Google's Smart Bidding strategies — Maximise Conversions, Target CPA, Target ROAS, Maximise Conversion Value — are objectively better than manual bidding once they have data. The challenge is using them correctly. Most accounts that switch to Smart Bidding too early or with the wrong settings see CPA jumps and conversion drops. This guide covers when to use each strategy, how to set them up, and how to avoid the failure patterns I've seen on dozens of client accounts.
Smart Bidding is Google Ads' suite of machine-learning-driven bidding strategies. Used right, it dramatically improves performance. Used wrong, it dramatically degrades it. The difference is whether you understand what each strategy actually does and when to use it.
This guide is built around real campaigns I've managed across the strategies. The goal is practical decision-making, not theoretical perfection.
What Smart Bidding actually does
Smart Bidding adjusts your bid for every individual auction based on real-time signals: device, location, time of day, day of week, browser, language, audience signals, and dozens more. Manual bidding gives Google one number; Smart Bidding gives Google a goal and lets it pick the right number for each auction.
Google's machine learning has access to data manual bidders never see — historic conversion patterns at the auction level, cross-device behaviour, predicted user lifetime value. When Smart Bidding has enough data to work with, it consistently outperforms manual.
The catch: it needs the data. New campaigns don't have it.
The Smart Bidding strategies
Maximise Conversions
What it does: Spends your full daily budget while maximising the number of conversions.
When to use: Early in a campaign's life, before you have enough data for Target CPA. New product launches. Lead-gen campaigns where every conversion has similar value.
Pitfalls: Doesn't tune for conversion value. If your conversions vary widely (some are ₹100 leads, some are ₹1 lakh leads), Maximise Conversions doesn't know the difference.
Setup:
- Daily budget set realistically (not too low, Google's algorithm thrashes when budget-constrained)
- Conversion tracking working correctly
- Minimum 30 conversions/month is ideal
Target CPA
What it does: Spends to bring you conversions at your specified target cost per acquisition.
When to use: After 30+ conversions on Maximise Conversions, when you know your target CPA, when conversions have similar value.
Pitfalls: Setting target too low (Google reduces volume rather than achieving impossibly low CPA). Changing target too often (resets learning).
Setup:
- Set target at your achieved CPA from Maximise Conversions, not your aspirational target
- Run for 14 days minimum before adjusting
- Adjust by 10-15% at a time, not 50%+
Maximise Conversion Value
What it does: Maximises total conversion value (revenue) within your budget.
When to use: Ecommerce. Lead-gen where lead value varies. Subscription businesses where customer LTV varies.
Pitfalls: Requires accurate conversion value reporting. Sending wrong values (e.g., averaging instead of actual order values) corrupts the optimisation.
Setup:
- Conversion tracking includes accurate value (not flat ₹0)
- For ecommerce: send transaction value via Enhanced Ecommerce
- For lead-gen: import value from CRM via offline conversions
Target ROAS (Return on Ad Spend)
What it does: Spends to achieve your target ROAS (e.g., 400% means ₹4 of revenue per ₹1 spent).
When to use: Mature ecommerce. Sufficient conversion volume (50+/month). Predictable conversion value patterns.
Pitfalls: Too aggressive a target reduces volume. Too modest a target leaves money on the table.
Setup:
- Achievable target based on historical data
- 14-day learning period
- Patient adjustments
Enhanced CPC (eCPC)
What it does: Modest automation layered over manual bids. Adjusts your manual bid up or down based on conversion likelihood.
When to use: Stepping stone from Manual CPC toward fully automated bidding. Useful when you want some manual control with some ML help.
Pitfalls: Less impactful than fully automated strategies. Often skipped now in favour of Maximise Conversions.
Manual CPC
What it does: You set the bid; Google charges that amount per click (subject to standard auction adjustments).
When to use: New campaigns with no conversion data. Tightly controlled budget situations. Specific testing scenarios.
Pitfalls: Slow to react to opportunity. Doesn't take advantage of Google's auction-level data.
A practical bidding roadmap
For a brand-new account with no conversion data:
Phase 1 (Days 1-30): Manual CPC or Maximise Clicks
Start manual. Set max CPC bids based on a hypothesis: target CPA × expected conversion rate. Generate clicks, validate the funnel, gather initial conversion data.
Phase 2 (Days 30-60): Maximise Conversions
Once you have 15-30 conversions in a 30-day window, switch to Maximise Conversions. Let it run for 14 days minimum. Monitor CPA, it should stabilise around what manual bidding achieved.
Phase 3 (Days 60-90): Target CPA
When Maximise Conversions has been stable for 30 days, switch to Target CPA. Set target at your achieved CPA, not lower. Run 14 days. Then incrementally lower target by 10% at a time if profitable.
Phase 4 (Days 90+): Target ROAS or Maximise Conversion Value
For ecommerce or value-variable conversions. Requires accurate value tracking. Patient optimisation.
What actually breaks Smart Bidding
A few patterns I've seen wreck campaigns:
1. Inadequate conversion tracking
The single biggest failure mode. Smart Bidding optimises toward whatever you call a conversion. If you're tracking page views, clicks, or low-value events, the algorithm chases those. Burning budget on wrong outcomes.
Verify conversions in GA4 and Google Ads. Use Google Tag Manager for proper deployment. Read the GA4 setup guide on this blog if conversion tracking is unclear.
2. Insufficient conversion volume
Smart Bidding needs about 30 conversions/month minimum for Target CPA to work well. Below that, the algorithm doesn't have enough signal to optimise. Stick with Maximise Conversions until volume scales.
3. Volatile target changes
Every change to your target CPA or ROAS resets the learning phase. Algorithm needs 7-14 days to stabilise after a change. If you change target weekly, the campaign never learns.
4. Insufficient budget
If daily budget is too low, Google can't fully execute the strategy. Symptoms: low impression share, high impression share lost to rank, inconsistent delivery. Either increase budget or accept reduced volume.
5. Poor account structure carried over
Smart Bidding doesn't fix bad ad-group structure or weak ad copy. If your structure has 100 keywords in one ad group, Smart Bidding amplifies that confusion rather than fixing it. Restructure first; then automate.
6. Seasonal disruption without seasonality adjustments
For seasonal businesses, set seasonality adjustments so Google knows your conversion rate is temporarily different. Without these, Smart Bidding misinterprets seasonal spikes as new normal.
Combining Smart Bidding with audiences
Smart Bidding considers audience signals automatically — if you've added audiences as observation, the algorithm uses them.
Layer audience signals:
- Customer match as exclusion (don't pay for current customers)
- In-market audiences as observation (algorithm adjusts up for these)
- Affinity audiences as observation
- Remarketing lists as separate campaigns with different bidding
Don't restrict targeting to specific audiences in Smart Bidding campaigns — let the algorithm decide who to show to.
Smart Bidding for Performance Max
Performance Max campaigns only run on Smart Bidding strategies (Maximise Conversions or Maximise Conversion Value). The same setup principles apply:
- Conversion tracking must be perfect
- Budget needs to be realistic
- Conversion values must be accurate (for value-based strategies)
- Learning period needs respect (14 days minimum)
Performance Max campaigns are particularly punishing of bad conversion tracking because they spread spend across Search, Display, YouTube, Discover, Gmail, and Maps simultaneously, bad signals in one channel pollute optimisation across all.
Common Smart Bidding myths
"Smart Bidding wastes budget"
Only when conversion tracking is bad or volume is too low. With proper setup, Smart Bidding consistently outperforms manual bidding for accounts with 30+ conversions/month.
"Smart Bidding can't be controlled"
You can layer bid modifiers at the campaign level (location, device, ad schedule). You can set target CPA/ROAS. You can exclude audiences. You have less granular control than Manual CPC but more total impact.
"Smart Bidding favours big advertisers"
Same algorithm available to all advertisers. The advantage scales with conversion volume; bigger accounts have more data, but small accounts still benefit.
"I should test Manual vs Smart Bidding"
Generally not worth testing once you have 30+ conversions/month. Smart Bidding wins consistently in mature accounts. The test was worth running in 2018-2020 when Smart Bidding was new and unreliable; not in 2025.
Reporting and monitoring
Once a campaign is on Smart Bidding, the metrics that matter shift:
Watch closely:
- CPA / ROAS vs target
- Impression share lost to budget
- Search lost impression share due to rank
- Conversion volume week over week
Watch loosely:
- CPC (Smart Bidding will pay variable CPCs across auctions; an average CPC view is misleading)
- Quality Score (still important but Smart Bidding optimises around it implicitly)
Ignore:
- Click-through rate (less meaningful when Smart Bidding chooses which auctions to show in)
- Average position (deprecated metric anyway)
Tools that help
A short list of tools that integrate with Smart Bidding workflows:
- Google Ads Editor. Bulk adjustments to conversion settings
- Looker Studio — Smart Bidding performance dashboards
- Optmyzr ($208+/month) — third-party tool with Smart Bidding-specific features
- Skai — enterprise PPC management with bidding intelligence
For ad-hoc work on smaller accounts, the link tracker on this site builds clean UTM URLs that flow proper attribution back to Smart Bidding.
Smart Bidding is the single biggest performance lever in modern Google Ads. The accounts that resist it are the ones leaving 20-50% of their potential ROAS on the table. The accounts that adopt it correctly. Proper tracking, sufficient volume, realistic targets, patience during learning, consistently outperform manual benchmarks. Make the switch when you're ready, then trust the algorithm for at least 30 days before second-guessing.
A 60-day Smart Bidding implementation plan
If your account is ready to switch:
- Week 1: Validate conversion tracking. Confirm conversion definitions match business goals.
- Week 2: Switch to Maximise Conversions. Don't change anything else.
- Weeks 3-4: Monitor daily. Don't make changes unless something is dramatically wrong.
- Weeks 5-8: Once stable, consider Target CPA. Set target at achieved CPA, not aspirational.
- Weeks 9-12: Optimise — landing pages, ad copy, audience signals. Let Smart Bidding do its job.
Resist the urge to constantly tweak. Smart Bidding rewards patience.
Final thoughts
Smart Bidding works when you let it. The pattern across every successful Smart Bidding implementation I've seen is the same — proper conversion tracking, enough conversion data, realistic targets, patience during learning. Skip any of these and Smart Bidding fails. Get all four right and the same campaign that broke even on Manual CPC starts producing 3-5x ROAS within 60 days.
Need help applying this to your own site? I'm Shani Maurya — a freelance web developer and digital marketer based in Delhi. If you'd like a hands-on audit or full implementation, get in touch — I usually reply within a few hours.
