Google's official line: "Smart Bidding always wins." My experience across 200+ accounts: it depends entirely on data volume, conversion quality, and how you set it up. This article gives you the actual decision matrix.
The 5 Smart Bidding strategies (and what each is for)
- Maximize Clicks - chases traffic, ignores conversions. Almost never the right choice.
- Maximize Conversions - chases conversion volume regardless of cost. Good for ramp-up.
- Maximize Conversion Value - chases revenue. Default for ecommerce with proper value tracking.
- Target CPA (tCPA) - bids to hit a specific cost per acquisition. Default for lead gen.
- Target ROAS (tROAS) - bids to hit a specific return on ad spend. Default for ecommerce at scale.
Manual CPC - when it still wins
Manual CPC is not dead. It outperforms Smart Bidding in five specific scenarios:
- Thin data: <30 conversions/month means Smart Bidding is essentially guessing.
- New campaigns: Until you have 14 days of conversion data, Smart Bidding has nothing to learn from.
- Brand campaigns: Branded keywords convert at known rates. Manual is cheaper and gives you full control.
- High-value, low-volume verticals: Legal, M&A consulting, luxury - where 3 leads/month are still profitable.
- Delayed conversions: If your sales cycle is >30 days, Smart Bidding cannot tie clicks back to revenue fast enough.
The conversion volume thresholds you must respect
Google publishes vague guidance. Here are the real thresholds based on what works:
| Strategy | Min conversions/month | Best for |
|---|---|---|
| Manual CPC / Enhanced CPC | 0-30 | New, low-volume, brand |
| Maximize Conversions | 15+ | Ramp-up, no CPA target yet |
| Target CPA | 30+ | Steady-state lead gen |
| Target ROAS | 50+ (with values) | Ecommerce at scale |
| Maximize Conversion Value | 50+ (with values) | Ecommerce ramp-up |
Below these thresholds, Smart Bidding adds noise instead of intelligence.
The decision tree - pick your bidding strategy in 30 seconds
- Do you have <30 conversions/month? → Manual CPC or Maximize Conversions.
- Are you ecommerce with conversion values flowing? → tROAS (or Max Conv Value if ramping).
- Are you lead gen with 30+ conv/month? → tCPA at 110-120% of current CPA.
- Are you running a brand-only campaign? → Manual CPC at low max bids.
- Do you have wildly varying conversion values? → tROAS, never tCPA.
How to migrate from Manual to Smart Bidding (the safe way)
The biggest mistake: switching cold and panicking when CPA spikes for 7 days. Use this 4-step migration:
- Audit your conversion tracking first. Smart Bidding amplifies bad data. If your tracking is broken, fix it before switching. See ROI tracking guide.
- Set tCPA = 110-130% of current CPA. Not your target - your current actual. Going aggressive on day one kills volume.
- Hold the line for 14 days. Learning phase volatility is normal. Do not adjust targets in the first 2 weeks.
- Tighten in 5-10% increments. Once stable, drop tCPA by 5-10% every 2 weeks until volume starts dropping.
The tCPA vs tROAS decision
This trips up most advertisers. The rule:
- Use tCPA when: All conversions have similar value (lead gen, demos, signups).
- Use tROAS when: Conversions have variable values (ecommerce, marketplace, multi-product SaaS).
If you sell products from ₹500 to ₹50,000, tCPA will optimise for the cheapest conversions - usually the wrong ones. tROAS forces the algorithm to chase revenue, not volume.
Common Smart Bidding mistakes that destroy performance
- Setting unrealistic targets: Setting tCPA at 50% of current CPA → algorithm pulls back drastically and volume collapses.
- Frequent changes: Adjusting target every 2-3 days resets learning. Hold each target for at least 2 weeks.
- Mixing weak conversions: Counting "newsletter signup" and "₹5L purchase" as equal-value conversions destroys signal. Use conversion values.
- Ignoring conversion lag: If your customers convert 3-14 days after click, Smart Bidding is optimising on incomplete data. Adjust the conversion window.
- Running tCPA on a campaign with <15 conv/month: Algorithm has nothing to learn from. Stay manual.
Real-world results - when each strategy wins
From actual accounts I have managed:
- B2B SaaS, 12 leads/month: Manual CPC beat tCPA by 38% on CPL. Switched back permanently.
- Ecommerce, ₹8L/month spend: tROAS beat Max Conversions by 2.4x ROAS. Massive win.
- Local clinic, 80 leads/month: tCPA dropped CPL by 22% after 3 weeks of learning.
- Real estate, ₹4L/month, 18 leads/month: Manual + Enhanced CPC beat tCPA - sales cycle was too long for Smart Bidding signal.
The bottom line
Smart Bidding is not magic. It is statistical inference on top of your conversion data. If your data is rich and your tracking is clean, it usually wins. If either is weak, Manual CPC + a good operator wins.
Want a recommendation specific to your account? Book a free 30-minute audit - I will look at your conversion volume, tracking quality, and pick the right bid strategy for the next 90 days.