Google is retiring standalone Display campaigns, and the only open question is whether you move on your schedule or on Google’s. Since June, a phased migration tool has let eligible advertisers migrate Display campaigns to Demand Gen as a new channel called “Google Display Network on Demand Gen”; later, new Display campaigns can only be created inside Demand Gen, and after that, remaining eligible campaigns get auto-migrated with no action from the advertiser 1. The feature-loss list is long enough that some Display setups need rebuilding, not porting, and the accounts that treat this as an audit instead of a checkbox will come out ahead.
The window is voluntary. The destination is not.
Every voluntary Google migration eventually stops being voluntary, and this one has a published sequence. The tool rolls out in phases starting June 2026, creation of new standalone Display campaigns then gets locked, and the process runs on a timeline Google says will continue into 2027 2. The end state is auto-migration of whatever remains, without advertiser action 1.
That sequencing is the whole strategic point. Migrating now means you choose which campaigns move first, when the learning period lands, and what gets rebuilt instead of ported. Waiting means Google chooses, and Google’s picker does not know your seasonality, your billing model, or which of your campaigns is quietly held together by a bid adjustment.
What the tool ports, and where it gets clumsy
The Demand Gen migration tool carries over roughly 42 days of performance history, and Google says to expect 1 to 2 days of learning after the move 1. Migrated campaigns get named “[Original name] #2,” originals are set to Removed but kept for reporting, and every ad goes back through review before serving again 1. The GDN channel is locked on during migration; YouTube, Discover, Gmail, and the Maps beta can be added afterward, one channel at a time, so nothing forces you into broader inventory on day one 1.
One pacing quirk deserves a calendar note: on migration day, budgets reset. A campaign that spent $10 of its $50 daily budget before migrating starts fresh with the full $50 that same day 1. On one campaign that is rounding error. Across a large account migrated mid-morning, it is a real overspend day, so migrate early in the day or trim budgets in advance.
Reporting continuity needs its own plan as well. The migrated campaign arrives with a truncated history window, the original stays behind under Removed for reference, and your year-over-year Display comparisons now span two campaign objects with different names. Decide before the move how the two get stitched together in your reporting, because three months from now nobody will remember that “Remarketing Core #2” and “Remarketing Core” are the same program.
Eleven features do not make the trip
The removal list decides who ports and who rebuilds. Demand Gen does not support Manual CPC, viewable-impressions bidding, bid adjustments, seasonality adjustments, portfolio bidding, combined audiences, the observation setting, shared budgets, lead form assets, or click-to-call assets, and pay-for-conversions billing gets switched to pay-for-clicks with Target CPA 1. That last one is the trap. It is not a targeting feature disappearing; it is a billing model changing underneath you, silently, from paying only when someone converts to paying for every click. If any of your Display campaigns run pay-for-conversions, check this before anything else.
Plenty survives the move. Target CPA, Maximize conversions, Maximize conversion value, Target ROAS, and Maximize clicks all carry over, along with placement and topic exclusions, contextual targeting, Customer Match, remarketing lists, and placement reporting 1. A Display setup built on smart bidding plus audience lists plus exclusions ports almost intact. A Display setup built on manual control does not port at all; it gets translated, and the translation loses the parts that made it work.
So the audit before the migration is one question asked eleven ways: which of the removed features is this campaign depending on? Manual CPC and bid adjustments are the common dependencies we find, and they tend to hide in old remarketing campaigns nobody has opened since 2024.
Who should migrate early
Simple setups should move first and take the free inventory. If the campaign is a remarketing list or a contextual play running Target CPA with a clean exclusion list, the port is low-risk and the destination is better than the origin: Demand Gen adds lookalike segments, carousel ads, target CPC bidding, campaign total budgets, Discover and Maps inventory, channel-level reporting, and generative image tools 1. For accounts that have wanted Discover placements without committing to a separate campaign type, this is the door.
The exception is rich media. HTML5 uploaded ads and third-party-served ads are planned for late 2026, not available now 1. If your Display program leans on custom HTML5 units or third-party ad serving, you wait, and the published 2027 runway means you can afford to.
Who should rebuild instead
Accounts built on manual bidding should not port; they should design again from the new primitives. A campaign holding a hand-set CPC ceiling on a remarketing list, tuned with device and geo bid adjustments, is not the same campaign once those controls vanish; migrating it just launches an untuned smart-bidding campaign wearing an old name. The honest move is a fresh Demand Gen build: pick the bid strategy from what exists (Target CPA or target CPC), set exclusions deliberately, and let the 42-day history stay behind with the removed original where it can still be read.
The same goes for anything structural: portfolio bidding strategies, shared budgets, and combined audiences are account-level architecture, and when the architecture is unsupported, porting the campaign strands it. Advertisers using lead form or click-to-call assets on Display lose those outright and need the conversion path rebuilt on a landing page instead, which is work we scope deliberately rather than patch mid-migration. Clients ask whether platform migrations like this are covered inside a management engagement; that answer lives in our FAQ.
Sequence matters whichever route you take. Move one campaign, let the learning period and the ad re-reviews settle, read a full week of stable delivery, then move the next tier. A phased migration costs a few weeks of calendar. Migrating an account in one afternoon stacks every learning period, every re-approval queue, and the budget-reset quirk onto the same day, and then asks you to diagnose which of them broke performance.
The same consolidation, one campaign type at a time
With Display campaigns retiring into Demand Gen, Google is running the identical play it has run all year: fewer campaign types, more automation inside each, and the manual controls traded for inventory and creative formats. It is the through-line we traced in our Google Marketing Live 2026 recap, and it is why the right response is not resentment about lost toggles. The toggles are not coming back. The work is deciding, per campaign, whether the automated replacement is good enough, and rebuilding the ones where it is not.
That evaluation is account work, not sentiment. In most Display programs we open, the automated replacement clears the bar on remarketing and fails it on tightly managed prospecting, so the port-versus-rebuild call gets made campaign by campaign rather than once for the account.
Google will migrate the stragglers without asking. Between now and then there is a window where the sequencing, the budget timing, and the rebuild-versus-port call all still belong to you. Accounts that spend the window auditing will go through the learning period once, on purpose. Accounts that wait will go through it twice, by surprise.