Google Ads Best Practices · June 22, 2026 · 13 min read

How To Choose The Right Google Ads Smart Bidding Strategy

Alexander PerelmanHead Of Product @ groas

A Google Ads smart bidding strategy is the automated bid approach you assign to campaigns so Google’s AI optimizes toward your business goal, whether that is a target CPA, a target ROAS, maximum conversions, or maximum conversion value. Choosing the right smart bidding strategy depends on three variables: your monthly conversion volume, your business objective, and how mature your account data is. The wrong strategy on the wrong account does not just underperform. It actively destroys performance by training Google’s algorithm on bad signals, burning budget during extended learning phases, and locking you into a feedback loop that gets harder to escape over time.

By the end of this guide, you will know exactly which smart bidding strategy to select, how to set realistic targets, and how to scale without crashing volume. You will also know when to switch strategies as your account matures.

Before You Start

You will need access to the Google Ads interface with at least Standard access to the account. Before selecting any bidding strategy, confirm these prerequisites:

At least 30 days of conversion data in the account (or a clear understanding that you do not have this yet)
A verified primary conversion action that tracks your actual business outcome (purchase, qualified lead, booked call), not a micro-conversion like page views
Access to Google Analytics 4 if you plan to import conversions from there
Clarity on your core business metric: is it cost per acquisition, return on ad spend, or raw lead volume?

If you are missing any of these, start there before touching your bidding strategy.

Step 1. Audit Your Monthly Conversion Volume Before Choosing Any Smart Bidding Strategy

Smart bidding strategy selection starts with one number: how many conversions does this campaign (or campaign group) generate per month? Google’s AI learns from conversion signals. Without enough signals, it cannot learn, and it will bid erratically.

Here is the practical threshold:

Maximize Conversions / Maximize Conversion Value: These uncapped strategies can function with as few as 15 conversions per month because they are not optimizing toward a fixed target. They simply try to get as many conversions (or as much value) as possible within your budget.
Target CPA (tCPA): Requires roughly 30 conversions per month per campaign to stabilize. Below that, the algorithm lacks enough signal to hit your target consistently.
Target ROAS (tROAS): Requires more data than tCPA because it is optimizing for value, not just count. Aim for 50 or more conversions per month with accurate revenue values attached before you trust tROAS.

How To Find This Data

In Google Ads, navigate to Campaigns, select your date range to the last 30 days, and look at the Conversions column. If you see “Limited by learning” in the Bid Strategy status column, your campaign does not have enough data for the strategy you have assigned. That status is not a temporary inconvenience. It means the algorithm is guessing, and you are paying for those guesses.

If your account generates fewer than 30 conversions per month at the campaign level, start with Maximize Conversions (not tCPA). Jumping to a target-based strategy too early is the single most common smart bidding mistake, and it is the one that costs the most money.

Step 2. Map Your Business Objective To The Right Bidding Strategy

Each smart bidding strategy is designed for a specific situation. Choosing based on what sounds best instead of what fits your data and goal is how accounts stall.

Maximize Conversions: When It Wins And When It Overspends

Use Maximize Conversions when you are building conversion volume in a new or low-volume campaign and your primary goal is acquisition count, not cost efficiency. It will spend your entire daily budget trying to get as many conversions as possible. That is its strength and its risk. If you set a $500/day budget, it will spend $500/day whether it finds good traffic or not. Keep budgets tight and monitor CPA daily.

Maximize Conversion Value: The Ecommerce Default

If you sell products at varying price points, Maximize Conversion Value tells Google to prioritize higher-value transactions. This is the right starting strategy for most ecommerce accounts before they have enough data to layer on a tROAS target. Its ceiling: it does not care about profitability, only revenue.

Target CPA: The Lead Gen Workhorse

Target CPA is the strategy for lead generation accounts that have cleared the 30-conversion threshold and know their maximum acceptable cost per lead. Introduce tCPA after running Maximize Conversions for long enough to establish a baseline CPA. Set the initial target at or slightly above that baseline, not at your ideal number. If your average CPA over the last 30 days is $45, set tCPA to $45 or $50, not $30. The algorithm needs room to learn before you tighten.

Target ROAS: The Ecommerce Scaling Lever

Target ROAS tells Google to bid for conversions that hit a specific return threshold. It is powerful at scale, but it constrains volume. The higher you set your tROAS target, the fewer auctions Google will enter. Introduce tROAS only after running Maximize Conversion Value and collecting 50 or more conversions with accurate value tracking. For B2B and SaaS accounts importing offline conversions, tROAS can also work if pipeline values are fed back to Google Ads consistently.

Enhanced CPC: Still Relevant For Manual-To-Smart Transitions

Enhanced CPC (eCPC) adjusts your manual bids up or down based on conversion likelihood. It is a bridge strategy if you are transitioning from fully manual bidding to automated strategies and want to maintain some bid control. It is not a long-term solution.

Portfolio Bid Strategies: When To Consolidate

If you have multiple campaigns targeting similar audiences or products, a portfolio bid strategy pools conversion data across them. This is useful when individual campaigns fall below conversion thresholds on their own but collectively have sufficient volume. Consolidating data this way gives the algorithm a bigger learning set without restructuring your campaigns.

Step 3. Set Targets That Reflect Reality, Not Ambition

Setting a tCPA or tROAS target is not about what you want to pay or earn. It is about where the algorithm can actually perform given current market conditions, competition, and your landing page conversion rates.

How To Calculate A Realistic Starting tCPA

Pull your last 30 to 60 days of conversion data. Calculate the median CPA (not the average, which outliers skew). Set your initial tCPA target at that median or 10-15% above it. This gives the algorithm headroom to optimize without starving it of auction opportunities on day one.

The tROAS Target Death Spiral

This is where ecommerce accounts crash most often. You set tROAS at 800% because that is what the business needs. Google cannot find enough auctions at that threshold, so volume drops. You panic and raise the target further or slash the budget. Volume drops more. The algorithm loses signal. Performance degrades further. The fix: start tROAS at your actual 30-day average ROAS and tighten in 10-15% increments over weeks, not days.

The 20% Rule For Changes

Any single change to your tCPA, tROAS, or daily budget should not exceed 20% at a time. Larger jumps reset the learning phase and wipe out the algorithm’s accumulated signal. Make incremental adjustments, wait five to seven days between changes, and track the impact before moving again.

Step 4. Give The Algorithm Clean Conversion Signals Before You Trust It

The best smart bidding strategy in the world produces bad bids if it is optimizing toward the wrong conversion action. This is the step most advertisers skip, and it is the one that matters most.

Why Bad Conversion Actions Produce Bad Bids

If your primary conversion action is “page view” or “button click” instead of “purchase” or “qualified lead,” tCPA will happily optimize for cheap page views all day. It will hit your target and report great numbers while generating zero revenue. Audit your conversion actions before selecting any smart bidding strategy. Your primary conversion should be the action closest to revenue.

Set Up Enhanced Conversions

Enhanced conversions improve match rates by sending hashed first-party data (email, phone, address) to Google alongside your conversion tags. This gives Google better signal quality, which directly improves bid accuracy. Without enhanced conversions, you are asking the algorithm to optimize with partial information. Set this up in Google Ads under Settings, then Conversions, and then select the conversion action to enable enhanced conversions. Accounts with signal quality issues frequently see performance recover once this is addressed.

Google Ads Native Tracking Vs GA4 Import

Use native Google Ads conversion tracking whenever possible. It captures view-through conversions and provides richer signal to the bidding algorithm. GA4 imported conversions work but have attribution model differences and a data lag that can slow learning. If you must use GA4 imports, be aware that your smart bidding strategies are working with slightly stale data.

Offline Conversion Import For B2B And Lead Gen

If your true conversion happens offline (a signed contract, a qualified appointment, a closed deal), import those outcomes back into Google Ads. Without offline conversion import, tCPA optimizes for form fills, not revenue. With it, the algorithm learns which clicks actually produce pipeline. This is non-negotiable for B2B accounts running smart bidding.

Step 5. Monitor Learning Phase Duration And Know When To Intervene

Every time you change a bid strategy, target, or budget by more than 20%, the campaign enters a learning phase. During this period, performance is volatile and CPA/ROAS numbers will swing.

The learning phase typically lasts five to seven days but can stretch to two weeks for low-volume campaigns. During this time, do not make additional changes. Stacking changes on top of a learning phase resets it and extends the instability.

What Resets The Learning Phase

Changing the bid strategy type
Adjusting tCPA or tROAS by more than 20%
Changing daily budgets by more than 20%
Adding or pausing a significant number of keywords or ad groups
Changing the primary conversion action

When To Override Before Learning Finishes

If spend is running away with no conversions for three or more consecutive days during a learning phase, intervene. The algorithm may be exploring traffic segments that will never convert. Pause, revert to the previous strategy, and investigate. Blind patience with a runaway learning phase is just burning budget with extra steps.

Signals The Algorithm Has Stabilized

The Bid Strategy status in Google Ads changes from “Learning” to “Eligible.” Daily CPA or ROAS variance narrows. Impression share stabilizes. At that point, you can begin incremental target adjustments.

Step 6. Scale Smart Bidding Targets Gradually Once Stable

Scaling is where the compound effect of smart bidding shows up. Once your strategy is stable and hitting targets, tighten incrementally.

For tROAS, increase the target by 10-15% every two weeks. For tCPA, decrease the target by 10-15% on the same schedule. Track volume alongside efficiency. If conversions drop by more than 20% after a target change, you have overtightened. Pull back to the previous target and wait.

When To Switch Strategies As The Account Matures

The natural progression for most accounts follows this path:

New/low-volume: Maximize Conversions (lead gen) or Maximize Conversion Value (ecommerce)
30+ conversions/month: Layer on tCPA (lead gen) or tROAS (ecommerce)
Mature/high-volume: Tighten targets, test portfolio strategies, and scale budgets

This is not a one-time decision. Revisit your bidding strategy quarterly or whenever conversion volume changes significantly. An account audit should include a bidding strategy review as a standard step.

Quick Reference: Smart Bidding Strategy Selection Matrix By Account Type

Lead gen with fewer than 30 conversions/month: Maximize Conversions. No target. Focus on building volume.

Lead gen with 30+ conversions/month: Target CPA. Set at historical median CPA. Tighten 10-15% every two weeks.

Ecommerce with fewer than 50 conversions/month: Maximize Conversion Value. No target. Let Google chase revenue.

Ecommerce with 50+ conversions/month: Target ROAS. Set at actual 30-day ROAS. Tighten gradually.

B2B with offline conversions: Target CPA after importing offline data. Do not run tCPA on form fills alone.

Agencies managing multiple client accounts: Portfolio bid strategies where campaigns share audience overlap. Consolidate learning data across campaigns to hit thresholds faster. Agencies scaling across 20 or more client accounts need systematized bidding strategy selection, not ad hoc decisions per campaign.

Common Mistakes To Avoid

Starting with tCPA or tROAS on a brand-new campaign. The algorithm has no data. It will overbid or underbid wildly. Start with an uncapped Maximize strategy first.

Setting targets based on what you want, not what the data says. A $20 tCPA goal on an account averaging $55 will get you zero impressions. Start at the historical baseline.

Changing targets too frequently. Every major change resets learning. Make one adjustment, wait a week, measure, then decide.

Optimizing toward micro-conversions. Page views, scroll depth, and video plays are not business outcomes. Your primary conversion action must be the event closest to revenue.

Ignoring the Bid Strategy report. Google Ads provides a Bid Strategy report (under Tools, then Bid Strategies) that shows top signals, auction insights, and performance trends. Most advertisers never look at it.

Running tROAS without accurate value tracking. If your product values in the data feed are wrong, tROAS optimizes toward the wrong revenue number. Verify your values match actual transaction revenue.

Using the same strategy across fundamentally different campaign types. A brand campaign and a non-brand prospecting campaign have different conversion rates, different intent signals, and different CPAs. They rarely belong on the same bid strategy.

How groas Handles Smart Bidding Strategy Selection For You

Every step in this guide, from auditing conversion volume to setting targets to monitoring learning phases to scaling incrementally, requires ongoing attention, judgment calls, and the discipline to not overcorrect. Most teams either lack the time to monitor daily or lack the data pattern recognition to know when a learning phase is going sideways versus working as intended.

groas eliminates this entire workflow. For DFY (Done For You) clients, a dedicated strategist owns your bidding strategy selection end-to-end, backed by a proprietary engine trained on over $500 billion in profitable ad spend. That engine identifies the optimal strategy and target for each campaign based on real-time conversion data, adjusts targets incrementally without resetting learning phases, and monitors performance around the clock. Your strategist does not guess at targets. The engine surfaces what the data supports, and the strategist validates and acts.

For DWY (Done With You) clients, your in-house team stays in control while the engine and a senior strategist work alongside you. You get the data analysis and recommendations without the guesswork. Your team makes the final calls, but with pattern recognition that no single media buyer can replicate manually.

For agencies using the DIY product, the groas engine runs underneath your campaigns across every client account. Your media buyers operate the engine directly, getting bid strategy recommendations and execution at a scale that would otherwise require hiring more staff. Agencies keep their clients, their brand, and their margin while the engine handles the computational heavy lifting.

No onboarding fees. Month-to-month, cancel anytime. groas earns the next month by performing.

What To Do Next

Smart bidding strategy selection is not a set-it-and-forget-it decision. It is a living process that requires regular audits, incremental adjustments, and the pattern recognition to know when the data supports a strategy change. Get it right and smart bidding compounds your results. Get it wrong and you train Google’s algorithm to waste your budget efficiently.

If you want this handled by a team that has seen the patterns across billions in ad spend, groas is the direct path. DFY clients can apply and have a strategist own this end-to-end. DWY clients can get started and keep their team in the driver’s seat with the engine and a strategist alongside them. Agencies can start a 7-day free trial and run the engine across every client account immediately.

The strategy framework is here. The question is whether you want to run it manually or let an engine trained on $500 billion in profitable spend do the heavy lifting.

Frequently Asked Questions

How Many Conversions Do I Need Before Using Target CPA?

Target CPA requires roughly 30 conversions per month per campaign to stabilize. Below that threshold, the algorithm lacks sufficient conversion signals to bid accurately, and you will see erratic CPA swings. If your campaign generates fewer than 30 conversions per month, start with Maximize Conversions instead. Build volume first, establish a baseline CPA over 30 to 60 days, then layer on tCPA set at or slightly above that baseline. Jumping to tCPA too early is the most expensive smart bidding mistake because it trains Google on insufficient data, producing bids that either overspend or suppress volume entirely.

When Should I Use Target ROAS Vs Target CPA In Google Ads?

Use Target CPA when your business measures success by cost per acquisition, which is typical for lead generation, SaaS, and service businesses. Use Target ROAS when you sell products or services at varying price points and need Google to prioritize higher-value transactions, which is typical for ecommerce. tROAS requires more data than tCPA (aim for 50 or more conversions per month with accurate revenue values), so it is generally introduced later in an account's maturity. B2B accounts can use tROAS if they import offline revenue values back into Google Ads consistently.

What Happens If I Set My Target CPA Too Low?

Setting tCPA below your historical average starves the algorithm of auction opportunities. Google will stop entering auctions it cannot win at your target price, which causes impression volume to collapse. Fewer impressions mean fewer clicks, fewer conversions, and less data for the algorithm to learn from. This creates a downward spiral: less data leads to worse bids, which leads to even less volume. The fix is to set your initial tCPA at or above your actual median CPA from the last 30 to 60 days and tighten by 10 to 15 percent every two weeks once performance stabilizes.

How Long Does The Google Ads Smart Bidding Learning Phase Last?

The learning phase typically lasts five to seven days but can stretch to two weeks for low-volume campaigns. During this period, performance will be volatile and CPA or ROAS metrics will swing. Do not make additional changes during the learning phase, as stacking changes resets it and extends instability. If spend is running with zero conversions for three or more consecutive days, intervene by reverting to the previous strategy. The learning phase ends when the Bid Strategy status changes from "Learning" to "Eligible" and daily performance variance narrows.

Should I Use Maximize Conversions Or Target CPA For A New Google Ads Campaign?

For new campaigns, always start with Maximize Conversions. A new campaign has no conversion history, so Target CPA has nothing to optimize toward. Maximize Conversions builds volume without a fixed target, generating the data the algorithm needs. Once you accumulate 30 or more conversions per month and have a clear baseline CPA, switch to Target CPA set at that baseline. This staged approach prevents the algorithm from making blind bids and gives you a foundation of real performance data to work from.

Can groas Handle Smart Bidding Strategy Selection Automatically?

Yes. groas eliminates the manual work of auditing conversion volume, selecting strategies, setting targets, and monitoring learning phases. For DFY clients, a dedicated strategist backed by a proprietary engine trained on over $500 billion in profitable ad spend owns bidding strategy end-to-end. The engine identifies the right strategy and target for each campaign based on real-time data and adjusts targets incrementally without resetting learning phases. For DWY clients, your team stays in control while the engine and a strategist provide recommendations. No onboarding fees, month-to-month commitment, cancel anytime.

What Is The 20% Rule For Google Ads Smart Bidding Changes?

The 20% rule means that any single change to your tCPA target, tROAS target, or daily budget should not exceed 20% at a time. Larger changes reset the learning phase, which wipes out the algorithm's accumulated optimization signal and introduces days or weeks of volatile performance. Make one incremental adjustment of no more than 20%, wait five to seven days for the data to stabilize, measure the impact, then make the next adjustment if needed. This disciplined approach compounds results over time instead of creating constant learning resets.

Is Enhanced CPC Still Worth Using In 2026?

Enhanced CPC is a bridge strategy for accounts transitioning from fully manual bidding to automated strategies. It adjusts your manual bids up or down based on conversion likelihood while keeping you in partial control. It is not a long-term solution. Once your account has enough conversion data (30 or more per month), you should move to Maximize Conversions or Target CPA, which give Google's algorithm full control over bidding and consistently outperform eCPC at scale. Use eCPC only as a temporary stepping stone, not a permanent strategy.

How Does groas Compare To Managing Smart Bidding In-House?

Managing smart bidding in-house means one person monitoring learning phases, calculating targets from historical data, making incremental adjustments on schedule, and catching runaway spend before budgets burn. That person is capped at what they can physically get through in a week. groas puts a senior strategist on top of an engine trained on hundreds of billions in ad spend that runs 24/7. It detects patterns across thousands of accounts, identifies optimal targets, and adjusts without resetting learning, all with $0 onboarding and no long-term contracts. The gap in execution quality shows up in the numbers within the first few weeks.

Should I Use Portfolio Bid Strategies Across Multiple Campaigns?

Portfolio bid strategies are valuable when individual campaigns fall below conversion thresholds on their own but share similar audiences or products. Pooling conversion data across campaigns gives the algorithm a larger learning set, which improves bid accuracy without requiring campaign restructuring. Use portfolio strategies when you have three or more related campaigns that individually generate fewer than 30 conversions per month but collectively exceed that threshold. Do not pool campaigns with fundamentally different intent signals (like brand and non-brand campaigns), as this muddies the data the algorithm uses to bid.