Google Ads Bidding Strategies: Manual CPC vs Smart Bidding Explained

There is a quiet argument inside almost every Google Ads account.

Should you control bids manually?

Or should you let Google’s Smart Bidding system make the decision?

On one side, you have the old-school media buyer.

They want control.

They want to decide what each click is worth.

They want to set a different bid for "enterprise crm software" and "crm software for small business."

They do not fully trust automation.

They remember when Google Ads was built around exact match keywords, manual CPC and tight campaign control.

On the other side, you have Google and the newer generation of performance marketers.

They argue that no human can calculate every auction signal in real time.

They believe Smart Bidding can read intent better than a person can.

They believe automation can see patterns that humans miss.

Both sides have a point.

Both sides can also be wrong.

The debate is not really "Manual vs. Auto."

The debate is Control vs. Context.

Manual CPC gives you control over the maximum bid.

Smart Bidding gives Google room to adjust bids based on context.

That context may include device, location, time of day, audience membership, query meaning, conversion history, browser signals and many other factors.

A human can look at a keyword.

Smart Bidding can look at the auction.

That difference matters.

If you are running Manual CPC in 2026 without a specific, data-backed reason, you may be limiting the account.

You may be bidding the same for weak users and strong users.

You may be underbidding for high-value auctions.

You may be overpaying for poor traffic.

But if you blindly apply "Maximize Conversions" to a new account with bad tracking, weak conversion data or the wrong goals, you can waste money quickly.

Automation is not magic.

It is only as good as the data and goals you give it.

This is not a generic overview.

This is the Mega-Authority Guide to Google Ads Bidding.

We are going to break down the maths behind Ad Rank, the role of Smart Bidding, the use cases for Manual CPC, and the Portfolio Bid Strategy settings that many advertisers miss.

The aim is simple.

Use manual bidding when control matters.

Use Smart Bidding when the data is strong enough.

Use portfolio strategies when you need both automation and guardrails.

Above all, make sure the bidding strategy matches the business goal.

---

Part 1: The Mathematics of the Auction

To master bidding, you must first respect the auction.

Google Ads uses an auction system where you do not simply pay your maximum bid.

You pay enough to compete based on Ad Rank, competition and other auction factors.

A simplified way to understand it is this:

$$ \text{Ad Rank} = \text{Max CPC Bid} \times \text{Quality Score} $$

$$ \text{Actual CPC} \approx \frac{\text{Ad Rank of Next Competitor}}{\text{Your Quality Score}} + 0.01 $$

This is simplified.

Modern Ad Rank uses more than Quality Score.

It can include ad quality, expected impact of assets, auction context, thresholds and other signals.

But the principle remains useful.

You do not win only by bidding more.

You win by bidding well and having a strong ad experience.

This formula shows that there are two broad ways to improve your economics:

1. Improve Quality: Better CTR, ad relevance and landing page experience.
2. Control Bids: Avoid overpaying for low-value traffic.

The mistake many advertisers make is focusing only on the bid.

They ask:

"How much should I pay per click?"

The better question is:

"What is this click likely to be worth?"

That is the difference between bidding and strategy.

The "Manual" Era

In the early days, "Max CPC Bid" was a static number.

You set it to $5.00.

If a user searched at 3:00 AM on a Tuesday from a low-converting location, your bid was $5.00.

If a decision-maker searched at 2:00 PM on a Wednesday from a high-value business district, your bid was also $5.00.

The keyword was the same.

So the bid was the same.

That was simple.

It was also limited.

Because two people can search the same keyword and have very different value.

One may be researching.

One may be ready to buy.

One may be a student.

One may be a purchasing manager.

One may be outside your service area.

One may have visited your pricing page yesterday.

Manual CPC does not naturally see that difference unless you build many manual layers around it.

This is the inefficiency that can hold campaigns back.

The "Smart" Era (Real-Time Bidding)

Smart Bidding changes the bid from a fixed number into a prediction.

$$ \text{Bid} = f(x_1, x_2, x_3, ... x_n) $$

Where $x$ represents signals such as:

• Device: Mobile, desktop, tablet.
• Operating System: iOS, Android, Windows, macOS.
• Time of Day: Morning, evening, working hours.
• Location: City, region, distance and local intent.
• Audience List: Past visitors, customers, cart abandoners.
• Query Context: The actual search term and its meaning.
• Demographics: Age, gender and household income where available.
• Conversion History: Patterns from previous converting users.

Smart Bidding estimates the Expected Conversion Rate and, where value data is available, the expected conversion value.

Then it adjusts the bid to help meet your target.

For Target CPA, the question is:

"How much can we bid and still average the target cost per action?"

For Target ROAS, the question is:

"How much can we bid and still achieve the target return?"

This is why Smart Bidding can be powerful.

It does not treat every searcher as the same.

But it also explains why Smart Bidding can fail.

If your tracking is wrong, the prediction is wrong.

If your conversion goals are weak, the system optimises for weak outcomes.

If your budget is too small, the model has limited room to learn.

If your account has very low volume, the system may not have enough data.

Smart Bidding is not a replacement for strategy.

It is a tool that needs clean inputs.

---

Part 2: Manual CPC (The Control Freak's Tool)

Manual CPC is not dead.

It is just no longer the default answer for every account.

Manual CPC is a precision instrument.

It gives you direct control over how much you are willing to pay for a click.

That can be useful in specific situations.

But it also limits the system’s ability to adjust for auction context.

How It Works

You set a maximum cost per click.

Google can charge up to that limit depending on the auction.

You choose the bid.

You accept the responsibility.

Manual CPC works best when the human has a clear reason to override automation.

It does not work well when it is used only because the advertiser does not trust Smart Bidding.

Distrust is not a bidding strategy.

Data is.

When to Use Manual CPC

1. Brand Protection: When bidding on your own brand name, you may want strong impression share at a controlled cost. Smart Bidding can sometimes overvalue Brand because conversion probability is high.
2. New Product Launches (Zero Data): If you are launching with no historical conversions, Smart Bidding has little account-specific data. Manual CPC or Maximise Clicks with a cap can help collect early data safely.
3. Strict Budget Caps: If you have a very small daily budget and cannot risk high individual CPCs, Manual CPC gives more direct cost control.
4. Low Volume, High Value B2B: In markets where clicks are expensive and conversion volume is very low, manual control may still be useful.
5. Competitor Testing: If you are testing competitor terms, Manual CPC can limit exposure while you learn whether the traffic is commercially useful.
6. Early Landing Page Testing: If you need controlled traffic to test whether a page converts, Manual CPC can prevent the account from overspending during setup.

Manual CPC is strongest when you know exactly what you are buying.

It is weaker when the value of each auction varies heavily.

For example, if you sell one local service in one town and only need a few leads per week, Manual CPC may still work.

If you run ecommerce across hundreds of products and customer values vary widely, Smart Bidding is usually better.

The decision should come from the business model.

Not habit.

The "Enhanced CPC" (eCPC) Trap

Google has historically offered "Manual CPC with Enhanced CPC."

This allows Google to adjust your manual bids when it predicts a conversion is more or less likely.

The problem is that it can become a half-measure.

You still manage base bids manually, but Google can still alter bids.

You do not get the full control of Manual CPC.

You also do not get the full structure of Smart Bidding.

• Recommendation: In most cases, go fully Manual or fully Smart. Avoid sitting in the middle without a reason.

If you use eCPC, know why.

Measure it.

Do not leave it on because it sounds safer.

---

Part 3: The Smart Bidding Ecosystem

There are several automated bidding strategies inside Google Ads.

Each has a different purpose.

The wrong strategy can damage performance even if the account structure is good.

The right strategy can unlock growth when the data is clean.

1. Maximize Clicks (The Traffic Firehose)

• Goal: Get as many visitors as possible within the budget.
• The Math: Google focuses on traffic volume, not conversion value.
• The Danger: It can find cheaper, lower-intent traffic to spend the budget.
• Use Case:
  • Early data collection with CPC caps.
  • Driving traffic to content for remarketing.
  • Technical testing.
  • Very early campaigns where conversion data is not yet reliable.

Maximize Clicks is often misunderstood.

It does what it says.

It maximises clicks.

It does not maximise leads.

It does not maximise sales.

It does not maximise profit.

That means it can be useful when you want traffic.

But it can be dangerous when you want customers.

If you use it, set a maximum CPC limit where available.

Without a cap, Google may still enter auctions that are more expensive than you expected.

Use Maximize Clicks as a controlled early-stage tool.

Not as a long-term performance strategy.

2. Maximize Conversions (The Volume Play)

• Goal: Get the maximum number of conversions within the daily budget.
• The Math: Google tries to spend the budget to drive as many conversion actions as possible.
• The Danger: It does not care about your target CPA unless you set one.
• Use Case:
  • Newer campaigns with some conversion data.
  • Accounts where volume matters more than strict efficiency.
  • Early learning before adding a target CPA.

Maximize Conversions can work well.

But it needs clean conversion tracking.

If your primary conversion is a weak action, it will optimise for that weak action.

If your budget increases too quickly, it may expand into more expensive traffic.

If you do not set a target, it may spend aggressively to find more conversions.

This is not always bad.

Sometimes you want volume.

Sometimes you want the system to explore.

But you need to know what you are asking it to do.

Do not use Maximize Conversions if the business needs strict CPA control from day one.

Use it when you are prepared for some fluctuation in exchange for learning and volume.

3. Target CPA (tCPA) (The Efficiency Play)

• Goal: Get as many conversions as possible at or around a specific cost per action.
• The Math: You tell Google: "I am willing to pay $50 for a lead."
  • If Google estimates a user has a 10% chance of converting, it may bid around $5.00.
  • If a user has a 1% chance, it may bid much less.
• The Danger: Setting the target too low can restrict delivery.
• Use Case: Lead Generation, SaaS, services, local businesses and many non-ecommerce accounts.

Target CPA is often the core strategy for lead generation.

But it must be realistic.

If your historical CPA is $100 and you set a target of $20, Google may struggle to find traffic.

The campaign can slow down.

Impressions can drop.

Volume can disappear.

The system is not being stubborn.

It is doing the maths.

If it cannot find enough auctions likely to convert at your target, it will reduce participation.

A good tCPA setup usually starts close to recent performance.

Then you adjust gradually.

For example:

1. Historical CPA: $80.
2. Initial tCPA: $80 to $90.
3. Stabilise.
4. Reduce by 10% after enough data.
5. Monitor conversion volume and lead quality.

Do not cut the target aggressively just because the business wants cheaper leads.

The market has a floor.

You can improve that floor with better landing pages, better offers, better tracking and better account structure.

But you cannot wish it away.

4. Target ROAS (tROAS) (The Profit Play)

• Goal: Maximise conversion value at a target return.
• The Math: You tell Google what return you need from ad spend.
• The Danger: Requires accurate value tracking and enough conversion value data.
• Use Case: Ecommerce and value-based lead generation.

Target ROAS is powerful when values vary.

For ecommerce, this is obvious.

One customer may spend $20.

Another may spend $800.

You do not want to treat those conversions equally.

For lead generation, tROAS can also work when you import values.

For example:

1. Raw lead = $10.
2. Qualified lead = $100.
3. Sales accepted lead = $500.
4. Closed deal = actual revenue.

This helps Google optimise towards quality.

Not just quantity.

The danger is bad value data.

If the values are wrong, Smart Bidding learns the wrong lesson.

If all leads have the same value even though some are much better, tROAS cannot understand the difference.

If revenue is missing from some purchases, bidding can become distorted.

Target ROAS is only as good as the values you pass back.

---

Part 4: Portfolio Bidding Strategies (The Enterprise Secret)

Most beginners apply bidding strategies at the Campaign Level.

More advanced advertisers often use Portfolio Bid Strategies.

A Portfolio Bid Strategy is a shared strategy that can apply across multiple campaigns.

This can be useful when several campaigns share the same goal.

Why Portfolio Bid Strategies are Superior

1. Shared Strategy: If 5 campaigns all have the same CPA goal, one portfolio strategy can manage them together.
2. More Stable Learning: Similar campaigns can benefit from being managed under the same goal.
3. Maximum CPC Cap: Some portfolio strategies allow bid limits that are not always available in standard campaign-level Smart Bidding.
4. Cleaner Management: One strategy can be adjusted instead of editing many campaigns separately.

The maximum CPC cap is especially useful.

In standard tCPA, Google may bid high for an auction it believes is valuable.

Sometimes that is fine.

Sometimes it creates uncomfortable outliers.

In Portfolio tCPA, you may be able to set a "Maximum CPC Limit."

For example, $20.

This allows automation, but adds a safety rail.

Be careful with bid caps.

Too low and you can choke performance.

Too high and they do not protect much.

A good cap should protect against extreme clicks without blocking normal auctions.

Use historical CPC data to guide the cap.

Portfolio strategies work best when grouped campaigns have similar goals.

Do not combine campaigns with very different value profiles.

For example, do not put Brand, Competitor and Generic campaigns into one portfolio strategy unless you have a very clear reason.

Different intent needs different targets.

---

Part 5: The "Learning Phase" Myth vs Reality

You will often see the status "Learning" in Google Ads.

This can worry advertisers.

Sometimes they touch the campaign too often because they think something is wrong.

That can make things worse.

What is actually happening?

The algorithm is recalibrating.

It is learning which auctions are more likely to produce the goal you selected.

It is testing different patterns.

It is comparing signals.

It is trying to find stable performance.

This period can be volatile.

That does not mean the strategy is broken.

It means the system is still adjusting.

The Rules of the Learning Phase

1. Do Not Touch It Too Often: Major changes can restart learning or disrupt the model.
2. The "Wobble": Performance may be unstable. One day can look strong. Another can look poor.
3. Data Thresholds: More conversion data usually means more stable bidding.

As a practical rule, many accounts benefit from around 30 conversions in 30 days for a target-based strategy to work more reliably.

This is not a hard law.

Some accounts work with less.

Some need more.

The point is simple.

More relevant conversion data helps.

Less data creates more variance.

How to Bypass the Learning Phase?

You cannot truly bypass learning.

But you can reduce the risk.

If you do not have enough conversions, you have two options:

1. Micro-Conversions: Optimise for actions that happen more often, such as add to cart or qualified page engagement.
   • Warning: This can train the system to find people who complete the micro-action, not people who buy.
2. Data Aggregation: Use portfolio strategies or consolidated account structures to pool relevant data.

Micro-conversions should be used carefully.

They are better than nothing when the account has no conversion data.

But they can create false confidence.

A person who spends two minutes on a page is not the same as a person who buys.

A person who starts checkout is not the same as a completed purchase.

A person who clicks a phone number is not always a qualified lead.

Use micro-conversions as stepping stones.

Not as the final goal.

---

Part 6: Value-Based Bidding (The Future)

Keyword bidding is no longer enough.

Audience bidding is not enough either.

Value bidding is the future.

Why?

Because not all conversions are equal.

In B2B, not all leads are equal.

• Lead A: Gmail address, 5 employees. (Low value)
• Lead B: Corporate email, 5,000 employees. (High value)

If you use standard tCPA, Google may treat Lead A and Lead B as the same conversion.

It will often optimise for the cheaper lead.

That can reduce cost per lead while damaging sales quality.

This is one of the most common problems in lead generation.

The marketing dashboard looks better.

The sales team feels worse.

The fix is to feed better data back into Google Ads.

Offline Conversion Import (OCI) is the solution.

You upload deeper funnel outcomes back into Google Ads.

Then, you can switch to Maximize Conversion Value or value-based bidding.

You tell Google:

• Lead = Value $1
• Qualified Lead = Value $100
• Closed Deal = Value $5,000

Now the system can learn which clicks become valuable customers.

Not just which clicks become cheap leads.

This is how you scale quality.

Not just quantity.

For ecommerce, value-based bidding often means passing actual order value.

For lead generation, it means connecting the CRM.

For hotels, it may mean booking value.

For finance, it may mean approved applications or funded deals.

For SaaS, it may mean pipeline value or closed revenue.

The closer your conversion data gets to real business value, the stronger your bidding decisions become.

---

Part 7: Advanced Calibration (The Tools You Aren't Using)

Smart Bidding is powerful.

But it is not clairvoyant.

It cannot automatically know your site will break tomorrow.

It cannot know your call centre will be closed for training.

It cannot know your Black Friday offer will double conversion rates unless you tell it.

This is where advanced controls matter.

Seasonality Adjustments

Scenario: You are launching a Black Friday promo. You expect Conversion Rate (CVR) to double.

Problem: Smart Bidding works from recent data. It may not immediately understand a short, sharp conversion rate spike.

Solution:

1. Go to Tools → Bid Strategies → Advanced Controls → Seasonality Adjustments.
2. Create a "New Event".
3. Tell Google: "From Nov 24 to Nov 27, expect CVR to increase by +100%."
4. Result: The algorithm can adjust more quickly during the promotion and return after it ends.

Use this for short events.

Good examples include:

1. Black Friday.
2. Flash sales.
3. Major limited-time promotions.
4. Short booking windows.
5. Product launches with heavy demand spikes.

Do not use seasonality adjustments for normal seasonal trends.

Google can often learn normal patterns.

Use it when something unusual and temporary is about to happen.

Data Exclusions

Scenario: Your tracking tag breaks for 3 days. You record 0 conversions.

Problem: The algorithm may think your ads stopped working. It may reduce bids or change behaviour based on bad data.

Solution:

1. Go to Tools → Bid Strategies → Advanced Controls → Data Exclusions.
2. Tell Google: "Ignore all data from Jan 10 to Jan 13."
3. Result: The algorithm avoids learning from broken tracking data.

Data exclusions are essential when tracking fails.

Use them when:

1. The conversion tag broke.
2. The site was down.
3. Payment tracking failed.
4. A CRM import failed.
5. Duplicate conversions were recorded.
6. A consent issue distorted data.
7. A checkout bug affected performance.

Do not use them to hide normal bad performance.

That is not the purpose.

They are for data errors.

Not disappointing days.

---

Part 8: Troubleshooting: Why Is My Bid Strategy Broken?

When campaigns flatline, it is usually one of several problems.

Do not panic.

Diagnose it.

1. "Misconfiguration" (The Conversion Action Error)

If you see this status, check your Conversion settings.

• Primary vs Secondary: Smart Bidding optimises for primary conversion actions included in the campaign goal. If your main "Lead" goal was moved to Secondary by mistake, the algorithm may see no meaningful target.
• Attribution Model: If you recently changed attribution, conversion numbers and timing may shift.
• Duplicate Goals: If duplicate conversions are included, the system may overestimate performance.
• Wrong Goal: If page views are primary, the system may optimise for low-value traffic.

Always start with conversion settings.

A bidding problem is often a tracking problem.

2. "Limited by Budget" (The Cap)

Standard tCPA respects your daily budget.

If your Daily Budget is $100 and your tCPA is $50, the maths suggests you may only afford around 2 conversions per day.

Google may throttle participation because the budget cannot cover enough auctions.

Raising budget can sometimes improve performance because the system has more room to enter auctions.

But this is not guaranteed.

You should only raise budget when:

1. CPA is acceptable.
2. Conversion quality is good.
3. Search terms are clean.
4. The campaign is limited by budget.
5. The business can handle more leads or orders.

Do not increase budget to solve a broken campaign.

Fix the foundation first.

3. "Limited by Inventory" (The Target)

If your tCPA is $10 and the market rate for a click is $25, you may be mathematically constrained.

Google cannot find enough conversion opportunities at the price you want.

Traffic drops.

Impressions fall.

Conversions slow.

Fix: Raise the tCPA target, improve conversion rate, improve landing page quality or change the economics.

There is no trick that makes an unrealistic target work forever.

The market has a price.

Your job is to improve the business case so that price becomes profitable.

4. "Bad Signal" (The Quality Problem)

Sometimes the bidding strategy is working exactly as instructed.

The problem is that the instruction is bad.

For example:

1. It is optimising for all leads, but only some leads are good.
2. It is optimising for purchases, but values are missing.
3. It is optimising for calls, but many calls are too short.
4. It is optimising for form submissions, but spam is included.
5. It is optimising for low-value products instead of profitable products.

In this case, changing bidding strategy is not enough.

You need better conversion data.

---

Part 9: Deep Case Study: The B2B SaaS Scale-Up

Let's look at a practical example from the "HR Software" space.

Context:

• Budget: $50,000/mo.
• Goal: Demo Requests.
• Starting Point: Client was using Maximize Conversions with no target.

The "Before" Metrics (Month 1):

The Diagnosis:

"Maximize Conversions" was spending most of the available budget, but it was buying too many low-quality leads.

The campaign was optimising for form submissions.

The sales team cared about qualified demos.

Those are not the same thing.

The account had three problems:

1. The conversion goal was too shallow.
2. The bidding strategy had no efficiency constraint.
3. Broad matching was pulling in weak searches.

The "Shift" (Month 2):

We switched to Portfolio tCPA with a Max CPC Cap.

• Strategy: Target CPA set to $250.
• Bid Cap: Max CPC $40.00.
• Broad Match: Paused initially. Phrase and Exact were used while quality stabilised.

The "After" Metrics (Month 3 - Post Learning):

Analysis:

By giving the algorithm an Efficiency Constraint instead of only a Volume Goal, we changed its behaviour.

It stopped chasing easy, low-quality conversions.

It focused more budget on searches with stronger commercial intent.

It reduced waste.

It gave the sales team better opportunities.

The lesson is simple.

The bidding strategy must match the business goal.

If you want volume, Maximize Conversions may work.

If you want controlled lead cost, use Target CPA.

If you want quality and revenue, move towards value-based bidding and offline conversion imports.

This is the power of the correct bidding strategy.

---

Part 10: Step-by-Step Implementation Guide

Here is how to set up a practical bidding structure in 2026.

This is your deployment checklist.

Step 1: The "Data Collection" Phase (Cold Start)

Objective: Get enough clean conversion data to give Smart Bidding a baseline.

1. Open your Campaign Settings.
2. Navigate to Bidding.
3. Select Maximize Clicks or Manual CPC for controlled early traffic.
4. Crucial: Set a maximum cost per click bid limit if available.
5. Set this based on your industry and expected conversion rate.

The number should not be random.

If your service sells for £10,000, a £10 click may be fine.

If your product sells for £30, it may not be.

Use basic economics.

Ask:

1. What is the product or lead worth?
2. What conversion rate do we expect?
3. What CPC can we afford during testing?
4. How much data do we need before switching?

The data collection phase is not about cheap traffic only.

It is about clean traffic.

Do not buy junk just because it is cheap.

Step 2: The "Soft Transition" (The Experiment)

Objective: Prove that Smart Bidding works better than your current strategy without risking the whole budget.

1. From the left sidebar, click Experiments → All Experiments.
2. Click the blue + button and select Custom Experiment.
3. Campaign Type: Search.
4. Base Campaign: Select your data collection campaign.
5. Experiment Name: "Test - Max Conv vs tCPA".
6. Configuration:
   • Change Bidding Strategy to Maximize Conversions with Target CPA.
   • Set the Target CPA close to your average CPA from the last 30 days. Do not set it much lower yet.
7. Split: 50% Base / 50% Trial.
8. Split Options: Use cookie-based where suitable.

Experiments are useful because they reduce risk.

They stop you making emotional account-wide changes.

They let the data speak.

Step 3: Analysis & Graduation

Objective: Validate the hypothesis.

1. Let the experiment run for 4 weeks where volume allows.
   • Week 1: Learning Phase. Treat early data carefully.
   • Week 2-4: More useful data.
2. Check the Experiment Scorecard.
3. Success Criteria:
   • Conversion Volume maintained or increased.
   • CPA decreased by more than 10%.
   • Lead quality stayed the same or improved.
   • Search term quality did not collapse.
4. If successful, click Apply to turn the experiment into your main campaign.

Do not judge only by platform conversions.

Check business quality.

Especially in lead generation.

Step 4: The "Portfolio" Move (Enterprise Grade)

Objective: Protect the account from volatility.

1. Go to Tools → Budgets and bidding → Bid strategies.
2. Click + → Target CPA.
3. Name it: "Portfolio - Core Service - tCPA $50".
4. Settings:
   • Target CPA: $50.
   • Advanced Options: Maximum bid limit: $15 where available.
5. Select your new campaign and apply this strategy.
6. Result: You now have automation with safety rails.

This is not always needed.

For small accounts, campaign-level bidding may be enough.

For larger accounts, portfolio strategies can make management cleaner and safer.

Use the right tool for the size of the account.

---

Glossary of Key Terms

To master bidding, you must speak the language of the algorithm.

Target CPA (tCPA):

A smart bidding strategy where you set a cost-per-action goal. Google predicts conversion probability and adjusts bids to try to achieve that average cost.

Target ROAS (tROAS):

"Return On Ad Spend." A value-based strategy where you set a revenue multiplier, such as 400%. Requires accurate conversion value data.

Portfolio Bid Strategy:

A shared bidding strategy that can control multiple campaigns. Some portfolio strategies allow maximum and minimum bid limits that give extra control.

Learning Phase:

A period after a major strategy or campaign change where the algorithm recalibrates. Performance can be volatile. Avoid unnecessary changes during this period.

Data Exclusion:

A tool that tells Google to ignore a specific date range, often because of tracking outages or data errors.

Seasonality Adjustment:

A tool that tells Google to expect a temporary change in conversion rate during short events, such as flash sales.

Signal Density:

The amount of useful data available to the bidding system. More relevant conversion data usually helps Smart Bidding perform more reliably.

Attribution Window:

The period of time after a click during which a conversion can be credited to an ad interaction.

---

Conclusion

Bidding is no longer about finding the perfect price for one keyword.

It is about building the right data flow.

Your job is not only to move bids up and down.

Your job is to make sure the bidding system receives the right signals.

That means:

1. Feed it data: Accurate conversion tracking and offline conversion imports.
2. Give it boundaries: Portfolio Strategies, realistic targets and sensible budgets.
3. Guide the goal: Correct CPA, ROAS or value-based targets.

If you do these three things, Smart Bidding will usually outperform Manual CPC in accounts with enough clean data.

If you do not, you may be better off bidding manually while you fix the data infrastructure.

That is the honest answer.

Automation wins when the foundation is strong.

Manual control helps when the foundation is weak, volume is low or risk needs to be limited.

The mistake is treating either side like a religion.

Good bidding is not ideology.

It is judgement.

Ready to audit your account setup? Check out our guide on Account Structure or read up on Keyword Research to ensure you're feeding the right queries into the machine.

---

The Bidding Strategy Hierarchy

Not all strategies are created equal.

We categorise them by maturity and control level:

• Tier 1 — Control: Manual CPC. Your bid is your bid. Good for strict control, low volume and specific tests.
• Tier 2 — Volume: Maximize Clicks and Target Impression Share. Useful for traffic, visibility or brand protection when chosen deliberately.
• Tier 3 — Efficiency: Maximize Conversions with Target CPA. Useful when you need controlled lead or sale costs.
• Tier 4 — Value: Target ROAS and Maximize Conversion Value. Best when you have reliable revenue or value data.

The "Truthful Bidding" Concept

If your break-even CPA is $100, do not set your tCPA to $20.

That is not ambition.

That is denial.

If you starve the algorithm with an unrealistic target, Google may stop entering many auctions.

It calculates that it cannot win a conversion at that price, so traffic drops.

Start with a tCPA close to your historical average.

For example:

Historical CPA: $50.

Initial tCPA: $55 or $60.

Then walk it down by around 10% every 2 weeks if volume and quality remain stable.

This is how you improve performance without shocking the account.

The "Competitor Conquesting" Play (Target Impression Share)

Target Impression Share is often treated as a brand strategy.

It can also be used for competitor campaigns.

• Keyword: [competitor name]
• Strategy: Target Impression Share, often top of page or absolute top where commercially justified.
• Outcome: Your ad appears when someone searches for your competitor.

This can create visibility.

But it is expensive.

Warning: CPCs can be several times higher. Conversion rates may be lower. Quality Score may be weaker. Only use this if you have high LTV, strong differentiation and a landing page built for comparison.

Do not send competitor traffic to a generic homepage.

If someone searches for a competitor, they need a reason to switch.

Give them one.

The "p(Conv)" Formula

Google estimates a probability of conversion for each auction.

A simplified version looks like this:

• If p(Conv) = 2% and your tCPA is $50 → it may bid around $1.00
• If p(Conv) = 20% and your tCPA is $50 → it may bid around $10.00

This is why Smart Bidding can outperform Manual CPC at scale.

It runs this logic across many auctions using signals that a human cannot calculate in real time.

But the formula only works if the conversion goal is meaningful.

If the conversion is weak, the system gets better at finding weak conversions.

If the conversion is valuable, the system gets better at finding valuable customers.

That is the whole game.

Final Rule

Do not ask:

"Which bidding strategy is best?"

Ask:

"What data do we have, what goal are we optimising for, and what level of control do we need?"

That question leads to better decisions.

For a new account, you may start with Manual CPC or Maximize Clicks with a cap.

For a growing lead generation account, you may move to Target CPA.

For ecommerce, you may move to Target ROAS.

For advanced B2B, you may move to offline conversion imports and value-based bidding.

For enterprise accounts, you may use portfolio strategies, bid caps, data exclusions and seasonality adjustments.

That is the maturity path.

Start with control.

Move to automation.

Then move to value.