest from observation. No click ID ties the view to the order — only a probabilistic device match plus a time window. It is the first number to inflate when tracking degrades, and the last number finance should set targets against.


Why view-through flatters retargeting most

View-through does not spread evenly. It concentrates exactly where customers already live: retargeting pools, brand audiences, existing buyers.

Retargeting pools see everything

Site visitors are served daily. When they buy on schedule, a view sits within 1 day almost by construction — coincidence scored as causation.

Brand searchers brush ads

A customer about to search your name scrolls past a Meta ad first. The search closes the sale; the scroll takes the view credit.

Heavy users collect views

Your most active buyers see the most impressions. View windows credit the channel for the customers who needed it least.

Numbers from a typical D2C month: prospecting reports 12% of conversions as view-through, retargeting reports 38%. Strip views and prospecting falls 12%, retargeting falls by more than a third. The pool that looked unbeatable was the pool most stuffed with views.

180 conversions — 35% of the headline — never involved a click.

The 4.1x ROAS that was 2.7x without views

Same week, one toggle
Spend₹5,60,000Meta prospecting + retargeting, one week
Claimed revenue (with views)₹22,96,000520 conversions at ₹4,415 AOV — 4.1x ROAS
Click-only revenue₹15,01,000340 conversions — 2.7x ROAS
View-through share35%180 conversions with no click behind them
Retargeting view share46%Prospecting view share just 14%
Breakeven ROAS3.1xClears with views, fails without them
Against a 3.1x breakeven floor the toggle flips the verdict: scale on the first number, kill on the second. Targets must be set on the second.

When to count them, when to ignore them

View-through is not always fiction. Video, display and social genuinely work by being seen — nobody clicks a 6-second bumper and buys a sofa the same second. The question is never count-or-ignore in general. It is which decision each version serves.

Optimise bids on views

🙈 Views inside the target

Bidding learns to buy coincidence. The algorithm discovers the cheapest views are past buyers — and fills your budget with them.

  • Retargeting hogs spend
  • Prospecting starves quietly
Optimise on clicks, read views aside

🔍 Clicks decide, views inform

Targets clear on click-only; views get a separate diagnostic column. Reach still shows up — it just cannot outvote behaviour.

  • Bids chase actions, not eyeballs
  • Creative read keeps its view lens

Count views where attention is the product. Never let them vote on budgets.

The rule, channel by channel:

ChannelCount view-through?How
Meta / TikTok prospecting videoYes, directionally1-day view as a secondary column, never in the ROAS target
Retargeting (any platform)NoClick-only. The pool already guarantees views; adding them double-counts intent
YouTube / Demand GenYes, cappedView conversions reported separately with a frequency floor note
Search / Shopping / PMax brandNoClick-only. Navigational intent needs no view assist story
Finance / board reportingNeverWarehouse revenue over warehouse spend (MER) — no windows at all

Never sum click and view conversions across platforms. One customer who saw a YouTube ad, scrolled past Meta, then clicked search is claimed as a view by two platforms and a click by a third. Summing manufactures three customers out of one.


Reporting them without fooling finance

The permanent fix is a three-column report, every Monday, from numbers you own. One query, no toggles to forget.

Click vs view split from platform exports

Three columns, one verdict

Read it in order: sort by view_share descending. Anything above 30% view share gets its ROAS read on the click-only column — the with-views column is commentary. Then check MER for the day: if platform ROAS climbs while MER stays flat, the extra conversions were reallocated credit, not created demand.

LIVE · DRAG ITHow much of your ROAS is views?
₹5.6L
340
180
ROAS click-only
2.68x
ROAS with views
4.1x
View share
34.6%
VIEW-STUFFED

Nearly a third or more of conversions never clicked. Set targets on click-only and move the views to a diagnostic column.

live mathsestimates are fine — this is a what-if sandbox

Say what to ignore: ignore cross-platform view sums, 28-day view windows on performance reports, and any creative test judged on views alone. Keep views for one job — reading whether video creative gets noticed — and take away their vote everywhere else. drag me — your split changes the verdict!


Quick gut-check

One question. Get it and the whole post clicks. 30 seconds, no maths!

A campaign reports 520 conversions, but 180 are view-through. Spend is ₹5,60,000 at ₹4,415 AOV. Which ROAS should set the budget?


Frequently asked questions

Is a 1-day view window honest but 7-day view dishonest?

Shorter is less wrong, not right. A 1-day window still credits coincidence — the buyer who scrolled past retargeting in the morning and bought on schedule at night. Prefer click-only for targets at any window length, and keep even 1-day views in the diagnostic column.

Should video campaigns be judged click-only too?

For budgets, yes. For creative, no. Use view-through rate, thumb-stop and watch time to judge whether the creative gets noticed — then judge whether the noticed creative creates orders via click-only ROAS and geo lift. Attention metrics pick winners; action metrics fund them.

Why do platforms default to including views?

Because bigger numbers retain budgets, and because views do contain some signal. Defaults serve the platform's optimisation loop, not your P&L. Changing the default column set once — click-only for targets — is a one-time fix that pays every Monday.

Can I compare view-through across Meta and Google?

No. Each defines view, window and matching differently, and both claim the same customer. Compare click-only within a platform over time, and compare platforms only on warehouse MER. Cross-platform view sums are triple-counting with extra steps.

What view share should worry me?

Above 35% the headline is mostly coincidence — act immediately. Between 15% and 35% it is padding — separate the columns and watch clicks. Below 15% it is a sane minority for video-heavy mixes. For retargeting, halve every threshold: 20% is already view-stuffed.


The summary

  • A view-through credits an ad the customer saw but never clicked. Same spend, same week, the toggle alone moves reported conversions 30–40%.
  • Views concentrate in retargeting and brand pools — the customers who needed the ad least generate the most view credit.
  • Bidding on views teaches the algorithm to buy coincidence: past buyers, cheap eyeballs, zero new demand.
  • Judge budgets on click-only ROAS, read views as a separate creative diagnostic, and report finance on warehouse MER with no windows at all.
  • Sort every Monday report by view share descending. Anything above 30% gets read on clicks — the with-views column is commentary.
  • Never sum views across platforms. One customer becomes three conversions the moment two platforms plus search all claim the journey.

Takeaways for your next report

  • A view is not a cause — one toggle moves reported conversions 30–40% with zero business change.
  • Views concentrate in retargeting and brand pools, flattering the harvester most.
  • Set targets on click-only ROAS; keep views as a creative diagnostic, never a budget vote.
  • Report finance on warehouse MER and never sum view conversions across platforms.
stick this on your Monday report
Free tool

MER Calculator

Total revenue divided by total marketing spend — the attribution-agnostic efficiency number, plus its contribution-margin-adjusted variant.

Chinmay Raibagkar

About author →

Founder of DataLens AI. He helps non-technical teams read their ad and database numbers with confidence — which number to trust, what to do next, and what to ignore.