AIuthority

Google Releases February Demand Gen Best Practices Update

By Charles Ryder

I’ve been tracking Demand Gen campaigns since the early access period in 2023. The trajectory hasn’t changed: more automation, more visual-first inventory, and a growing emphasis on the combination that actually drives results—feeds, creative, and clean data. With the February 24, 2026 Demand Gen Drop, Google is making that point more official by packaging what it believes is delivering the biggest wins for advertisers right now.

This isn’t a feature dump. It’s a best-practices refresh built on Google’s internal performance analysis (April 2024 through December 2025). Google says advertisers who adopted three or more Demand Gen best practices saw 40% more conversions. It’s a bold headline, but also a useful checklist for anyone deciding how seriously Demand Gen belongs in a 2026 media plan.

Google's February Demand Gen update infographic, showing best practices for AI-driven campaigns, automation, and visual inventory. Why Demand Gen keeps getting more attention

Demand Gen is Google’s AI-driven, visual demand-creation format across YouTube (including Shorts), Discover, Gmail, and the Display Network. It’s also positioned as the successor to Discovery Ads and increasingly the destination for upgrades from Video Action Campaigns. Put simply: this is where Google wants upper- and mid-funnel growth budgets to land.

Google’s “Drops” series (introduced in late 2025) has been its way of nudging marketers toward a more consistent operating model. February’s Drop reinforces that push by organizing recommendations into four pillars: Audiences, Bid/Budget, Creative, and Data Strength.

The February update: Google’s ABCD pillars (and what I’m taking from them)

Google frames the guidance as an “ABCD” approach. The framing is simple; the implication is not. In AI-heavy campaign types, performance usually comes down to inputs and signal quality—what you feed the system and how reliably it can measure outcomes.

A: Audiences — scale with signals, not rigid segments

Google’s audience guidance keeps moving in the same direction: fewer hard boundaries, more room for the system to learn.

  • Optimized targeting is a central recommendation, with Google citing 20% higher conversions at the same cost when it’s enabled.
  • Lookalike segments continue to function less like strict “match this profile” targeting and more like an AI signal—typically trading manual control for reach.
  • New Customer Acquisition (NCA) goals get another push. Google reports an 11.5% improvement in new-to-returning customer ratio and a 3% decrease in acquisition cost.

My read: teams used to tight audience guardrails will have to recalibrate. The upside is scale. The downside is waste if your conversion tracking and creative aren’t strong enough to steer the system.

B: Bid and budget — stop starving the model

Google is blunt here: automation needs enough volume to learn.

  • Use tCPA or tROAS to point automation toward efficiency goals.
  • Fund campaigns at a level that supports learning volume—Google repeats the common rule of thumb: budget at least 10x your CPA.
  • Use Performance Planner to model scenarios rather than guessing.

The unspoken qualifier matters: if you can’t support volume—Google often references needing 50+ conversions in 30 days—you shouldn’t expect the same uplift others report. Demand Gen can work on tests, but it’s not designed for tiny budgets that never graduate.

C: Creative — “Excellent” Ad Strength isn’t optional anymore

This is where many teams underinvest, and Google is clearly trying to force the issue.

  • Aim for “Excellent” Ad Strength, which usually means supplying enough high-quality assets across formats and aspect ratios.
  • Run multiple asset types (images and video), not just one.
  • For commerce, product feeds via Merchant Center keep showing up as a compounding performance lever.
  • Google also reiterates its ABCD-style video guidance (Audience, Branding, Connection, Direction), a reminder that assets need to win attention—especially in Shorts and Discover.

If your Demand Gen approach is “repurpose a few banners and hope,” Google is telling you that era is over.

D: Data strength — the quiet make-or-break factor

It’s the least glamorous pillar and often the one that decides whether automation helps or hurts.

  • Confirm sitewide tagging is correct and resilient (Google mentions options like tag gateways).
  • Use Data Manager to import offline or CRM conversions when it’s relevant.
  • Strengthen integrations with Google Analytics and keep conversion definitions clean and consistent.

In AI-led campaigns, weak data doesn’t just muddy reporting—it breaks the feedback loop. When that loop degrades, automation doesn’t “sort of work.” It can fall apart.

Google Demand Gen ABCD pillars: Audiences, Bid/Budget, Creative, Data Strength, illustrating key strategies for campaign optimization. Cropp’s results: a case study that fits the playbook

Google spotlights a Demand Gen case study from Cropp (LPP Group), a Gen-Z-focused apparel brand. The setup mirrors Google’s recommended mix: tROAS bidding, Merchant Center feeds, strong creative coverage, and audience expansion through lookalikes and optimized targeting.

The reported outcomes are the kind of stats performance teams pay attention to:

  • 50% ROAS uplift
  • 4.5% conversion lift (reported for women shoppers)
  • 8% lift in search interest, measured through a Conversion Lift study and view-through optimization

This example lands because it reflects what Google’s systems keep rewarding: feed integration, enough creative volume to iterate, and sufficient conversion data for the model to stabilize.

What this means for marketers heading into 2026

February’s update reinforces the bigger shift already in motion: Google Ads continues moving away from “you target, you control” and toward “you provide strong inputs, the system finds outcomes.” Marketers aren’t less important—the job just looks different.

If you want the practical version, it’s this:

  • Treat Demand Gen as a creative and measurement discipline as much as a media discipline.
  • Consolidate structure where possible (fewer campaigns with stronger signals).
  • Build budgets around learning requirements, not perpetual micro-tests.
  • Expect audience controls (especially lookalikes) to behave more like guidance than guardrails.

Conclusion

Google’s February Demand Gen best-practices update draws a clear line between advertisers who dabble and advertisers who win. The winners commit to the full stack: scalable audience signals, budgets that allow learning, asset-rich creative built for modern placements, and measurement that’s actually trustworthy. If you want to keep pace with the changes and track what matters as Demand Gen evolves, a dedicated resource like AIuthority can help you stay aligned with the latest guidance and real-world performance patterns.