Most Google Ads accounts we audit have the wrong bidding strategy running—not because someone made a careless choice, but because they selected whatever Google recommended during setup and never reviewed it again. Many businesses also rely on a performance marketing agency to optimize bidding strategies and improve campaign performance over time. Bidding strategy is one of the few levers in a modern, largely automated Google Ads account that still requires a genuine human decision, and getting it wrong can quietly limit results for months before anyone notices. This guide explains what Target CPA, Target ROAS, and Maximize Conversions actually optimize for, when each strategy works best, and the common mistakes that cause even the right bidding strategy to underperform.

Why Bidding Strategy Still Matters in an Automated Account

A common misconception is that once Smart Bidding is switched on, the strategy chosen barely matters because “the algorithm handles it anyway.” That’s a costly assumption. Smart Bidding strategies aren’t interchangeable settings — each one tells Google’s machine-learning model a completely different objective to chase, and the model will genuinely optimize toward whatever goal it’s given, even when that goal doesn’t match what the business actually wants.

An account set to Maximize Conversions with no target will happily spend the entire daily budget chasing volume, even if the cost per conversion climbs well past what’s profitable. An account set to an unrealistically tight Target CPA will throttle delivery down to a trickle, because the algorithm won’t bid into auctions it predicts will miss the target. Neither is a flaw in the system — both are the system doing exactly what it was told, which is precisely why the choice of strategy, and the number attached to it, deserves more attention than most accounts give it.

What Is Maximize Conversions Bidding?

Maximize Conversions is Google’s most straightforward automated bidding strategy: it sets bids automatically, within the daily budget, to get as many conversions as possible without a specific cost target attached. There’s no CPA or ROAS ceiling guiding the algorithm — the only constraint is the budget itself.

When Maximize Conversions Makes Sense

This strategy works best for accounts still building conversion history, where the priority is generating enough data for Smart Bidding to learn from rather than protecting a specific cost-per-acquisition from day one. A brand-new campaign, or one entering a new market with no prior performance data, often starts here deliberately, accepting a less efficient early period in exchange for the conversion volume needed to make Target CPA or Target ROAS viable later.

Where Maximize Conversions Goes Wrong

The risk with this strategy is that “as many conversions as possible” says nothing about whether those conversions are profitable. An account can hit a personal-best conversion count while quietly bleeding money on a cost-per-acquisition that’s climbed past what any of those conversions are actually worth. Businesses running Maximize Conversions long-term without a ceiling on acceptable cost are usually the same accounts surprised, months later, by how thin their margins have become.

What Is Target CPA Bidding?

Target CPA (Cost Per Acquisition) tells Google’s algorithm to get as many conversions as possible while trying to hold the average cost per conversion at or near a specific number the advertiser sets. Unlike Maximize Conversions, this strategy introduces a real cost constraint the algorithm has to respect.

How Target CPA Actually Behaves

Google doesn’t guarantee every single conversion costs exactly the target — some will cost more, some less — but the algorithm works to keep the average across the campaign close to that number. Setting the target too aggressively below the account’s realistic historical average is one of the most common mistakes advertisers make: rather than forcing efficiency, an unrealistic target usually just restricts how many auctions the algorithm is willing to enter, shrinking volume sharply.

When Target CPA Makes Sense

Lead-generation businesses — where every conversion (a form fill, a call, a demo request) carries roughly the same value regardless of which keyword or audience produced it — are the clearest fit for Target CPA. If a real estate consultancy values every qualified lead at approximately the same amount, optimizing toward a consistent cost per lead is a more direct match for the business goal than optimizing toward conversion volume alone.

Setting a Realistic Target CPA

The safest starting point is the account’s trailing 30–60 day average cost per conversion, adjusted slightly based on what margin the business can actually sustain — not an aspirational number pulled from a competitor’s case study or an executive’s target for the quarter. Once the campaign stabilizes at that realistic number, gradual adjustments downward, watched carefully over several weeks, work far better than a single aggressive cut.

What Is Target ROAS Bidding?

Target ROAS (Return on Ad Spend) tells the algorithm to optimize for conversion value relative to spend, expressed as a percentage — a Target ROAS of 400% means the goal is generating ₹4 in conversion value for every ₹1 spent. This is a fundamentally different optimization goal from Target CPA, because it accounts for the fact that not all conversions are worth the same amount.

Why ROAS Matters More Than CPA for Certain Businesses

An e-commerce store selling products ranging from ₹500 accessories to ₹15,000 appliances has conversions with wildly different values. Optimizing purely toward a low cost-per-conversion, the way Target CPA does, could push the algorithm toward cheap, low-value purchases while ignoring the more valuable ones. Target ROAS corrects for this by weighing conversion value directly, which is why it’s the standard choice for e-commerce and any business with meaningfully variable order or deal sizes.

The Data Requirement Target ROAS Needs

Target ROAS depends entirely on accurate conversion value tracking, not just conversion counts — this means dynamic value passback from a Shopping feed or CRM, not a flat, static value assigned to every conversion regardless of size. An account that hasn’t set up value tracking properly will see Target ROAS perform poorly, not because the strategy is wrong for the business, but because the algorithm is optimizing against inaccurate or missing data.

Setting a Realistic Target ROAS

As with Target CPA, the account’s own historical blended ROAS is the right starting point, not an ambitious number set by finance without reference to what the account has actually delivered. A target set meaningfully above historical performance usually restricts delivery rather than lifting efficiency, for the same reason an unrealistic Target CPA does — the algorithm simply won’t enter auctions it predicts will miss the number.

Target CPA vs. Target ROAS vs. Maximize Conversions: A Direct Comparison

• Optimization goal: Maximize Conversions chases volume with no cost ceiling; Target CPA chases volume while holding average cost near a set number; Target ROAS chases conversion value relative to spend.
• Best fit: Maximize Conversions suits new campaigns building data; Target CPA suits lead-gen businesses where conversions carry similar value; Target ROAS suits e-commerce and variable-value conversions.
• Data requirement: Maximize Conversions needs basic conversion tracking; Target CPA needs consistent conversion tracking; Target ROAS needs accurate, dynamic conversion value tracking.
• Risk if misconfigured: Maximize Conversions can overspend on low-value conversions; an unrealistic Target CPA throttles volume; an unrealistic Target ROAS restricts delivery just as sharply.
• Typical account stage: Maximize Conversions early, before enough data exists; Target CPA or Target ROAS once 30–50 conversions (Google’s commonly cited threshold) have accumulated for the algorithm to optimize against meaningfully.

Many mature accounts don’t pick just one — they run Maximize Conversions on newer campaigns still gathering data, Target CPA on lead-gen campaigns with consistent conversion value, and Target ROAS on e-commerce campaigns with variable order values, all within the same account.

How Much Conversion Data Do You Actually Need Before Switching?

One of the most common questions advertisers ask is when it’s safe to move off Maximize Conversions and onto a target-based strategy. Google commonly cites roughly 30–50 conversions within the prior 30 days as a rough threshold for a Smart Bidding strategy to have enough signal to optimize meaningfully — below that, the algorithm is working with too little data to reliably hit a specific cost or value target, and switching too early usually produces erratic, disappointing results that have more to do with insufficient data than a flawed strategy choice.

Accounts with low overall conversion volume — a high-ticket B2B service generating five qualified leads a month, for instance — may never comfortably clear that threshold on a single campaign. In these cases, consolidating conversion actions, extending the attribution window, or using a broader “micro-conversion” as the optimization signal (like a landing page scroll depth or a pricing page visit) can give Smart Bidding more frequent signal to learn from, while still keeping the ultimate business goal in view through separate reporting.

Common Mistakes That Undermine Every Bidding Strategy

Switching Strategies Too Frequently

Every time a bidding strategy changes, Google Ads treats it as a meaningful shift and re-enters a learning period, typically one to two weeks, during which performance is often less stable. Advertisers who switch strategies every time performance dips for a few days rarely give any single strategy a fair chance to prove itself.

Setting Targets From Aspiration Instead of Data

A Target CPA or Target ROAS chosen because it sounds like a good number for the business, rather than one grounded in the account’s actual historical performance, usually restricts delivery instead of improving it. The algorithm can’t out-optimize a target that doesn’t reflect the reality of the auction landscape it’s bidding into.

Poor Conversion Tracking Undermining the Whole Strategy

No bidding strategy can perform well against inaccurate data. Duplicate conversion tracking, missing value data for Target ROAS, or conversions counted from bot traffic all feed bad signal into an algorithm that has no way of knowing the data is flawed — it simply optimizes toward whatever it’s told is a conversion.

Ignoring Seasonal Adjustments

A Target CPA or ROAS set during a normal sales month may be completely unrealistic during a high-competition period like a festive sale season, when auction prices rise across the board. Google Ads offers seasonality adjustments specifically for these situations, and skipping them means the algorithm keeps chasing a target that no longer matches the current auction environment.

How to Choose the Right Strategy for Your Business

The decision usually comes down to three questions. First, does every conversion carry roughly the same value, or does value vary significantly by order size or deal type — the former points toward Target CPA, the latter toward Target ROAS. Second, does the account have enough historical conversion volume to set a realistic target, or is it still early enough that Maximize Conversions makes more sense while data accumulates. Third, is conversion value tracking accurate and dynamic, because Target ROAS without reliable value data will underperform regardless of how sound the strategy is in theory.

Getting this right isn’t a one-time decision either — as an account matures, conversion volume grows, and value tracking improves, the right strategy often shifts from Maximize Conversions to Target CPA to eventually Target ROAS, each stage building on cleaner data than the last.

Portfolio Bidding Strategies: Managing Multiple Campaigns Together

Beyond the single-campaign choice between Target CPA, Target ROAS, and Maximize Conversions, Google Ads also allows these strategies to run as portfolio bids — a shared strategy applied across multiple campaigns rather than one campaign optimizing in isolation.

Why Portfolio Bidding Changes the Calculation

A standard, single-campaign Target CPA optimizes only within that one campaign’s auctions. A portfolio Target CPA, applied across several related campaigns, lets Google’s algorithm shift budget and bidding aggressiveness between those campaigns based on which one is currently delivering conversions more efficiently — something a single-campaign strategy can’t do on its own.

When Portfolio Bidding Helps

Businesses running several campaigns targeting overlapping audiences or similar product categories — say, separate campaigns for different city locations of the same service business — often see more consistent overall performance from a shared portfolio strategy than from managing each campaign’s target independently. The algorithm effectively has a larger pool of auction data to learn from and more flexibility to move budget toward whichever campaign is currently converting best.

The Trade-Off Worth Knowing

Portfolio bidding sacrifices some campaign-level predictability for account-level efficiency. A campaign that’s part of a portfolio might see its individual spend fluctuate more than it would under its own standalone target, because the algorithm is balancing performance across the whole group rather than protecting each campaign’s individual number. For businesses that need to report on each campaign’s performance independently to different stakeholders or budget owners, this trade-off needs to be weighed carefully before combining campaigns into one portfolio strategy.

A Practical Example: Choosing the Right Strategy as an Account Matures

Consider a Faridabad-based B2B staffing agency running Google Ads for the first time, generating leads through a contact form with a single, consistent value regardless of which job role or industry the lead relates to. In month one, with zero historical conversion data, the campaign launches on Maximize Conversions with no target, deliberately accepting a higher early cost per lead in exchange for the conversion volume needed to build a usable dataset.

By month two, the campaign has accumulated roughly 45 conversions — close to the threshold where Smart Bidding has enough signal to work with a specific target. The account switches to Target CPA, set at the trailing 30-day average cost per lead rather than an aspirational number pulled from an industry benchmark. Volume dips slightly in the first two weeks after the switch, consistent with the typical learning period, before stabilizing at a cost per lead close to the target.

Six months in, the business expands into placing candidates for higher-value executive search roles alongside its standard staffing services — two service lines with meaningfully different deal values. At this point, a single Target CPA optimizing toward “cost per lead” regardless of which service line the lead relates to starts working against the business, since it has no way to distinguish a ₹2,000-value staffing lead from a ₹40,000-value executive search lead. The account restructures into separate campaigns by service line, assigns differentiated conversion values through the CRM, and migrates the executive search campaign to Target ROAS, while the standard staffing campaign — where lead values stay fairly consistent — remains on Target CPA.

This progression, from Maximize Conversions to Target CPA to a mixed Target CPA and Target ROAS setup, isn’t a sign of instability — it’s what a bidding strategy should look like as an account’s data, conversion value structure, and business complexity genuinely evolve over time. Businesses that pick one strategy at launch and never reconsider it as circumstances change are usually the ones leaving performance on the table months or years later.

Diagnosing an Underperforming Bidding Strategy

When a Smart Bidding strategy stops delivering results, the instinct is often to switch strategies immediately — but the actual cause is usually one of a handful of specific, diagnosable issues rather than a fundamentally wrong strategy choice.

Check Whether the Target Has Drifted From Reality

Auction dynamics shift over months as competitors enter or exit a category, seasonal demand changes, or the account’s own quality signals improve or decline. A Target CPA or ROAS set accurately six months ago can become unrealistic without a single deliberate change to the campaign — simply because the market around it moved. Comparing the current target against the account’s trailing 30-day actual performance is the fastest way to spot this kind of drift.

Look for Conversion Tracking Gaps Before Blaming the Algorithm

A sudden, unexplained performance drop is more often a broken conversion tag, a website update that disrupted tracking, or a duplicate conversion action inflating numbers than a genuine failure of the bidding strategy itself. Reviewing conversion tracking health should always come before adjusting a target or switching strategies, since a strategy can’t be judged fairly against broken data.

Rule Out External Budget or Landing Page Changes

A reduced daily budget, a paused campaign extension, or a landing page redesign that hurt conversion rate can all look like a bidding strategy problem when the actual cause sits elsewhere in the account or on the website. Isolating what specifically changed around the time performance shifted usually points to the real cause faster than assuming the algorithm itself has degraded.

Give Genuine Underperformance Time to Confirm Before Reacting

Even after ruling out tracking issues and target drift, normal day-to-day and week-to-week variance in a Smart Bidding campaign is wider than most advertisers expect. A dip over three or four days rarely justifies a strategy change; a consistent decline held over two to three full weeks, with tracking and targets both confirmed accurate, is the point where a genuine strategy reassessment is warranted.

Why Growthkul Gets This Right

A lot of accounts we take over are running whatever bidding strategy Google’s setup wizard suggested at launch, with a target CPA or ROAS number nobody has revisited since — often set months or years earlier, against auction conditions that no longer exist. That’s the gap between an account that’s technically “on Smart Bidding” and one that’s actually being managed.

Growthkul starts every account with a full audit of conversion tracking accuracy before touching bidding strategy at all — because no strategy, however well chosen, can perform well against inaccurate or incomplete conversion data. From there, targets are set from the account’s actual trailing performance, adjusted gradually rather than in aggressive swings, and revisited on a real cadence tied to seasonality and business changes, not left untouched for a year. Wasted spend identification — the exact keywords, placements, and ad sets quietly burning budget under a misconfigured target — is part of every audit, because that’s usually where the real inefficiency in an account hides, not in the choice of algorithm itself.

Conclusion

Target CPA, Target ROAS, and Maximize Conversions aren’t ranked from worst to best — they’re built for different stages of an account and different business models. A brand-new campaign, a lead-gen business with consistent conversion value, and an e-commerce store with wide-ranging order sizes each have a genuinely different right answer, and the mistake most accounts make isn’t picking the wrong algorithm outright, it’s picking a reasonable strategy and then setting an unrealistic number, or never revisiting it as the business and the auction landscape change.

The businesses that get consistent value from Smart Bidding treat the target number as a living decision tied to real account data, not a one-time setting configured at launch and forgotten. If it’s been more than a quarter since anyone looked at your bidding strategy and the number attached to it, that’s usually the first place worth checking before spending another rupee on a campaign that might be quietly capped by a target set for a version of the account that no longer exists.