First, segment view rate by format, device and placement
View rate is views divided by impressions, and it only means something for formats that count views. The ad formats guide sets out which formats are bought per view and which per thousand impressions; remove bumpers and non-skippable ads from the calculation before judging anything. Then break the remaining figure down three ways in the reports. By ad, to see whether one cut is failing while another holds. By device and placement, to see whether mobile feeds, Shorts or partner inventory account for the weak impressions. And by audience segment, including observation segments, to see whether the loss sits with one group.
Two reading habits save time. Compare like with like: a skippable ad bought per thousand impressions under a reach goal and the same ad bought per view under a consideration goal are optimised for different outcomes and will show different rates. And read the retention curve beside the view rate where the report offers it: a curve that falls off a cliff in the opening moments is a creative problem, while a curve that holds for some audiences and not others is a targeting problem. The YouTube ads hub draws the same line between format and targeting.
Diagnosis table
Diagnosis
| Symptom | Check | Likely cause | Fix |
|---|---|---|---|
| Viewers leave almost immediately in every audience and placement | Retention in the opening moments, by ad | Opening seconds that do not hold attention | Re-cut the opening to show the subject and the reason to watch at once |
| Viewers stay a little, then skip before anything about the brand appears | When the brand and the point first appear in the edit | Brand or point arriving too late in a skippable ad | Move the brand and the core claim before the skip button appears |
| View rate weak across most segments, with spend spread thinly | Audience report: segment sizes, optimised targeting share, stacked layers | Audience too broad or layered without a reason | Narrow to one intent per ad group, or remove the layers that add nothing |
| Rate far lower on mobile feeds and Shorts than on desktop in-stream | Placement and device split; whether a vertical asset exists | Landscape ad cropped on vertical and Shorts placements | Supply a vertical cut built for the placement |
| Low view rate on a campaign whose goal is awareness or reach | Campaign goal, subtype and bidding model | Campaign goal or subtype built for reach, not views | Judge it on reach, or rebuild under a view-based goal |
| Rate collapses on a remarketing campaign over time | Reach and frequency report; list size; frequency cap | Frequency too high on a small remarketing list | Cap frequency, widen or rotate the list, refresh the creative |
Fix, verify and prevent each cause
An opening that loses people at once
A skippable ad competes with the video the viewer chose, and the opening moments decide whether they wait. Establishing shots, slow logo animations and a narrator warming up are skipped before anything happens. The symptom is consistent: retention drops steeply in the first moments for every audience and placement, so the problem travels with the creative.
Fix: re-edit so the first frame shows the subject, a person or the product in use, and a reason to keep watching, and make it work without sound; the creative constraints in the ad formats guide set out what each format must do before the skip button appears.
Verify: run the new cut beside the old in the same ad group and compare retention in the opening moments and view rate on the same audiences.
Prevent: review every new edit as a muted phone viewer would, watching only the opening, before it is uploaded.
The brand and the point arrive too late
Some ads hold attention well enough but save the brand, the product or the claim for the end, so the people who skip, often the majority, see a story with no name attached. A view-rate problem and a brand problem at once.
Fix: bring the brand and the single most important claim forward, before the point at which the viewer can skip, and repeat the brand visually through the ad rather than only in the end frame; the YouTube ads hub puts the same rule among its creative principles.
Verify: view rate holds or improves, and where a Brand Lift study is available, recall among exposed viewers rises.
Prevent: write the opening as its own short ad first, then extend it, so the brand is designed in rather than added last.
An audience that is too loose or stacked without a reason
A broad affinity segment run per view will reach many people with no reason to watch, and a campaign that stacks an in-market segment, a topic and a placement list in one ad group may be forced into whatever inventory is left. Both show as a weak rate spread across delivery. Optimised targeting can add to the spread.
Fix: read the audience report, then give each ad group one intent, such as a search-term custom segment or an in-market segment, and remove layers that do not change who is reached, following the layering table in the audience targeting guide. The audience targeting help page describes the segment types.
Verify: view rate rises in the narrowed ad groups and cost per view (CPVCost of each video view.) holds or falls.
Prevent: add new segments in observation mode first and promote only those whose view rate beats the campaign.
A landscape film cropped in vertical placements
Video campaigns can serve into the Shorts feed and mobile feeds, and a landscape ad placed there is letterboxed or cropped, with text cut off and the subject small on a phone held upright. Viewers swipe on. The placement and device split shows the damage: in-stream on desktop and television screens looks fine, mobile feeds and Shorts look poor.
Fix: supply a vertical cut with on-screen text and the subject framed for the upright screen, built rather than cropped, as the ad formats guide describes for Shorts. The video ad formats help page is the reference for current asset requirements.
Verify: view rate in vertical placements moves towards the in-stream figure once the vertical asset serves.
Prevent: make a vertical version part of every video brief, not an afterthought.
A goal that was never trying to win views
The campaign goal decides how Google buys. Reach subtypes buy cheap impressions in short formats and aim to show the ad to as many people as possible; they do not chase viewers who will watch. Judging a reach campaign by its view rate compares it with a target it was never set.
Fix: check the goal and subtype; if the brief is awareness, report the campaign on ReachUnique people who saw an ad., frequency and cost per thousand impressions instead, and if the brief genuinely needs engaged viewing, rebuild it under a consideration goal bought per view, as the YouTube ads hub explains in its section on goals and subtypes.
Verify: each campaign is reported against the measures its goal optimises.
Prevent: write the goal and its success measure into the campaign name or brief, so nobody later judges it by another goal's numbers.
The same viewers, again and again
Remarketing lists are small, and a video ad served to the same people many times a week wears out quickly: the first exposures are watched, later ones are skipped on sight. The view rate on the campaign declines steadily while reach barely grows.
Fix: set a frequency cap per person per week, widen the list or rotate several ads through it, and exclude recent converters, as the frequency section of the audience targeting guide recommends.
Verify: the reach and frequency report shows exposures per person falling to the cap and the view rate recovering.
Prevent: put a FrequencyAverage times a person saw an ad. cap on every remarketing campaign at launch and review it when the list size changes.
Prevention: creative built to the opening rules
Most low view rates are decided in the edit suite, so prevention starts with the brief. For every skippable ad, the brief names the opening image, the moment the brand appears and the single claim, all of which must land before a viewer can skip. Every ad is delivered in at least two shapes, landscape for in-stream and vertical for Shorts and mobile feeds, and every edit is checked on a phone with the sound off. Shorter companion cuts, such as a bumper for frequency, are written as their own films rather than trimmed from the long one.
The campaign side has its own habits. Choose the goal from the brief and report each campaign on that goal's measures. Keep one intent per ad group so the audience report can show which group watched, and run new segments in observation first. Cap frequency on every remarketing campaign. Then set a review rhythm: in the first days of a flight, read view rate by ad, placement and device and pause the cut or placement that is failing; afterwards, read it fortnightly beside the retention curve, so a tired creative is replaced before the rate slides. When a new cut is made, test it against the current one on the same audience rather than replacing it outright.
A fitness app's skippable ad recovers its viewers
A fitness app runs a forty-second skippable ad under a consideration goal at £0.05 per view, with a 14% view rate against the 25% to 30% its previous campaign achieved. Segmented, desktop and TV in-stream impressions show a 22% rate, while mobile feeds and Shorts, which carry 58% of impressions, show 8%: the ad is landscape only. The retention curve also shows half of viewers leaving within three seconds, during a slow sunrise shot, and the app's logo first appears at second 31. The team re-cuts the opening to start on a person mid-workout with the app on screen, puts the logo in the corner from the first frame, and adds a vertical twenty-second cut. Over the next three weeks the view rate reaches 27% overall and 24% in vertical placements, and cost per view falls to £0.03.
Because view rate links the two pricing models, the CPM versus CPV comparison shows how a weak rate changes the effective cost of a buy on views.
Common mistakes
Averaging bumpers into the view rate. Judging a reach campaign by views. Fixing the audience when the opening loses everyone. Cropping the landscape master for Shorts. Leaving remarketing without a frequency cap. Replacing the whole ad when only the first moments need re-editing.
