Video Action Campaigns are gone and five features went with them

Video Action Campaigns finished migrating to Demand Gen in April 2026, and five capabilities did not survive the move: lead forms, seasonal event targeting, new customer acquisition, shared budgets and frequency caps. If your video programme relied on any of them, the workaround is now your problem to build.

Most of the coverage treated this as a rename. It is not. The campaign type changed, the feature set narrowed, and the reporting history broke.

What actually happened to your campaigns

The upgrade preserved creative assets and rough targeting intent, and reset almost everything else. Campaign IDs changed, which means your historical reporting no longer joins cleanly to the new campaigns in any dashboard built on campaign ID. Learning restarted. Bidding history did not carry over in the way most people assumed.

Before you do anything else, check whether your reporting layer is silently comparing two different things. If a Looker Studio dashboard or a warehouse table keys on campaign ID, your year-on-year video numbers are currently wrong, and nobody has told you.

The five losses, and what to do about each

Lead forms. Video campaigns could capture a lead inside the ad unit. Without that, the traffic has to land somewhere and convert there. Which means the landing page is now doing the work the ad unit used to do. The practical implication is speed and form length: a video-sourced visitor is lower intent than a search visitor and abandons faster. Cut the form to the minimum viable fields, put it above the fold, and make sure the page loads properly on a mid-range Android over mobile data, because that is who is watching.

Seasonal event targeting. You can no longer tell the campaign that a specific date range is special and have it adjust. The replacement is manual and it works: ad scheduling for the window, a deliberate budget step-up, and separate campaigns for the peak so the reporting stays clean. Put the dates in a calendar with an owner, because this is the feature most likely to be forgotten until the week after the event.

New customer acquisition. The goal setting that let you bid differently for people who had never bought is gone from this campaign type. Rebuild it with customer list exclusions plus value rules. Upload your customer list, exclude it from prospecting campaigns, and use value rules to weight new-customer segments where they are still available. It is coarser than the native goal and it captures most of the benefit.

Shared budgets. Managing a pooled budget across campaigns has to be done by hand now, or through portfolio bid strategies where they apply. If you were using shared budgets to let a group of campaigns self-allocate, expect to spend real time each week doing that allocation yourself. Set a fixed review day rather than reacting.

Frequency caps. This is the one that quietly costs money. Without a cap, exposure management moves to the two levers you still control: audience size and creative rotation. A small audience with one creative will hammer the same people. Widen the audience, run at least three distinct creatives per audience, and watch reach against impressions rather than impressions alone. If impressions climb while unique reach flatlines, you are paying to annoy people who already said no.

The creative shift nobody budgeted for

Demand Gen is a different surface mix. It runs across YouTube, Shorts, Discover and Gmail, which means a single 16:9 hero video is no longer a complete asset set. You need vertical, you need square, and you need static images that stand on their own.

Businesses that had one polished brand film and treated it as the campaign are the ones seeing the sharpest performance drop, and the cause is not the algorithm. The asset set is incomplete, so delivery concentrates on the few placements the assets fit, and the campaign never reaches the inventory that was supposed to make it worth running.

Budget for a proper asset set. Vertical cut, square cut, three thumbnail-quality stills, and at least two hooks tested against each other. That is a production brief, not an export setting.

The transition dip, and how long to tolerate it

Performance usually drops through a forced migration and recovers over several weeks as learning re-establishes. That is expected and it is not a reason to intervene on day four.

What is a reason to intervene: reach flat while frequency climbs, spend concentrating on one placement, or conversion rate falling while click-through holds. The first is your missing frequency cap. The second is an asset gap. The third is the missing lead form showing up as landing page friction.

Give it 30 days on a stable budget with a complete asset set, then judge it. If you are still down after that with all three diagnostics clean, the campaign is not the problem and the offer probably is.

Pull your video reporting join keys before your next board pack, because explaining a data break after the meeting is much harder than fixing it before.