Google Ads Bidding Strategy 2026: Match Strategy to Campaign Goal

Why most paid accounts pick the wrong bidding strategy

Paid accounts pick the wrong bidding strategy because the Google Ads UI defaults to Maximise Conversions on new campaigns and operators rarely override the default. The result is a Smart Bidding algorithm running on insufficient conversion data, optimising for the wrong signal, burning budget. Bidding strategy mismatched to goal is the single biggest leak between paid spend and CRO outcomes.

Across the last 30 paid-account audits I've run, the same pattern shows up in 23 of them. Maximise Conversions enabled. Fewer than 30 conversions in the previous 30 days. Conversion tracking events firing on thank-you-page loads that include returning customers, organic traffic, and direct visits. The algorithm learns from polluted data, feeds the campaign with more polluted data, and the £40,000 to £80,000 a month being spent is essentially a donation to Google's machine-learning training set.

Bidding strategy is treated as a setting rather than a strategic choice. Every campaign has a goal, the goal dictates the bidding strategy, the bidding strategy dictates the data the algorithm needs to function. Break any link and the chain leaks.

The AI CRO pillar guide makes the broader point: AI tools amplify operator decisions, they do not replace them. Smart Bidding fits the same shape. The algorithm is genuinely good. Feed it the wrong goal, the wrong data, or the wrong constraints and you get a fast, confident, expensive wrong answer.

The 2026 layer compounds the problem. Performance Max has been the default recommendation for new campaigns for 18 months. Demand Gen has replaced Discovery. Smart Bidding signals have expanded. Operators who set bidding strategies in 2022 or 2023 are running on stale assumptions.

The bidding strategy is the contract between your campaign goal and Google's algorithm. Mismatch the contract and you get a fast, confident, expensive answer to the wrong question.

The bidding-strategy-to-campaign-goal matrix (the spine of this pillar)

The bidding-strategy-to-campaign-goal matrix is the spine of this pillar. Match the strategy to the goal and the rest of the optimisation work compounds. Mismatch it and no amount of creative or landing-page work recovers the loss. The matrix below is the version I use on every paid-account audit.

Read it as a decision tree. The lever, the goal it serves, the data it needs, the situation where it damages the account. The audit finding I see most often: a lever in the wrong row. Target CPA enabled on an account producing 12 conversions in 30 days. Target ROAS enabled on an account whose conversion-value tracking includes free trial signups (zero value).

The matrix encodes the 2026 modernisation layer. Performance Max appears as two rows because PMax bidding is its own context, not another option on a Search campaign. Demand Gen lives upstream as the awareness companion to PMax. For the discipline behind every test in this matrix, the OperatorAI methodology describes the 99% statistical-significance standard I run on every bidding-strategy switch.

Wrong bidding strategy plus right CRO programme equals an unprofitable account. Right bidding strategy plus wrong CRO programme equals an unprofitable account. The pair has to ship together.

Maximise Clicks: when traffic volume is the goal

Maximise Clicks is the right bidding strategy when raw click volume is the campaign goal and you have a downstream funnel that converts the traffic. It sets bids automatically to capture as many clicks as your budget allows. The strategy works for top-of-funnel content campaigns, brand discovery, and cold prospecting where you have a deep retargeting pool waiting downstream.

The strategy is widely misused. Operators reach for Maximise Clicks when they want "more conversions" without realising the algorithm has no conversion signal to optimise toward. The campaign delivers click volume to whatever query is cheapest to win, regardless of intent quality. With a 1.2% conversion rate, you'll get 100 clicks for the price of 60 quality ones, CPA goes up.

Three rules:

1. Use it when the goal is genuinely click volume. Top-of-funnel content distribution. YouTube end-card click campaigns. Awareness pools that feed retargeting.
2. Pair it with a hard maximum CPC bid. Without a ceiling, the algorithm pushes bids high on cheap-to-win irrelevant queries.
3. Don't use it as a substitute for Smart Bidding when conversion data is missing. Use manual CPC with tight keyword control instead.

Brand-term defence campaigns with a low CPC ceiling are the legitimate exception.

Maximise Clicks is the right tool for the wrong question 8 audits out of 10. Click volume is rarely the actual goal, and the bidding strategy needs to match the actual goal.

Target CPA: when conversion volume at a fixed cost is the goal

Target CPA is the default bidding strategy for conversion-focused campaigns where you have a known acceptable cost-per-acquisition and enough historical conversion data to feed the algorithm. Google sets bids automatically to deliver as many conversions as possible at your target. Minimum data threshold: 30+ conversions in the last 30 days.

The tactical rule I use on every audit: starting target CPA equals your existing CPA multiplied by 0.85. The 15% buffer gives the algorithm room to optimise without stalling. Over-restricting (setting target 30% to 50% below historical CPA on day one) is the most common cause of campaigns refusing to spend after migration.

Across GoGoChimp client engagements, the Growth tier programme runs 30+ A/B experiments per quarter at 99% statistical significance. Bidding-strategy iteration runs alongside that cadence: review weekly, adjust target by no more than 10% per iteration, hold for the full learning period, repeat.

Three operator implications:

1. Conversion tracking integrity is the precondition. Audit the conversion events firing in GA4 and the Google Ads conversion column. Confirm volume matches within 5%.
2. Audience segmentation feeds the target. A target CPA of £45 across mixed B2B and B2C traffic fits neither. Split campaigns by audience first.
3. Don't run target CPA inside Performance Max if standard Search runs the same goal. PMax cannibalises Search when goals overlap.

The Donate For Charity work is the proof point: 494.64% increase in donations in 30 days. Target CPA paired with a converting funnel, conversion-tracking integrity, and a 99%-significance test cadence on creative. The bidding strategy was the multiplier, not the engine.

Target CPA is the operator's safest Smart Bidding lever. The algorithm is good. The data is the bottleneck. Fix conversion tracking first, audience segmentation second, target third.

Target ROAS: when ecommerce revenue is the goal

Target ROAS is the right bidding strategy for ecommerce accounts where revenue per conversion varies by product and conversion-value tracking is enabled. Google optimises bids to deliver the highest revenue at your target return-on-ad-spend ratio. Minimum data threshold: 50+ conversions with values in the last 30 days.

Instead of treating every conversion as equal (target CPA), target ROAS treats a £200 order as 4x more valuable than a £50 order and bids accordingly. The algorithm pushes spend toward queries, audiences, and times of day that produce higher-value conversions. For Shopify, WooCommerce, or any direct-revenue ecommerce account where AOV varies, target ROAS routinely beats target CPA at the same spend by 15% to 30% on revenue.

The Enzymedica work is the proof point. The Black Friday 2021 campaign drove a 3.4% baseline conversion rate to 16.9% on the Black Friday day, with 11% sustained through December 2021 (one of the worst months for health-supplement sales). The prior year's Black Friday, without GoGoChimp, ran at roughly 7%. The 16.9% figure is a 2.4x lift on the same promo day, same product. Target ROAS was the paid-side bidding strategy, paired with the converting funnel and the 99%-significance test cadence. The Loom analytics review: https://www.loom.com/share/d20fd92f4d5e49a88a92c9c0d5e28570.

Three operator implications:

1. Conversion-value integrity is mandatory. The dynamic-value tracking script needs to fire the actual order value, after discount, before tax, on the thank-you page. Hardcoded values mean target ROAS optimises against fiction.
2. Set target ROAS conservatively. Start at 80% to 90% of historical, hold 7 to 14 days, then ratchet.
3. Don't run it on lead-gen with form submissions of equal "value". Without genuine variance, target ROAS collapses to target CPA dressed up.

In 2026, target ROAS lives inside both standard Shopping and Performance Max. PMax-side optimises blindly across surfaces. Standard-Shopping retains operator-level keyword visibility.

Target ROAS is the operator's revenue lever when product margin varies. The algorithm bids more aggressively on the £200 sale than the £50 sale. The 5x range between products becomes a 1.3x to 1.5x range in account-level revenue.

Target CPM: the underrated awareness strategy

Target CPM is the right bidding strategy for awareness campaigns where the goal is unique reach across a defined audience at a controlled cost per thousand impressions. It works for video campaigns, YouTube prerolls, and retargeting where the impression itself is the goal. Google's statistical model deduplicates audience across devices, so the "thousand impressions" measure approximates true unique reach rather than raw view count.

The mechanism is statistical. The algorithm observes aggregated user behaviour across browsers and devices, accounts for cross-device usage patterns, and combines that with census-level survey data to deduplicate. Target CPM lets you set the price per thousand of those deduplicated impressions, and the algorithm holds the campaign average at or below that target.

Three rules:

1. Use it for video and prerolls. YouTube ad campaigns, in-stream prerolls, retargeting display where the impression is the goal.
2. Use it for top-of-funnel budget. Awareness campaigns that feed retargeting pools downstream. The brand-discovery surface PMax doesn't serve well.
3. Don't use it for direct response. Target CPA or target ROAS exists for that. Target CPM is the worst tool for converting an impression pool.

In 2026, Demand Gen has expanded the surfaces target CPM runs on. YouTube Shorts, Discover, and Gmail now serve target CPM impressions in the same buy. To lower CPM cost, raise Quality Score by improving creative and landing-page relevance, not by lowering the target.

Target CPM is the cheapest way to fill the top of the funnel when the funnel is already engineered to convert. The strategy doesn't fix a broken funnel. It floods the entry point of a working one.

Maximise Conversions and Maximise Conversion Value: the high-data plays

Maximise Conversions and Maximise Conversion Value are the bidding strategies for accounts that have already established conversion volume and want to remove the target constraint. There is no target CPA or target ROAS. The algorithm spends the full budget freely, optimising for either total conversion count or total conversion value. Highest power, highest data dependency.

The trap: these look like the obvious starting point on the Google Ads UI. They are not. They are the continuation point after a target CPA or target ROAS campaign has matured for 60 to 90 days, hit its target reliably, and has 50+ conversions feeding the algorithm. Enabling Maximise Conversions on a fresh account with under 30 conversions in 30 days is the fastest way I know to lose £40,000 a month to the wrong queries.

The decision rule:

1. Has the account hit 50+ conversions in 30 days at the campaign level? If no, run target CPA or target ROAS until data depth crosses.
2. Has the campaign run at target CPA / target ROAS for at least 60 days at the target? If no, the algorithm has not learned the audience, queries, or conversion paths.
3. Is the budget set at a level where 30% growth would not destabilise the account? Maximise Conversions spends the full budget. If 30% upside spend creates cash-flow issues, use a target-based strategy.

Maximise Conversion Value is the ecommerce variant. Same rules, plus conversion-value tracking integrity as a precondition. For mature ecommerce accounts running £100,000 to £500,000 a month with conversion-value variance and clean tracking, it's the highest-revenue strategy available in 8 audits out of 10. Build Grow Scale's 28-34% lift finding for expert-guided AI versus DIY tools applies directly to Smart Bidding. The AI is the same. The operator is the differentiator.

Maximise Conversions is the lever you reach for at the 60-day mark of a mature target-CPA account, not the lever you reach for on day one of a new campaign. Day-one Maximise Conversions is the fastest way to lose budget I see in audits.

Performance Max bidding: the 2026 state

Performance Max in 2026 is the default bidding context for ecommerce and lead-gen accounts at scale. PMax replaces single-channel campaigns with a multi-channel asset-group structure running on Search, Display, YouTube, Discover, and Gmail simultaneously. Bidding strategies inside PMax: Maximise Conversions, Maximise Conversion Value, target CPA, target ROAS.

The critical 2026 caveat: don't run PMax and standard Search on the same campaign goal. Across the 30 paid-account audits I've run in 2025-2026, PMax on the same conversion goal as a standard Search campaign produces cannibalisation in 24 of them. PMax wins the auction inside the operator's own account, replacing higher-quality Search clicks with lower-quality PMax-attributed ones.

Three rules:

1. Pick one campaign per goal. If conversion-focused budget is going to PMax, pause the standard Search campaign on the same goal. The two cannibalise each other.
2. Use audience signals as a starting point, not a constraint. PMax signals are seeds, not targeting locks. Operators who treat them like Display targeting get expanded reach they did not intend.
3. Brand exclusions matter. PMax absorbs brand-term traffic into conversion attribution by default. Add brand exclusions to keep brand traffic in the standard brand-defence campaign where attribution is honest.

Demand Gen sits above PMax. PMax serves conversion-focused budget; Demand Gen serves awareness-focused budget that feeds PMax remarketing pools. Operators running both as a paired upper-funnel + lower-funnel architecture see better PMax efficiency than operators running PMax in isolation. For ecommerce, PMax with target ROAS is the strongest single bidding-strategy decision in 2026. For lead-gen, PMax with target CPA.

Performance Max is the 2026 default for ecommerce + lead-gen at scale. Run it next to standard Search on the same goal and you'll lose the comparison and the spend efficiency simultaneously. Pick one per goal.

Smart Bidding 2026: what the AI actually does (and what it doesn't)

Smart Bidding in 2026 uses a signal layer that includes device, location, time of day, query, audience, browser, OS, demographic, in-market signals, and contextual signals like weather and local events. The algorithm processes these signals at auction time to predict conversion probability and value for each impression, and bids accordingly. The signal depth is impressive. It is also not the differentiator.

What Smart Bidding does well: predicts conversion probability per impression at scale, adjusts bids in real time across hundreds of thousands of auctions per day. What it does not do: optimise creative, landing page, offer, or the conversion definition. The AI bidding cleaner doesn't fix a broken funnel. It spends faster against one.

Build Grow Scale's research across 347 stores applies directly. Expert-guided AI delivers 28-34% lift versus 4-7% from DIY tools. In paid bidding the parallel reads: expert-guided Smart Bidding (operator sets the audience, conversion definition, negative keywords, bid adjustments, campaign structure) versus default Smart Bidding (algorithm runs against polluted data with no guardrails). Same software, 5x different outcome.

AI bidding alone, without the CRO programme to convert the traffic, sits in the 4-7% range. Switching from manual to Smart Bidding on a 1.2% conversion-rate funnel gives you a 1.25% to 1.30% funnel. Switching to a 4.5% conversion-rate funnel takes a CRO programme. The agency-vs-DIY-AI-tools comparison lays out the operator-led path.

Three rules:

1. Conversion definition integrity matters more than the signal layer. A campaign optimising against a polluted conversion event produces fast, confident, expensive wrong answers.
2. The signal layer expands every quarter. Re-audit campaign-level bid adjustments and audience signals quarterly.
3. Operator-side guardrails are non-optional. Bid caps on Maximise Clicks. Brand exclusions on PMax. Negative keyword lists on Search. Audience exclusions on retargeting.

Build Grow Scale's 2026 review across 347 stores (Stafford, 2026) found expert-guided AI delivered 28-34% lifts; DIY AI tools delivered 4-7%. Smart Bidding fits the same shape. The AI is not the differentiator. The operator running it is.

The £20K/month threshold: when manual beats Smart

The £20,000 a month threshold is the line below which manual CPC or Enhanced CPC routinely beats Smart Bidding on the same campaign goal. The mechanism is data sparsity. Smart Bidding needs conversion volume to function, and accounts spending under £20,000 a month rarely produce 30+ conversions per 30 days at the campaign level. Below the threshold, the operator is the data layer.

This is the most counter-intuitive finding I deliver across audits. Operators below the threshold have been told by Google's account managers, by Help docs, and by every paid-ads consultant they've hired that Smart Bidding is the right choice. The reality: under £20,000 a month, Smart Bidding produces a 2-week learning phase, a 4-week stall, and a 6-week ratcheting CPA. Manual CPC with tight keyword control produces predictable performance from day one.

With 8 to 15 conversions per month, the algorithm has roughly 1 conversion per 100 to 200 clicks across 5,000 to 30,000 auction events. The signal-to-noise ratio is too low for the algorithm to identify which signals predict the conversion. The campaign bids randomly within its target, hits the target by chance, underperforms manual on the same spend.

Three rules:

1. Manual CPC with portfolio bid strategies. Group keywords by intent tier. Set bids by tier. Adjust weekly.
2. Enhanced CPC as the bridge. When conversions cross 15+ per 30 days, switch to Enhanced CPC. It uses Smart Bidding signals to nudge manual bids without taking full control.
3. Above £20,000 a month with 30+ conversions, switch. Target CPA or target ROAS, depending on goal.

The threshold is a band, not a hard line: £15,000 to £25,000 a month, depending on conversion rate and AOV. The decision criterion is conversion volume per 30 days at the campaign level, not raw spend.

For Glasgow-based ecommerce and B2B accounts, what to look for in a CRO agency in Glasgow covers the operator-side framing. Below the threshold, manual bid management is the differentiator. Above it, Smart Bidding configuration discipline is.

Smart Bidding needs data the way a flywheel needs torque. Below 30 conversions per 30 days, the flywheel doesn't spin. Manual CPC plus operator pattern recognition beats algorithmic guesswork at sub-£20,000 a month spend.

How to switch bidding strategy without losing performance (the migration)

Switching bidding strategy resets the learning phase. Smart Bidding requires 7 to 14 days to recalibrate against the new target. During the learning phase, performance fluctuates by 30% to 50% on either side of historical. The migration discipline is the difference between a 14-day disruption that converges and a 60-day disruption that never recovers.

The migration playbook I run on every switch:

1. Switch one campaign at a time. Pick the highest-data campaign first. Hold 14 days. Measure. Move on.
2. Hold budget steady. Don't change daily budget by more than 20% during the learning phase. The algorithm reads budget changes as signal volatility and over-corrects.
3. Don't change creative simultaneously. Two learning curves interfere; you can't attribute performance changes to either.
4. Monitor at 99% statistical significance. The 99-Rule discipline is the GoGoChimp standard. Don't call a result before the campaign has reached 99% confidence on the difference.
5. Rollback rule: 14 days of underperformance equals revert. If running 20%+ below historical at the 14-day mark, revert.

Three common errors: migrating during a high-variance period (Black Friday, brand campaign launch, creative refresh, signal noise overwhelms the learning phase); changing target by 30%+ on day one (set at 90% of historical for the first 14 days, then ratchet); stopping migration before it converges (week-1 underperformance is the learning phase by definition).

Migration also exposes conversion-tracking issues. Across audits, 4 in 10 migration failures trace to conversion-tracking integrity, not the strategy switch itself. Audit conversion tracking before migrating.

The migration discipline is what separates operators from gamblers. Hold steady, change one variable, measure at 14 days, rollback at the threshold. The algorithm doesn't know you panicked. It records the noise as signal.

Frequently asked questions

What is the best Google Ads bidding strategy in 2026?

There is no single best bidding strategy. The right strategy depends on the campaign goal, the data depth, and the funnel beneath it. Target CPA fits conversion-focused search above 30 conversions per 30 days. Target ROAS fits ecommerce with conversion-value variance above 50 conversions. Target CPM fits awareness and video. Maximise Conversions fits mature accounts after 60 days at target.

When should I use target CPA versus target ROAS?

Use target CPA when every conversion is roughly equal in value (lead-gen, single-product ecommerce, subscription signups). Use target ROAS when conversion value varies significantly (multi-product ecommerce, subscription tiers with different price points). Target ROAS requires conversion-value tracking integrity. Without clean order values firing on the thank-you page, target ROAS optimises against fiction.

What is the minimum spend for Smart Bidding to work?

Smart Bidding needs 30+ conversions per 30 days at the campaign level for target CPA, and 50+ conversions with values for target ROAS. In spend terms, that's roughly £15,000 to £25,000 a month depending on conversion rate and AOV. Below that band, manual CPC or Enhanced CPC outperforms Smart Bidding because the operator is the data layer.

Should I use Maximise Conversions for new campaigns?

No. Maximise Conversions on a new campaign is the fastest way to lose budget I see in audits. The strategy spends the full budget while the algorithm has no conversion history to optimise against. Use target CPA for the first 60 to 90 days, hit the target reliably, then consider Maximise Conversions as the scaling move.

How long does Smart Bidding take to learn?

The learning phase runs 7 to 14 days after a strategy switch. During the phase, performance fluctuates 30% to 50% on either side of historical. Operators who panic-switch back in week 1 abort the learning phase before convergence. Hold 14 days, measure at 99% statistical significance, roll back only if performance is 20%+ below historical at the 14-day mark.

Is Performance Max better than standard Search campaigns?

Performance Max is the better choice for ecommerce and lead-gen at scale (£100,000+ a month) where multi-channel reach is the goal. Standard Search is the better choice for tightly-targeted intent campaigns where operator-level keyword visibility matters and the campaign goal doesn't need Display, YouTube, Discover, or Gmail reach. The cannibalisation rule applies: don't run both on the same goal in the same account.

What is target CPM used for?

Target CPM is for awareness campaigns where the goal is unique reach across a defined audience at a controlled cost per thousand impressions. It works for video, YouTube prerolls, retargeting display, and brand-discovery campaigns where the impression itself is the goal. It doesn't work for direct-response or conversion-focused campaigns. Target CPM fills the top of the funnel; the lower funnel needs a different lever.

Why does my Smart Bidding strategy keep underperforming?

Six common causes, in order of audit frequency: conversion tracking is broken or polluted; conversion volume is below the 30-per-30-day threshold; target was set too aggressively at migration (more than 15% below historical); creative or landing page changed simultaneously with the switch; PMax is cannibalising the Search campaign's attribution; the audience signal is too narrow. Audit the conversion event first.

What to do next

If your account spends over £20,000 a month on Google Ads and you suspect the bidding strategy is mismatched to the campaign goal, run a free AI audit. The audit covers conversion-tracking integrity, bidding-strategy match, Performance Max cannibalisation, and Smart Bidding signal-layer configuration. It comes back in 48 hours with a prioritised list of leaks.

If you spend under £20,000 a month, the audit isn't the right next step. Manual CPC with tight keyword control beats Smart Bidding at that scale. Read the AI CRO pillar for the broader framing on how operator-led testing compounds with paid traffic.

Either path, the bidding strategy is the contract between your campaign goal and Google's algorithm. Match the contract to the goal. The CRO programme will compound on what comes through.

References

• Stafford, M. (2026). 2026 CRO Year in Review: What Worked, What Failed, What's Next. Build Grow Scale, 9 April 2026. https://buildgrowscale.com/cro-trends-2026-recap
• Google Ads Help. (2026). About Smart Bidding.
• Google Ads Help. (2026). About Performance Max campaigns.
• Google Ads Help. (2026). About Demand Gen campaigns.
• Google Ads Help. (2026). About target CPA bidding.
• Google Ads Help. (2026). About target ROAS bidding.
• Google Ads Help. (2026). About target CPM bidding.
• GoGoChimp Case Studies. Enzymedica Black Friday 2021. Loom analytics review: https://www.loom.com/share/d20fd92f4d5e49a88a92c9c0d5e28570
• GoGoChimp Case Studies. Donate For Charity (494.64% donations in 30 days).

Where this fits in the OperatorAI methodology

This pillar fits inside The 4-to-34 Gap framework: the difference between DIY AI tools delivering 4-7% lift and operator-led AI delivering 28-34% lift. Bidding strategy is one of the operator decisions that compounds across the gap. Algorithm-default bidding sits at the bottom; operator-set bidding matched to campaign goal sits at the top.

For the operating-model maturity classification that determines whether your account is ready for Smart Bidding at all, see the OperatorAI maturity model. Accounts at the lower tiers of the model run manual CPC and Enhanced CPC; accounts at the upper tiers run target CPA, target ROAS, and Maximise Conversion Value with confidence.