Google’s AI Max and Performance Max and Meta’s Advantage+ now control most of the levers that used to separate good media buyers from average ones: bids, audiences, placements and keyword matching. Every platform is pushing advertisers toward automation.
So what is left for a marketer to do? Quite a lot, but the work has moved.
1. Feed the algorithm better data
Automated bidding is only as smart as the conversion signal behind it. If you optimize for form fills, the AI finds cheap, low-quality leads. If you optimize for qualified leads or purchases, it finds buyers.
What to do:
- Set up server-side tracking and Conversions API (Meta) or enhanced conversions (Google).
- Import offline conversions, such as a closed sale or a verified lead, back into the platforms.
- Use value-based bidding. Tell the platform that a ₹10,000 order matters more than a ₹500 one.
For D2C and fintech brands especially, this one change often does more than any bid tweak.
2. Creative is the new targeting
With broad targeting now the default, the creative itself decides who sees your ad and who responds to it. Meta’s own system reads your visuals, hook and copy to find the right people.
What to do:
- Test 4-6 genuinely different concepts per month, not 10 versions of the same idea.
- Vary the angle: problem-led, social proof, offer-led, founder story, UGC.
- Make the first 3 seconds of every video count.
- Review creative performance weekly and kill fatigue early.
3. Structure campaigns simply
Over-segmented accounts starve the algorithm of data. Many accounts do better with fewer campaigns and more budget per campaign.
What to do:
- Consolidate campaigns that share the same goal.
- Give each ad set or asset group enough volume to exit the learning phase.
- Avoid editing budgets and targeting every day. Let changes settle for 5-7 days.
4. Control the guardrails
Automation will spend whatever you allow. Your job is to set the boundaries.
What to do:
- Add negative keywords and brand exclusions where the platform allows.
- Separate brand and non-brand search so branded conversions don’t hide weak prospecting.
- Use placement and audience exclusions, and check search term and placement reports regularly.
- Set clear CPA or ROAS targets, and make sure they are realistic.
5. Measure beyond last click
Platform dashboards tend to over-credit themselves. Cross-check with:
- GA4 and your CRM or order data
- Blended metrics, such as total revenue against total ad spend (MER)
- Incrementality tests, such as geo or holdout tests, for larger budgets
If the platform says ROAS is 5x but your bank balance disagrees, trust the bank balance.
6. Don’t ignore the landing page
AI can improve who clicks, but it can’t fix a slow, confusing page. A faster page, a clear offer, trust signals, and easy mobile checkout or WhatsApp contact (very effective for Indian audiences) lift conversion rate on every rupee of spend.
Quick checklist for this month
- Audit your conversion tracking for accuracy
- Switch to value- or quality-based conversion goals
- Launch at least 3 new creative angles
- Consolidate fragmented campaigns
- Review search terms, placements and exclusions
- Compare platform ROAS to real business revenue
The takeaway
Automation hasn’t made PPC experts obsolete. It has moved their value from button-pushing to strategy: data quality, creative, structure and measurement. Brands that get these right will outperform competitors who just “turn on the AI” and hope.