You might remember Google Discovery Ads — the visually rich ads that appeared in the Discover feed, the YouTube Home feed and Gmail. In 2023, Google folded Discovery Ads into a new campaign type called Demand Gen. If you were running Discovery campaigns, they were migrated automatically, and there's no going back.

So what are Demand Gen ads, how do they differ from what Discovery was, and — the question that matters — should a UK business actually bother with them? Here's the honest verdict.

What Discovery Ads were

Discovery Ads were Google's answer to social advertising. Instead of waiting for someone to search, they pushed visually engaging single-image and carousel ads into Google's feed-based surfaces:

  • The Google Discover feed on Android and the Google app (that endless scroll of articles and content).
  • The YouTube Home feed.
  • Gmail promotions and social tabs.

They were interruption marketing in the Google ecosystem: you weren't capturing existing demand, you were creating it. Targeting was audience-first — custom segments, in-market audiences, your own remarketing lists — rather than keyword-first. The format looked a lot like a Facebook or Instagram ad: big image, headline, description, brand name.

For businesses with visually appealing products — home furnishings, fashion, travel, food — Discovery worked well as an awareness and prospecting channel. For a plumber or a solicitor, it was mostly a curiosity.

What changed: enter Demand Gen

Demand Gen kept the same inventory (Discover, YouTube Home, Gmail) but rebuilt the campaign machinery around it. The important changes:

  • Video is now first-class. Demand Gen supports video assets alongside images, and Google heavily favours campaigns that include both. Short vertical video, in particular, performs well because the inventory skews mobile.
  • Lookalike segments. You can now build lookalike audiences from your converters, high-value customers or YouTube subscribers. This is the single biggest upgrade for UK advertisers with decent first-party data.
  • Conversion bidding that actually works. Demand Gen supports Maximise Conversions and tCPA bidding properly tuned for feed inventory. Discovery's optimisation was always a bit blunt; Demand Gen is noticeably better at finding people who convert.
  • More placement control. You can now see and exclude specific placements, which matters because Gmail inventory and Discover inventory behave very differently.

Existing Discovery campaigns were upgraded to Demand Gen automatically, and Google retired the old Discovery campaign type. If you're setting up from scratch today, Demand Gen is the only option in this inventory.

How Demand Gen differs from Search and Display

It's worth being clear about where Demand Gen sits, because misallocating budget between Google campaign types is a classic UK SME mistake:

  • Search captures existing demand. Someone types "emergency plumber Manchester" and you appear. High intent, high cost-per-click, obvious attribution.
  • Display puts banner ads across the web. Cheap, vast, and mostly ignored — useful for remarketing and cheap reach, weak for prospecting cold audiences.
  • Demand Gen sits between them: it's prospecting to people who aren't searching right now but match your audience, in premium Google-owned inventory. It borrows the audience logic of social ads and the bidding machinery of Google Ads.

The honest comparison is with Meta Ads (Facebook/Instagram), not with Search. Demand Gen and Meta fight over the same budget line: visual, interruption-based prospecting. If your Meta prospecting is already working well, Demand Gen is worth testing as a second channel. If neither is working, the problem is probably your creative or your offer, not the platform.

When Demand Gen works well

Based on how the inventory behaves and what Google's own data shows, Demand Gen tends to perform for:

  • Visually driven products. Fashion, beauty, homeware, furniture, travel, food and drink. If your product photographs badly, feed inventory will punish you — the whole surface is visual.
  • Considered purchases with a research phase. Demand Gen is good at planting the seed. Someone sees your kitchen ad in Discover in March, Googles you in June, converts in July. Attribution will under-credit Demand Gen every time, so you need to accept assisted conversions as real.
  • Businesses with strong creative production. The advertisers winning on Demand Gen refresh creative monthly and run multiple image and video variants. If "creative production" means one product photo from 2021, lower your expectations.
  • Retargeting layered on prospecting. Running Demand Gen to warm audiences (site visitors, cart abandoners, past customers) is one of its strongest use cases — the inventory is premium and the audience is yours.

When it doesn't (the honest part)

Demand Gen is not a magic bullet, and Google's sales materials oversell it. Skip it or deprioritise it when:

  • You're a lead-gen business with no visual hook. Solicitors, accountants, B2B software, industrial suppliers — feed inventory will burn budget showing your ad to people with zero intent. Search and LinkedIn are better homes for that money.
  • Your budget is under about £1,000/month for the test. Demand Gen needs volume to optimise. Spreading £300/month across it will teach you nothing except that £300 disappears quickly. Consolidate before you diversify.
  • You can't track assisted conversions. If your reporting only credits last click, Demand Gen will look like a failure even when it's driving the research phase. You need proper attribution (data-driven attribution in Google Ads, plus GA4 path analysis) before you judge it.
  • Your creative is weak. This is the uncomfortable truth. Demand Gen is a creative channel. Bad creative on Search can still convert because intent is high; bad creative on Demand Gen converts nobody, because nobody asked to see it.

Setting up a sensible UK test

If Demand Gen fits your business, run it as a disciplined test, not an open-ended spend:

  1. Start with lookalikes and remarketing. Build a lookalike of your best customers (top 20% by value if you have the data) and layer site visitors. These outperform broad prospecting in the first 60 days.
  2. Use at least 3 image assets and 1 video. Google's guidance says more assets give the machine more to optimise with, and in practice campaigns with video materially outperform image-only ones.
  3. Set a test budget you can lose. £1,000–£2,000 over 4–6 weeks is a fair test for most UK SMEs. Judge it on assisted conversions and view-through impact, not last-click ROAS alone.
  4. Exclude underperforming placements. After two weeks, check placement reports. Gmail inventory often delivers cheap clicks and few conversions; don't be afraid to cut it.
  5. Refresh creative monthly. Feed audiences fatigue fast. Plan a creative refresh cadence before you launch, or performance will decay in month two and you'll blame the platform.

The verdict: should you bother?

For UK ecommerce and consumer brands with visual products, decent creative resources and at least a few thousand pounds a month in ad spend: yes, test it. Demand Gen is a legitimate second prospecting channel alongside Meta, with Google's bidding machinery behind it.

For everyone else — B2B, lead gen, small budgets, weak creative — it's a distraction from the channels that will actually move revenue: Search, well-run Meta remarketing, and a website that converts. Come back to Demand Gen when you've got the fundamentals right.

The pattern is always the same: new ad formats reward advertisers with strong fundamentals (tracking, creative, offer) and punish everyone else. Demand Gen is no exception.

If you'd like an honest read on whether Demand Gen fits your account — or whether your budget is better spent elsewhere — our PPC advertising team will tell you straight, including when the answer is "not yet". For a related read on how social platforms influence buying behaviour, see our article on how social media affects SEO and brand discovery. Ready to talk? Contact us for a no-obligation review of your current campaigns.

FAQs

What happened to Google Discovery Ads?

Google retired the Discovery Ads campaign type and migrated all Discovery campaigns to Demand Gen. Discovery as a separate product no longer exists. Demand Gen covers the same inventory (Discover feed, YouTube Home, Gmail) with added video support, lookalike audiences and improved bidding.

Are Demand Gen ads better than Facebook ads?

Neither is universally better — they're different tools for the same job (visual prospecting). Meta usually has broader reach and stronger audience data for consumer prospecting; Demand Gen has Google's bidding and first-party search-intent signals. Many UK advertisers run both and compare assisted conversions. Test both before declaring a winner.

How much should I spend testing Demand Gen in the UK?

A fair test is £1,000–£2,000 over 4–6 weeks. Less than that and the algorithm can't optimise properly; you'll get noise, not a verdict. Make sure your conversion tracking is solid before you start, or you won't be able to judge the results.

Can Demand Gen work for B2B?

Rarely as a primary channel. Feed inventory reaches consumers in leisure mode, which suits B2C products. B2B can sometimes work for very visual products or broad brand awareness, but Search and LinkedIn are usually better uses of a B2B budget in the UK.

Do I need video creative for Demand Gen?

Technically no — image-only campaigns are allowed. In practice, campaigns with video significantly outperform image-only ones, because the inventory skews heavily to mobile and short-form video. If you can't produce even simple video creative, temper your expectations or fix that gap first.

Digital Marketing Visuosofts Get a Quote