Display Campaigns Are Moving to Demand Gen: How to Migrate Without Losing Performance in 2026
Standalone Display campaigns are going away. Google is moving the entire Google Display Network inventory inside Demand Gen: the in-account migration tool went live in mid-2026, creating new Display campaigns is being switched off later in the year, and whatever is left gets migrated automatically through 2027. The Display to Demand Gen migration is not a rename of a campaign type — it is a change in the set of levers you get. Some settings carry over one-to-one, some quietly disappear, and budget pacing plus reporting start behaving differently.
This is the working version for advertisers and media buyers: the exact timeline, a complete list of what transfers and what breaks, a step-by-step plan for migrating on your own terms, and the metrics to watch for the first two weeks. Written for accounts where Display actually drives conversions, not for the ones where it just burns leftover budget.
The single most useful takeaway: do not wait for the automatic migration. A campaign moved with Google’s tool, manually, during a quiet week keeps its learning history and rides out the transition in one or two days. A campaign auto-migrated mid-season loses its pacing rhythm and needs its audiences rebuilt by hand.
What the Display to Demand Gen migration actually changes
Demand Gen was built for demand that has not yet turned into a search query: YouTube, Shorts, Discover, Gmail. In 2026 Google added Google Display Network inventory — roughly 2 million sites, apps and video placements — to the same campaign type, and used that as the basis for retiring Display as a standalone product.
Google’s logic is consolidation: one campaign type, one asset set, one model that distributes impressions across channels. The advertiser’s logic is different: Display was a predictable channel with manual levers, while Demand Gen is an algorithm-led campaign where several levers have been replaced by signals. That is the real risk of the transition — not a drop in volume, but a loss of control.
One detail that gets missed: in Demand Gen you can serve on GDN only. Channel controls let you reproduce your previous configuration first and expand into YouTube and Discover later, as a separate test. Google reports that advertisers who add GDN to existing Demand Gen campaigns see roughly a 9.5% average ROI lift, and cites GoFood at −24% CPA and +19% conversions. Treat both as vendor benchmarks, not as forecasts for your account.
The timeline worth planning around
| Phase | When | What it means in practice |
|---|---|---|
| Migration tool in your account | from June 2026 | You trigger the move manually; history (about 42 days) and budget carry over |
| No new Display campaigns | second half of 2026 | New launches happen in Demand Gen only; “old way” tests are no longer possible |
| Automatic migration of the rest | 2026–2027, in waves | Google moves campaigns for you, on its schedule; settings with no equivalent are dropped |
The practical reading: the window in which the migration is under your control is open now and it is finite. The later you start, the fewer options you have to do it cleanly.
What transfers intact
The migration tool carries over more than sceptics expect. Per Google’s documentation, you keep:
- Campaign goal and performance history — around 42 days of data, so Smart Bidding continues rather than restarting.
- Budget — the daily amount transfers, but the daily counter resets: spend before the migration does not count towards that day.
- Audience targeting — remarketing lists, custom segments, contextual signals.
- Bid strategies — Target CPA, Maximize Conversions, Target ROAS, Maximize Clicks.
- Creative assets — responsive display ads, uploaded images, carousel ads.
The migrated campaign is renamed “[original name] #2”. Small thing, big consequence: it breaks automated rules, scripts and any dashboard that filters campaigns by name. Fix that first, or your media buyer reporting stack will silently drop half the account.
What you lose or what behaves differently
This is the part that matters. Below are the levers that either do not exist in Demand Gen or work on different rules.
| In Display | In Demand Gen | What to do |
|---|---|---|
| Manual CPC | Not available | Move the campaign to Target CPA or Maximize Conversions well before migrating and let it gather data |
| Viewable impression (vCPM) bidding | Not available | Move brand-awareness goals into dedicated video or reach campaigns |
| Brand Lift and Search Lift | Not supported on migrated campaigns | Measure impact with geo experiments instead |
| Combined audiences | Replaced by the audience builder | Rebuild the combinations manually before migrating and compare reach |
| Placement and app exclusions | Available, but inside a different delivery logic | Re-check your lists after the move; watch app inventory for junk traffic |
| Hard channel separation | The algorithm splits channels | Start GDN-only, expand as a separate, measured step |
A note on exclusions. In Display, junk in-app inventory was handled with long exclusion lists at account level. In Demand Gen the algorithm decides how impressions are split across channels, so the same exclusion list produces a different outcome. Review the placement report every two or three days for the first fortnight and keep extending the lists. It also helps to refresh how brand lists and brand exclusions work, because the Demand Gen mechanics are closer to those than to old placement exclusions.
Three reasons not to wait for the automatic move
You control the date. Auto-migration arrives in waves and does not ask whether you are mid-promotion. A daily budget counter reset plus one or two days of re-learning during peak season is a real, quantifiable cost. If you have promo peaks like Black Friday ahead, finish the migration before the promo window, not inside it.
You keep the history. The migration tool carries about 42 days of data. Rebuilding the campaign from scratch in Demand Gen does not: you get the full Smart Bidding learning period, with all the CPA volatility that comes with it.
You get an audit first. Auto-migration silently drops anything without an equivalent. Migrating yourself means you see in advance which combined audiences will fall apart and which exclusions need duplicating. It is also a natural moment to run the account against a Google Ads account audit checklist.
The step-by-step migration plan
- Freeze a baseline. Export 30-day and 90-day figures: spend, conversions, CPA/ROAS, CTR, conversion rate by placement and by device. Without this export you will not be able to prove later whether things improved.
- Get off manual bidding. If the campaign runs on manual CPC, switch it to Target CPA or Maximize Conversions at least two or three weeks before the migration so the strategy has data.
- Deconstruct combined audiences. For every combination, write down which segments built it and what volume it delivered. Rebuild in the audience builder afterwards and compare reach. A 20–30% gap is normal; a multiple is a signal to rebuild differently.
- Duplicate exclusions. Placement, app, topic and brand lists — at account level, not just on the campaign, so they survive the move.
- Complete the asset set. Demand Gen underperforms on banners alone: you need logos in the required ratios, vertical and square images, and at least one video. If you have no video, build a simple one from statics — still better than an empty slot. Practical approaches are in the piece on ad creatives for feeds.
- Narrow the channels. In the new campaign’s settings, keep GDN only if the goal is to reproduce your previous logic. Expanding into YouTube and Discover is a separate test later.
- Run the migration tool. Early in the week, not at month-end: you want working days to react before nobody is watching the account.
- Fix names and automation. The “#2” suffix breaks rules, scripts, labels and dashboard filters. Rename to your convention and update conditions in your automated rules.
- Limit big edits. For the first 7–10 days, keep budget and target changes within ±15% per step, otherwise you will blend the migration effect with your own changes.
- Set checkpoints. Day 3, day 7, day 14: compare against the frozen baseline on CPA, channel spend share and traffic quality.
The first two weeks: what to watch
- Channel spend split. Even on GDN-only, check every couple of days where budget goes. Read it the way you read Performance Max channel-level reporting: not by blended CPA, but by CPA inside each channel.
- Traffic quality. Share of sub-three-second sessions, share of in-app impressions, anomalous placements. This is your earliest signal that new exclusions are needed.
- Overlap with other campaigns. After the move, Demand Gen overlaps more eagerly with remarketing and PMax. Check audience overlap — the mechanics mirror the piece on PMax and Search cannibalization.
- Conversion stability. A 1.5–2x CPA swing in the first days is expected during a short re-learn. If it has not returned to the baseline corridor after 14 days, the problem is configuration, not learning.
- Change history. Log every edit — during a performance investigation it is the only reliable source; the method is in the article on change history and diagnosing performance drops.
Common mistakes
Migrating everything in one day. With 15 Display campaigns, a bulk move destroys your ability to tell which one dropped and why. Sensible pace: two or three campaigns at a time, a few days apart.
Opening all channels immediately. “Let the algorithm decide” is tempting, but then you cannot compare the new state to the old one. Reproduce Display first, expand second — ideally through Experiment Center experiments rather than by flipping a live campaign.
Judging success on blended CPA. The channel mix changes, and with it the conversion mix: cheap video conversions can mask GDN conversions getting more expensive. Always look at the splits.
Forgetting the daily budget reset. On migration day the campaign technically starts its day over. Migrate at 6pm after the daily limit is already spent and actual 24-hour spend will overshoot the plan. A small thing that produces an unexplained overspend in the monthly report — especially if you run to the budget pacing rules with a hard monthly cap.
Who gains and who loses
Likely to gain: accounts where Display carried remarketing and dynamic retargeting on a large base, and teams that can produce video. Demand Gen gives them lookalike segments, wider formats and channel-level reporting. If your dynamic remarketing already runs off a feed, the transition should land neutral or positive.
Likely to lose: anyone who ran Display on manual CPC plus tight placement lists, squeezing cheap traffic out of a narrow inventory slice. That approach does not reproduce in Demand Gen. You either accept the new logic and learn to steer with audience signals — covered in the piece on optimized targeting and audience signals — or reallocate that budget elsewhere.
How this fits the rest of 2026
The Display migration is one move in a pattern: Google is removing standalone campaign types and manual levers, replacing them with a single AI-led campaign per demand channel. The same pattern includes upgrading Dynamic Search Ads into AI Max, turning on local inventory ads by default in Shopping campaigns, and shipping text guidelines for AI-generated assets as partial compensation for lost copy control.
The planning conclusion: in 2026 the winners are not the advertisers clinging longest to manual settings, but the ones who translate their logic fastest into the signals the algorithm accepts — feed, audiences, assets, constraints — and who build measurement honest enough to see the result. If that stage needs a stable ad account with history plus hands-on setup help, look at Google Ads agency account rental and the rest of the PPC Rebels services; campaign-type fundamentals live in the complete Demand Gen guide.
Settings map: where everything lives after the move
Half the post-migration panic is not about lost capability — it is about a familiar setting sitting somewhere else under a different name. Here is the practical map.
| In Display | Where to find it in Demand Gen | What to watch |
|---|---|---|
| Ad groups split by audience | Ad groups with audiences at group level | The structure survives, but the algorithm reallocates impressions between groups more aggressively |
| Remarketing lists | Audience builder → your data segments | Confirm the lists are active and growing: minimum size thresholds still apply |
| Similar audiences | Lookalike segments | You choose the similarity setting: narrow, balanced or broad |
| Topics and interests | Interest segments inside the audience builder | Contextual topics act as a signal, not a hard constraint |
| Responsive display ads | Demand Gen asset groups | Aspect-ratio requirements are stricter: you need 1:1, 4:5 and 1.91:1 |
| Frequency capping | Campaign-level frequency settings | Not available for every objective — check on the specific campaign |
| Placement report | Placement report plus channel breakdown | Read both: the channel explains where the placement came from |
Budget arithmetic: why monthly spend drifts
Three effects stack up and push week one away from plan.
The daily counter reset. On migration day the campaign restarts its day, so calendar-day spend can exceed the daily budget. Migrate several campaigns on the same day and the effect compounds.
A different pacing profile. Demand Gen allocates budget across several channels, so the hourly spend curve differs from Display. A campaign that used to finish its budget by evening may finish it by lunchtime.
Auction mispricing during re-learning. Bids are less precise in the first days, which produces both underspend and local CPC spikes.
Practical move: during migration week set the daily budget 10–15% below target and restore it after stabilisation. If you work to a hard monthly cap, recalculate the remaining runway using the logic in the article on budget pacing.
Remarketing and frequency: where the money usually leaks
Display was the remarketing workhorse for years, and that is exactly the use case that transitions least comfortably — not because of targeting, but because of frequency and overlap.
- Overlap with other campaigns. After the move, Demand Gen competes for the same user with PMax and search remarketing. Without audience-level exclusions you pay twice for one contact.
- Frequency. The inventory widened, so the same user can meet your ad on a site, in Discover and in Gmail. Set a frequency cap immediately rather than waiting for complaints.
- Excluding existing customers. The classic slip: the “already purchased” list stayed on the old campaign and never moved. Verify exclusion audiences campaign by campaign, not “in general”.
- Recency segmentation. A 7 / 30 / 90-day structure still works, but the algorithm prefers spending on broad segments. If recency matters to you, separate it by campaign rather than by ad group.
Pre-migration checklist
- Baseline metrics exported for 30 and 90 days, saved outside the interface.
- Campaign off manual CPC for at least two weeks.
- Combined audiences documented by composition and volume.
- Exclusion lists duplicated at account level.
- Assets include a logo in the required ratios, a vertical and a square image, and at least one video.
- Day chosen: early in the week, not month-end, not the eve of a promotion.
- A 14-day watch plan exists: who looks, at what, and at what deviation they intervene.
- “Already purchased” exclusion audiences in place and PMax overlap checked.
- Daily budget reduced 10–15% for migration week.
- Everyone who reads reports has been told campaign names will change.
If that list leaves you feeling the account is overdue a clear-out, it probably is. Migration is a good reason to rebuild the structure properly instead of dragging accumulated mess into a new campaign type; the reference points are in the piece on Google Ads account structure for scale.
FAQ: Display to Demand Gen migration
Can I opt out and stay on Display?
No. Creating new Display campaigns is switched off in the second half of 2026, and existing ones migrate automatically in waves through 2027. Your only real choice is whether you move at a time you pick or wait for the automatic wave.
Will I lose the campaign’s learning history?
With the official migration tool you keep roughly 42 days of history and typically see a one-to-two-day re-learn. Rebuilding the campaign manually from scratch carries no history at all.
What if my campaign runs on manual CPC?
Switch it to an automated strategy two or three weeks before you migrate, so the strategy collects data on familiar inventory and the transition does not coincide with the start of learning.
Can Demand Gen serve on GDN only?
Yes — channel controls let you restrict delivery to the Google Display Network. That is the recommended first step: reproduce the old configuration, then expand channels as a separate test.
Do my placement exclusion lists survive?
Exclusions exist in Demand Gen but operate inside a different delivery logic. Keep the lists at account level and re-check the placement report every two or three days for the first two weeks.
Why does the migrated campaign have “#2” in its name?
That is how the migration tool labels the new campaign. Rename it to your own convention and update every rule, script and report filter that depends on campaign names.
Is video really required in the asset set?
Formally no, practically yes. Demand Gen allocates impressions by format, and without video the campaign reaches a smaller slice of inventory. A simple video built from statics and text is enough to start.
How do I measure migration impact without Brand Lift?
Use geo experiments: run the new configuration in a subset of regions and compare against holdout regions. That is far more reliable than a before/after calendar comparison, where the effect blends with seasonality.
How long until things stabilise?
Plan for 7–14 days. A 1.5–2x CPA spike in the first days is normal; if performance has not returned to the previous corridor after two weeks, the cause is settings, assets or audiences rather than learning.
Should I migrate every campaign at once?
No. Two or three campaigns at a time, a few days apart, keeps the diagnosis clean — you can see which campaign dropped and what to fix inside it.
What if junk app traffic appears after migration?
Work the placement report, add app and category exclusions, and exclude mobile apps as an inventory type if needed. Always validate against conversion quality, not CTR alone.
How do I know the migration succeeded?
Three conditions at once: CPA back in the baseline corridor within 14 days, channel spend split matching the plan, and placement-level traffic quality no worse than before. If only the first holds, you probably just bought cheap conversions from a different channel.
