How to Choose and Optimise Performance Max Bidding for Ecommerce in 2026

If you’re running Performance Max for an ecommerce brand in Melbourne and you’re not sure whether your bid strategy is helping or hurting, the answer in most accounts I look at is hurting. The default advice from Google, from the industry blogs, and from most agencies starts with Target ROAS as the goal and works backwards. That sequence is wrong for most accounts, and the reasoning behind why it’s wrong is rarely discussed honestly.

This guide walks through how I actually choose bid strategies for ecommerce PMax campaigns, what conversion volume and product value signals drive the decision, why I don’t use Target ROAS at all in my own accounts, and how I control ROAS through budget pacing instead. It assumes you’ve launched PMax before, know what an asset group is, and understand the difference between Maximize Conversions and Maximize Conversion Value at a surface level. For the campaign-structure side, see the guide to Performance Max campaign structure. For audience signals and creative setup, the PMax Asset Groups and Creative guide handles that. And for the underlying feed and campaign setup on Shopify, see the Shopify Google Ads & Shopping setup guide. This one is purely about bidding decisions and the budget mechanics that actually move profit. If you would rather hand this over, that is what my ecommerce PPC management covers.

The PMax bid strategy menu has four options, and the labels are changing in 2026

Performance Max gives you four Smart Bidding options, organised under two parent strategies. The June 2026 label change rolling out across Google Ads collapsed the names but kept the underlying behaviour identical. Worth knowing so you don’t get confused when you see different naming in different parts of the interface.

Maximize Conversions spends your daily budget to acquire as many conversions as possible, treating each conversion as equally valuable. You can attach an optional Target CPA, which Google now labels “Target CPA” as of June 2026.

Maximize Conversion Value spends your daily budget to acquire the highest total conversion value, weighting bids based on the value signals you send through conversion tracking. You can attach an optional Target ROAS, which Google now labels “Target ROAS” rather than “Maximize Conversion Value with Target ROAS.”

The distinction that matters operationally is not the four-way menu. It’s the two-way decision between volume-optimised bidding (Max Conversions) and value-optimised bidding (Max Conversion Value), with the optional targets sitting on top as constraints.

The other change worth flagging: starting 17 August 2026, Google is updating how its bidding systems handle campaigns limited by budget, and the company has warned that Target CPA and Target ROAS campaigns may see temporary performance fluctuations. A new Bid Target Adjustment Tool launched on 6 July 2026 to help advertisers manage the transition. If you’re running tROAS or tCPA campaigns through that update window, expect noise and don’t react to short-term swings.

The fundamental decision: Maximize Conversions vs Maximize Conversion Value

The industry default for ecommerce is to start with Maximize Conversion Value because value signals are the whole point of ecommerce optimisation. That advice is correct in principle and wrong in practice for most Melbourne accounts I work with, because the data threshold to make Maximize Conversion Value work properly is significantly higher than the threshold for Maximize Conversions.

Maximize Conversions learns inside three to seven days. The signal is binary: did the conversion happen or not. The algorithm can calibrate quickly because the data it’s working with is high-frequency and unambiguous.

Maximize Conversion Value takes two to four weeks. The signal is continuous: how much value did this conversion produce. The variance is higher because conversion values differ by order, by product mix, by promotion timing, by customer segment. The algorithm needs more data points to learn what a high-value conversion looks like and how to bid for it.

Industry-pattern observation from Smarter Ecommerce’s analysis of over 14,000 PMax campaigns: under 30 conversions per month, value-based bidding produces inconsistent results and frequently underperforms targets. At 60 to 90 conversions per month, hitting Target ROAS becomes roughly a coin flip. Only at 150-plus conversions per month does Target ROAS reliably meet or exceed its goal.

For most Melbourne ecommerce accounts I see, monthly conversion volume sits somewhere between 30 and 100. That puts them squarely in the range where value-based bidding is technically eligible but practically unreliable. Maximize Conversions, with a much lower data threshold, produces more stable performance and faster learning at that volume range.

My rule of thumb: if the account has some conversion history but isn’t generating more than 100 to 150 conversions per month in the campaign you’re optimising, start with Maximize Conversions. Switch to Maximize Conversion Value only when the data supports it.

The two gates: conversion volume AND average product value

Conversion count alone doesn’t decide the bid strategy. Average product value matters equally, and it’s the variable most operator guides skip.

Think about why. If your average order value is $30 and you’re targeting 50 conversions per month, your total monthly conversion value is $1,500. That’s enough data points for Maximize Conversions to learn well, but the value variance per conversion is small, so the upside of switching to value-based bidding is also small. The algorithm can’t optimise much harder for a $35 order versus a $25 order because both fall inside a narrow band.

Now flip it. If your AOV is $300 and you’re targeting the same 50 conversions per month, your monthly conversion value is $15,000. The value variance per conversion is much larger, often $100 versus $700, and value-based bidding has meaningful room to optimise. Higher AOV makes the value signal more useful even at the same conversion count.

The practical decision matrix for Melbourne ecommerce accounts:

This is operator judgment, not a formula Google publishes anywhere. Use it as a starting framework and adjust based on what you actually observe in the account.

Maximize Conversions is the accessible default for accounts with some conversion history

For most ecommerce accounts I onboard in Melbourne, Maximize Conversions is where I start, and it’s where I stay for longer than the agency industry tends to recommend. The reasoning:

Lower data threshold for stable learning. As covered above, Max Conversions can produce predictable performance at 30 conversions per month, while MCV often can’t.

Faster learning phase. Three to seven days for Max Conversions versus two to four weeks for MCV. For Australian SMB ecommerce accounts where the operator can’t afford a month of erratic performance during learning, the shorter learning phase is a material advantage.

Cleaner attribution signal. Max Conversions optimises against the binary conversion event, so the algorithm isn’t trying to learn from noisy value data while it’s also trying to learn what audience signals matter. Strip out a layer of complexity and the algorithm learns faster.

Works without value tracking maturity. Plenty of Melbourne ecommerce stores I look at have conversion tracking that captures the order but doesn’t pass back accurate revenue values. Maybe the value is missing for first-time customers, or it’s reporting the gross sale instead of the net after discount codes, or it’s the AOV midpoint rather than the actual order. Max Conversions doesn’t care. MCV does.

The downside is real and worth naming. Max Conversions treats a $50 order and a $500 order identically, so the algorithm can over-allocate spend to lower-value conversions if your product range is varied. Compensate by structuring your campaigns by margin tier and AOV band, which the campaign structure guide walks through.

Maximize Conversion Value when the data threshold is genuinely met

I do use MCV. The threshold I look for before switching:

• 100-plus conversions per month, consistently, for at least two months
• Conversion tracking confirmed accurate (value passing correctly, no duplicate firing, no missing customer types)
• AOV variance wide enough that value-based optimisation has room to add lift (typically AOV ranging across at least a 3x band)
• No major upcoming changes to the product mix, pricing, or attribution model that would invalidate the learning

When those conditions are met, MCV does add value. The algorithm bids more aggressively for users whose purchase signal correlates with higher value, and it pulls back on users whose signals suggest lower-value conversions. The lift over Max Conversions at that data volume is real, typically 10 to 20 per cent on revenue at similar spend.

Google’s official eligibility criteria for MCV in Demand Gen, by way of comparison, requires at least 50 conversions with value in the past 35 days. PMax doesn’t enforce this minimum at the campaign level, but the underlying machine learning logic is similar. Don’t switch to MCV just because you’re allowed to. Switch when the data volume actually supports it.

Why I don’t recommend Target ROAS, and what I use instead

This is the position that diverges most sharply from typical PMax guidance, and it’s worth explaining the reasoning carefully because it sounds counterintuitive if you’ve absorbed the standard Google playbook.

Reason one: Target ROAS makes Google reduce your spend to hit the target. Google’s own documentation says this plainly. Maximize Conversion Value is designed to spend your daily budget. Target ROAS is designed to hit the target ROAS while spending as much as it can within that constraint. If your target is set above where the auction can comfortably deliver, Google reduces volume to protect the target. You end up with a campaign that hits its ROAS goal but spends 60 per cent of its budget. That’s a volume problem dressed up as a performance success.

Reason two: ROAS is attributed and not 100 per cent accurate. Performance Max attributes conversions across Search, Shopping, Display, YouTube, Discover, and Gmail inventory. The attribution windows and credit allocation across those channels are decided by Google’s model, which has known incrementality issues. PMax campaigns regularly claim credit for branded search clicks, retargeting clicks, and other interactions that would have produced revenue regardless. Tuning the campaign to hit a specific ROAS number when that number is partially fictional means you’re optimising against a goal that doesn’t fully reflect business reality. Foundry CRO’s 2026 ROAS benchmark research noted explicitly that PMax’s reported ROAS is often inflated by attribution claims that wouldn’t survive incrementality testing.

Reason three: ROAS targets create unstable optimisation cycles. Setting Target ROAS too high constrains volume. Setting it too low produces wasteful spend on low-value conversions. The sweet spot is a moving target as the auction environment, your product mix, and your conversion tracking evolve. Operators end up adjusting ROAS targets every few weeks, each adjustment resetting the learning phase. The campaign never genuinely stabilises.

Reason four: budget is a more honest lever for ROAS control. What Target ROAS actually does is restrict spend to protect efficiency. Adjusting budget does the same thing more directly, more predictably, and without the attribution noise. Cut the budget, the campaign spends less, the cherry-picked highest-intent conversions get priority, and the effective ROAS rises. Increase the budget, the campaign spends more, it reaches further down the auction quality curve, and the effective ROAS falls. The lever is direct and the cause-effect relationship is clean.

The position is contrarian. It’s also operator-defensible. Some industry sources do support it. Sarah Stemen’s budget pacing guide notes that constrained bidding strategies like tROAS will often underspend if targets are too aggressive. PPC Hero’s pacing guide describes tROAS-driven underdelivery as one of the most common causes of misdiagnosed campaigns. The Optmyzr-cited PMax expert Jyll Saskin Gales runs Maximize Conversions for new campaigns explicitly because the bid strategy decision is volume-gated, not theory-gated.

What I use instead is straightforward: Maximize Conversions or Maximize Conversion Value without a target, with the daily budget set deliberately to produce the effective ROAS the account needs.

How to set your budget to control effective ROAS

If you’re going to use budget pacing as your ROAS lever instead of a tROAS target, the budget calculation matters. The framework I use:

Step one: identify the conversion floor. PMax needs roughly 30 to 50 conversions per month per campaign to optimise effectively. If you’re spending less than that produces, the campaign will struggle regardless of bid strategy. The industry-pattern threshold from operator guides and platform-side analysis sits at $3,000 AUD per month for most ecommerce accounts, though the actual floor varies by AOV and conversion rate.

Step two: set the daily budget to at least 3x your target CPA. This is the industry “3x CPA rule” that surfaces in most PMax operator guides. If your target CPA is $30 AUD, your daily budget should be at least $90 AUD, which gives the algorithm headroom to explore and learn without getting starved.

Step three: observe the natural ROAS the campaign produces at that budget. Run for two to three weeks without changing anything. Note the average ROAS, the daily ROAS variance, and the spend efficiency curve. PMax campaigns typically fluctuate 20 to 30 per cent day-over-day, so look at week-on-week averages not single-day numbers.

Step four: adjust budget up or down based on profit math. If the campaign is producing ROAS above your break-even, increase the budget incrementally to capture more volume. If it’s producing ROAS below break-even, reduce the budget incrementally to force the algorithm onto higher-intent inventory. The increment matters: never adjust budget by more than 15 to 20 per cent in a single change, because larger swings reset the learning phase and the noise from the reset will obscure the signal you’re trying to read.

Step five: hold the budget stable for at least two weeks per adjustment. Stable budget creates stable learning. Stable learning creates predictable performance. Every change resets the clock, so the operators who make too many adjustments end up with campaigns that never genuinely settle into their optimal state.

For Melbourne ecommerce accounts at common AUD spend tiers, the budget pacing patterns look like this:

These ratios are approximate and depend heavily on category, AOV, and conversion rate. They illustrate the operator framing: budget is the lever, conversion volume is the gating variable for bid strategy, and CPA emerges from the interaction.

Target CPA for PMax is mostly lead-gen territory

Target CPA exists for PMax and works the same way as Target ROAS but with a per-conversion cost target instead of a return target. For ecommerce, I rarely use it. The reason is the same as the tROAS critique: tCPA restricts volume to protect efficiency, which means you’re paying the algorithm to underspend.

Where tCPA genuinely earns its place is in lead-generation campaigns where the conversion isn’t a purchase. A free consultation booking, a downloaded resource, a demo request. The value variance per conversion is low, conversion volume is typically higher than ecommerce, and the operator usually has a specific cost-per-lead they’re underwriting. In that environment tCPA makes more sense than for ecommerce, where AOV and order economics make a value-based approach more appropriate when the data supports it.

For ecommerce specifically: if you find yourself reaching for tCPA, you probably want Max Conversions without the target. Same volume optimisation, less algorithmic restriction.

Bid strategy transitions are triggered by volume and stability, not by the calendar

A lot of PPC guides recommend transitioning from Max Conversions to MCV or to a target-based strategy after six weeks, or eight weeks, or some other time-based threshold. Time alone doesn’t tell you whether the campaign is ready.

The criteria I look for before transitioning:

Volume gate. At least 30 conversions in the most recent 30-day window. Below that threshold, neither MCV nor tROAS will work reliably.

Stability gate. Daily ROAS variance settling into a predictable band, typically plus or minus 25 per cent of the rolling 30-day average. If daily ROAS is still swinging wildly week to week, the campaign isn’t stable enough for a target-based strategy to add value.

Attribution confidence gate. Conversion tracking confirmed accurate, no missing values, no duplicate firing. If you’re going to optimise to value, the values better be right.

Margin tier consistency. The campaign’s product mix isn’t about to change. Switching bid strategies right before a new product launch or a promotional period guarantees the learning phase will collide with the change.

When all four gates are open, the transition makes sense. When any of them are closed, hold off. Time on the calendar doesn’t override these gates.

The actual transition mechanics: change the bid strategy in the campaign settings, expect two to four weeks of learning where performance drops 30 to 50 per cent below pre-change levels, and don’t make any other changes during that window. Plan for the dip, don’t react to it.

Bid strategy and audience signals work as a system, not as independent levers

The bid strategy decides how the algorithm spends. The audience signals decide where the algorithm looks first. Both matter, and they interact in ways that aren’t obvious until you’ve watched the same account through multiple bid strategy changes.

When you’re on Max Conversions, audience signals influence early exploration more strongly because the algorithm hasn’t yet learned which value-conversion correlations matter. Strong audience signals (your Customer Match list, your highest-intent custom segments) compress the learning phase and push the algorithm toward higher-quality inventory faster.

When you’re on Maximize Conversion Value, audience signals matter less in the steady state because the algorithm has built its own model of who’s likely to produce high-value conversions. They still matter during learning, but the relative weight shifts toward the value model the algorithm has developed.

When you’re on Target ROAS, audience signals matter as much as on MCV, but the algorithm is more constrained in how aggressively it can explore beyond them because the ROAS target restricts inventory expansion. If you run a target ROAS, a sale raises the break-even it has to clear, so work through the discount maths before BFCM.

Practical implication: feed PMax good audience signals regardless of bid strategy, and feed them more aggressively when you’re on Max Conversions (because the algorithm leans harder on them). The PMax Asset Groups and Creative guide covers how to construct the audience signal stack.

Why I leave New Customer Acquisition off

Google introduced the New Customer Acquisition (NCA) goal in Performance Max several years ago, expanded it through 2025 and 2026, and aggressively promotes it as a way to bid more competitively for net-new customers. The case studies Google publishes (Dime Beauty, Luxury Escapes) report significant lifts. I explain the test and what I do instead in new customer acquisition in Performance Max.

I don’t use it. Three reasons:

Reason one: the 540-day “new customer” definition has structural selection bias. Google’s NCA logic considers anyone who hasn’t purchased in 540 days as a new customer. That window catches a lot of users who weren’t actually acquired by the ad. They were already aware of your brand, they came back through search or direct, and PMax got credit because it served them an ad somewhere in the path. The “incremental new customer” the feature reports is often not incremental at all.

Reason two: the conversion value inflation distorts reporting. When you set NCA to bid higher for new customers (the recommended mode), Google reports the inflated conversion value in the Google Ads interface. Your reported revenue from PMax goes up. Your actual revenue does not. This is a reporting integrity issue if you’re using Google Ads numbers to make business decisions.

Reason three: the case studies aren’t incrementality-tested. Google’s published lifts are pre-versus-post comparisons or campaign-versus-campaign splits, not proper incrementality tests with holdout groups. Operator-side analysis (ZATO Marketing’s writeup, Adchieve’s research, and several independent agency tests) has consistently flagged that the apparent NCA lift partially disappears when subjected to proper testing.

The honest position: NCA might add some genuine incremental value for some accounts, but Google has not published evidence that survives a proper incrementality framework, and the cost of running it without that evidence is reporting noise that makes other optimisation decisions harder. For Melbourne ecommerce accounts where the operator needs clean data to make budget decisions, I leave it off.

If you do want to test NCA in your account, do it as a campaign experiment with proper measurement, not as a default-on setting. And read the actual reporting carefully: if conversion value is being inflated for new customers, separate the reported and actual numbers before drawing conclusions.

Budget pacing patterns for Melbourne ecommerce accounts

Heavy AU-specific framing for this section because budget pacing decisions are deeply contextual to the account’s spend tier, and Melbourne SMB ecommerce sits in a specific range that the US-focused PMax guidance doesn’t cover well.

The under-$3K AUD per month tier. Most ecommerce accounts spending less than $3,000 AUD per month don’t have enough volume to make PMax work as the primary campaign type. The 30-conversion floor isn’t being met, the learning phase doesn’t stabilise, and the algorithm can’t differentiate between auction opportunities. Recommendation at this spend tier: run Standard Shopping as the primary campaign, supplement with targeted Search for high-intent commercial terms, and consider PMax only once the account scales past $3K AUD monthly with sustained conversion volume.

The $3K to $6K AUD per month tier. This is where PMax starts to be viable for Melbourne ecommerce. Run one PMax campaign, Max Conversions bid strategy, daily budget around $100 to $200 AUD. Don’t segment into multiple campaigns at this spend tier because the volume per campaign would fall below the optimisation floor. Hold the budget stable for at least 30 days before judging performance.

The $6K to $15K AUD per month tier. Now you have room to segment by margin tier. Two PMax campaigns: hero products in one (higher AOV, better margins), standard catalogue in the other. Both on Max Conversions if the per-campaign conversion count is under 100 per month. Switch the hero campaign to MCV first when the volume permits, because it’ll have higher AOV variance and more upside from value-based optimisation.

The $15K AUD per month and up tier. Three to four PMax campaigns segmented by margin tier and possibly by customer segment. MCV becomes the default bid strategy for the high-volume campaigns. Use budget pacing to control effective ROAS across the portfolio: if you need overall ROAS to lift, reduce budget on the lower-margin campaigns first while protecting hero spend. The 70-20-10 budget allocation framework (70 per cent to proven performers, 20 per cent to scaling candidates, 10 per cent to testing) works well at this spend tier.

The Melbourne-specific operator note: AUD spend tiers behave slightly differently to USD equivalents because the AU search inventory is thinner. The auction is less crowded at the entry tiers, which can produce higher CPMs paradoxically because Google’s algorithm has fewer auction signals to work with. Budget pacing tends to be more volatile in the early weeks of a campaign than in larger US markets. Expect day-to-day swings of 30 to 50 per cent in the first month, settling toward the 20 to 30 per cent range as the algorithm builds inventory knowledge.

Bid strategy interaction with Search and Shopping cannibalisation

If you’re running PMax alongside Standard Shopping or Search campaigns (which most Melbourne ecommerce accounts should be doing for portfolio reasons), the bid strategy choices interact across campaign types.

PMax has a known tendency to claim credit for branded search clicks, especially in accounts that don’t apply brand exclusions. When PMax is on Maximize Conversion Value or Target ROAS, the value-weighted bidding amplifies this behaviour: branded clicks convert at high rates, so the algorithm bids aggressively for them even when they would have converted through your standalone brand Search campaign.

The mitigation, regardless of bid strategy: apply brand exclusions to your PMax campaigns. Google added campaign-level brand exclusions in 2023 and expanded them through 2024 and 2025. They now support up to 1,000 entries per account and apply to both Search and Shopping inventory inside PMax. Without them, your PMax ROAS is inflated by cannibalised brand traffic, and your operator decisions get distorted.

The bid strategy interaction with brand exclusions: when you apply exclusions and remove the brand traffic from PMax, the campaign’s reported ROAS will drop. This is correct. The campaign was previously taking credit for revenue it didn’t generate. The new, lower ROAS is the actual non-brand performance. Use that number for budget decisions, not the inflated pre-exclusion number.

The same principle applies to retargeting inventory. If PMax is serving Display ads to people who already visited your product pages, the conversions it claims are partially incremental and partially not. There’s no clean exclusion lever for this in PMax, but you can monitor it through the channel performance report and adjust audience signals to reduce the weight on existing visitors.

Common bidding mistakes in Melbourne PMax accounts

Pattern-matched failures I see consistently across SMB ecommerce accounts:

Setting Target ROAS too high at launch. A new campaign with no conversion history gets handed a 600 per cent ROAS target because that’s the operator’s profitability goal. The algorithm has no data to bid against, so it spends nothing, generates no conversions, and the operator declares PMax broken. Start without a target.

Adjusting Target ROAS or Target CPA too frequently. Google’s recommendation is no more than 20 per cent change every 14 days. Operators routinely adjust weekly. Every change resets learning, the campaign never stabilises, and performance trends are misread as bid-strategy failures when they’re actually learning-phase artefacts.

Switching bid strategy too early. Switching to MCV or tROAS at 15 conversions per month because Google says it’s the minimum. The Google minimum is a permission gate, not a recommendation. Wait for stable volume above 50 conversions per month before switching.

Treating “Limited by Budget” as an alert. For Max Conversions and MCV campaigns, “Limited by Budget” is the design state, not an error. The bid strategy is supposed to spend the full daily budget. The warning is misleading and should be ignored unless you’ve decided you want to add budget.

Pausing campaigns when CPA spikes. PMax CPA fluctuates significantly week to week because the algorithm is exploring inventory. A single bad week of CPA is not a reason to pause or restructure. Look at four-week rolling averages, not single-day or single-week numbers.

Conflating bid strategy with the actual performance lever. The bid strategy is one decision. Budget, audience signals, asset quality, conversion tracking accuracy, and campaign structure are equally important. Operators who obsess about bid strategy while ignoring the other levers consistently underperform operators who use a clean bid strategy correctly with everything else dialled in.

Reporting limitations for bid strategy decisions

The Performance Max reporting layer has improved substantially through 2024 and 2025 but still doesn’t give you what you need to make bid strategy decisions cleanly.

What you can see: channel performance (Search vs Shopping vs Display vs YouTube), conversion volume and value, asset performance scores (Best/Good/Low), search term categories (partial), and audience signal insights.

What you can’t see directly: which audience signal produced which conversion, which placement type within each channel converted (e.g. Discover vs Gmail within the Display bucket), how Smart Bidding allocated spend across asset groups, and the true incremental lift of any optimisation.

This matters for bid strategy decisions because the right strategy depends partly on understanding where your conversions are coming from. If 80 per cent of your PMax conversions are coming from branded Shopping queries that would have converted through your brand Search campaign anyway, switching to value-based bidding will optimise harder for traffic that’s already converting. The “lift” you measure isn’t incremental.

Operator workaround: triangulate with backend revenue data, GA4 attribution paths, and incrementality testing where possible. Don’t make bid strategy decisions purely from the Google Ads interface, because the interface tells you what PMax thinks it did, not what actually happened in the business.

The senior operator workflow for ongoing bid strategy management

A practical workflow that captures the principles above:

Weekly review (30 minutes per campaign): Check 7-day vs 30-day rolling ROAS and CPA. Note any week-on-week deltas larger than 25 per cent. Check the channel performance distribution. Verify the conversion volume is on track for the monthly target. Don’t change anything yet unless something is genuinely broken (tracking issue, account suspension risk, major delivery anomaly).

Fortnightly review (60 minutes per campaign): Decide whether the budget needs adjustment. Look at the conversion volume curve over the past four weeks: is it stable, growing, or declining? If conversion volume is stable and ROAS is healthy, hold. If ROAS is above target and volume is constrained, increase budget by 15 to 20 per cent. If ROAS is below target and volume is sufficient, decrease budget by 15 to 20 per cent. Hold for two weeks before next adjustment.

Monthly review (2 hours per campaign): Evaluate whether the bid strategy still fits. Has conversion volume crossed the threshold for switching from Max Conversions to MCV? Has the AOV mix shifted enough to change the calculus? Are audience signals still aligned to the strategy? Make at most one structural change per monthly cycle.

Quarterly review (4 hours per campaign): Step back. Compare current performance to the previous quarter. Check whether the account-level strategy still makes sense. Has the business changed direction? Is PMax still the right campaign type for this product mix? Are there structural changes (additional campaigns, consolidation, channel shifts) that the data now supports?

The discipline that separates senior operators from agency-default management: making fewer changes, holding for longer, and using budget as the primary lever rather than tinkering with bid strategy settings every week. That cadence is how I run the ecommerce PPC service for Melbourne stores.

FAQ

Should I start PMax campaigns with Maximize Conversion Value or Target ROAS?

Neither, for most ecommerce accounts in Melbourne. Start with Maximize Conversions if the account has any conversion history at all, because the data threshold to make Max Conversions stable (around 30 conversions per month) is much lower than for Maximize Conversion Value (around 100 conversions per month) or Target ROAS (around 150 conversions per month for reliable target adherence). Switch to value-based bidding only when the data volume genuinely supports it.

How do I control ROAS in Performance Max without using Target ROAS?

Use budget as the lever. Lower the daily budget to force the algorithm onto higher-intent inventory and higher effective ROAS. Raise the budget to capture more volume at lower effective ROAS. The relationship is direct and predictable, unlike Target ROAS which restricts volume in unpredictable ways while also being measured against potentially inflated attribution numbers. Adjust budget no more than 15 to 20 per cent per change, hold for at least two weeks between adjustments.

Is Target ROAS ever the right choice for ecommerce PMax?

For very high-volume accounts (150+ conversions per month per campaign, sustained for at least two months) with mature conversion value tracking and a clearly understood break-even ROAS, Target ROAS can work. Even then, I’d recommend running a controlled experiment comparing Max Conversion Value without a target against MCV with a target ROAS before committing, because in many cases the untargeted version produces better total revenue at acceptable efficiency.

How long should I wait before switching from Maximize Conversions to Maximize Conversion Value?

Wait for four gates to be open: 30-plus conversions per month sustained for at least two months, conversion tracking confirmed accurate with value passing reliably, AOV variance wide enough to justify value-based optimisation (3x range across orders), and no major product or pricing changes pending. Time on the calendar doesn’t override these gates.

Should I enable New Customer Acquisition in my Performance Max campaigns?

I leave it off. The 540-day new-customer definition has structural selection bias, the conversion value inflation distorts the reporting in the Google Ads interface, and the case studies Google publishes aren’t incrementality-tested. If you want to test it, do it as a campaign experiment with proper measurement and a holdout group, not as a default-on setting.

What budget tier do I need to run Performance Max effectively for ecommerce in Melbourne?

The practical floor is around $3,000 AUD per month per campaign. Below that, you typically don’t generate the 30 conversions per month minimum for stable PMax learning. At $3,000 to $6,000 AUD per month, run one PMax campaign. At $6,000 to $15,000 AUD per month, segment by margin tier into two campaigns. At $15,000 AUD per month and above, three to four campaigns with structural segmentation make sense.

What’s the difference between Maximize Conversions and Target CPA?

Mechanically, Target CPA is now just Maximize Conversions with an optional Target CPA value attached. Google’s June 2026 label change collapsed the names but the underlying behaviour stayed the same. Max Conversions without a target spends your full daily budget to acquire as many conversions as possible. Max Conversions with a Target CPA spends your daily budget while trying to keep the average CPA at or below the target, which often results in underspending if the target is set too aggressively.

How often should I adjust my PMax budget?

No more than once per two-week cycle, with adjustments of no more than 15 to 20 per cent per change. Larger or more frequent changes reset the learning phase and the resulting noise obscures the signal you’re trying to read. Stable budget creates stable learning. Operators who change budget weekly almost always underperform operators who hold for longer.