Google Ads AI Bidding: Which Strategy to Use in 2026

Google's AI bidding has come a long way from the days when "automated bidding" meant handing over control and hoping for the best. In 2026, smart bidding strategies are genuinely sophisticated — they process dozens of auction-time signals including device, location, search query intent, time of day, audience segment, and even browser type, all within milliseconds. But that sophistication cuts both ways. When you choose the right strategy with the right setup, results compound fast. When you get it wrong, the algorithm optimises confidently toward the wrong outcome and your budget disappears quietly.

This is a topic we deal with every week at Workflow AI Advisors. Across our client base — spanning e-commerce, SaaS, professional services, and lead generation — we've seen the same mistakes repeated: campaigns forced into Target ROAS before they have enough data, Max Clicks used as a "safe default" that generates zero conversions, and Enhanced CPC being treated as a smart bidding strategy when it barely qualifies. Let's cut through the noise.

How Google's Smart Bidding Actually Works in 2026

Smart bidding uses machine learning to set bids at the individual auction level. Unlike manual CPC, where you set a single bid that applies broadly, smart bidding adjusts in real time based on predicted conversion probability. Google's models pull from your account's historical conversion data, broader Google signals, and contextual factors at the moment of each search.

The critical point most advertisers miss: the model is only as good as the conversion data you feed it. If your conversion tracking is broken, delayed, or measuring the wrong actions, the algorithm optimises toward whatever it can measure — and that's rarely what you actually want. Before you touch your bidding strategy, audit your tracking. Every time, without exception.

In 2026, Google has also expanded its use of "broad match + smart bidding" pairing as the default recommendation. This is contextually powerful when the account has strong data volume, but it's a liability in low-volume or niche accounts where the model doesn't have enough signal to filter irrelevant queries. Keep that in mind as we go through each strategy.

The Full Breakdown: Every Google Ads AI Bidding Strategy

1. Maximise Conversions

What it does: Spends your entire daily budget with the goal of getting as many conversions as possible, without a specific cost constraint.

When to use it: New campaigns with limited conversion history. If you have fewer than 30–50 conversions in the past 30 days, Target CPA will starve itself trying to find an efficiency target it can't yet compute. Max Conversions lets the algorithm explore and accumulate data while still optimising directionally toward your conversion goal.

When to avoid it: When your budget is large relative to historical spend. Without a CPA constraint, Max Conversions will find ways to spend every pound or dollar — including on lower-quality traffic. We've seen this strategy drive strong conversion volume at CPAs that make no commercial sense. Always set a Target CPA within Max Conversions if you have any sense of your acceptable cost per conversion.

2026 note: Google now allows you to set a Target CPA directly within the Maximise Conversions strategy, effectively blending it with tCPA. Use this. It gives the algorithm exploration room while preventing runaway costs.

2. Target CPA (tCPA)

What it does: Sets bids to achieve your goal cost per conversion. Some conversions will come in above target, some below — the algorithm averages across the campaign over time.

When to use it: Lead generation businesses with a clear understanding of acceptable customer acquisition cost. If you know a qualified lead is worth £80 to your business and you're currently converting at £60 CPA, tCPA is your natural home. It's also the right strategy for accounts that have hit that 30–50 conversion/month threshold and want to scale efficiently.

Common mistake: Setting the target CPA too aggressively below your current average. If your historical CPA is £90 and you set a target of £45, the algorithm will throttle impressions so severely the campaign barely runs. A better approach is to set the target at or slightly below your current average, let it stabilise for 2–3 weeks, then compress the target gradually — no more than 10–15% at a time.

When to avoid it: Low-volume campaigns, or businesses where conversion values vary significantly between customers. In those cases, ROAS-based bidding is more appropriate.

3. Target ROAS (tROAS)

What it does: Optimises bids to achieve a specific return on ad spend. Unlike tCPA which treats all conversions equally, tROAS weights bids based on predicted conversion value — so higher-value clicks get more aggressive bids.

When to use it: E-commerce accounts where transaction values vary. If someone is equally likely to buy a £20 product or a £200 product, you want the algorithm bidding more aggressively for the higher-value customer. tROAS makes this possible. At Workflow AI Advisors, our e-commerce clients running tROAS with proper value tracking average a 4.2x ROAS across campaigns — but that number only holds because the conversion value data is clean and complete.

Data requirement: Google recommends at least 50 conversions in the past 30 days for tROAS to function properly — and honestly, we'd push that to 100+ for meaningful stability. Below that threshold, you'll see erratic spend patterns and target misses.

When to avoid it: Service businesses where all conversions have the same nominal value (e.g., form fills assigned a fixed £1 value). In that scenario, tCPA is simpler and equally effective.

4. Maximise Conversion Value

What it does: The value-equivalent of Maximise Conversions — spends your full budget chasing the highest possible total conversion value without a ROAS constraint.

When to use it: When you're launching a new e-commerce campaign and need to build revenue data before setting a ROAS target. It's also useful during high-volume periods (like Q4) when you want to capture demand aggressively without being throttled by an efficiency target.

Risk: Same as Maximise Conversions — without a ROAS target set within the strategy, the algorithm will spend freely. Always pair with a target ROAS value once you have enough data to set one sensibly.

5. Target Impression Share

What it does: Bids to achieve a specific share of impressions in a chosen location (top of page, absolute top, or anywhere on page).

When to use it: Brand defence campaigns where your goal is presence, not direct conversion efficiency. If a competitor is bidding on your brand terms and you want to hold the top position, impression share bidding makes sense. Also useful for brand awareness plays where you're tracking upper-funnel metrics rather than direct conversions.

When to avoid it: Performance campaigns. Using Target Impression Share for a direct response campaign is one of the most reliable ways to burn budget without results. It completely ignores conversion probability — you'll win impressions from users who will never convert, at whatever price is needed to hit your share target.

6. Enhanced CPC (eCPC)

What it does: Adjusts your manual CPC bids up or down based on conversion likelihood. It's the lightest form of automation.

When to use it: Honestly? Rarely in 2026. eCPC is a transitional strategy — it was valuable when smart bidding models were less mature, but Google's own data shows that full smart bidding strategies outperform eCPC in the vast majority of cases. The exception might be accounts where you have strong manual bidding expertise and a specific reason to maintain granular bid control (certain regulated industries, very low conversion volumes).

The honest take: If you're using eCPC as a long-term strategy rather than a stepping stone, you're probably leaving performance on the table.

The Data Foundation: Why Strategy Selection Is Only Half the Battle

No smart bidding strategy compensates for weak conversion tracking. Before optimising bidding, confirm:

• Conversion actions are correctly tagged — Google Tag Manager implementation verified, no duplicate tracking, no unverified tags counting as conversions
• Conversion values are accurate — For e-commerce, dynamic values passed at checkout. For lead gen, assigned values that reflect actual lead quality differences
• Attribution model matches your sales cycle — Data-driven attribution (DDA) is now the default and generally the right choice, but verify it has enough data to function (typically 300+ conversions and 3,000+ clicks in 30 days)
• Offline conversion imports are set up — If your conversions happen offline (calls, in-person sales), importing those signals dramatically improves bidding model accuracy

Our AI automation services include building the data pipelines that feed these signals back into Google Ads automatically — CRM integrations, offline conversion imports, and real-time value tracking that most agencies set up once and forget. The difference in bidding performance is significant.

Campaign Structure Considerations That Affect Bidding

Bidding strategy and campaign structure are inseparable. A few principles that hold in 2026:

Consolidate campaigns to feed data pools. Fragmented account structures — dozens of small campaigns each with limited conversion volume — starve smart bidding models. Where possible, consolidate to give each bidding strategy enough signal. A campaign with 15 conversions a month will never perform as well under tCPA as one with 60.

Performance Max and smart bidding. PMax campaigns run their own automated bidding within the campaign type — you set a ROAS or CPA target and Google handles the rest across all inventory. The structural question is whether PMax should run alongside Search campaigns or replace them. Our current recommendation for most accounts: run PMax alongside branded and high-intent Search campaigns, not as a wholesale replacement. Monitor search term insights carefully.

Portfolio bid strategies. If you're managing multiple campaigns with the same goal, portfolio bid strategies let a single tCPA or tROAS target govern all of them collectively, pooling conversion data across campaigns. This is underused and often delivers better stability than individual campaign-level strategies.

Choosing Your Strategy: A Decision Framework

Here's the framework we apply across client accounts:

1. Under 30 conversions/month per campaign: Maximise Conversions (with a CPA target if you have any cost benchmark)
2. 30–100 conversions/month, uniform conversion value: Target CPA
3. 50+ conversions/month, variable conversion values: Target ROAS
4. Brand defence or awareness campaign: Target Impression Share
5. New e-commerce campaign building value data: Maximise Conversion Value, transitioning to tROAS at volume
6. Regulated industries with manual control requirements: eCPC as a transitional measure

The key word in all of this is transitional. Smart bidding strategies should evolve as your account matures. An account that starts on Max Conversions in January should, ideally, be on tCPA or tROAS by Q3 — with a clear plan to compress targets as efficiency improves.

What to Monitor After Switching Strategies

Every bidding strategy change triggers a learning period of approximately 1–2 weeks. During this time, performance will fluctuate — sometimes significantly. Do not make further changes during the learning phase. Evaluate performance over at least 30 days (longer for low-volume campaigns) before drawing conclusions.

Key metrics to track post-switch:

• Impression share lost to budget vs. lost to rank
• Average CPA or ROAS vs. target
• Conversion volume week-over-week
• Search term quality (via search terms report)
• Auction insights — are you gaining or losing competitive position?

If you're running paid media across multiple markets — a common scenario for our clients in the US, UK, UAE, and Australia — separate campaigns by market and apply bidding strategies independently. CPAs and ROAS targets that work in the UK may be completely wrong benchmarks for Australian or Singapore markets. Don't average across geographies at the strategy level.

For deeper guidance on how paid media integrates with organic growth, our SEO and GEO services page covers how we build full-funnel visibility that makes paid investment more efficient over time.