Google Ads Automated Bidding: Conversion Data Needed

Google Ads automated bidding needs at least 15 to 30 conversions per month per campaign to work reliably, with 30 or more being the recommended benchmark for stable performance. This article explains why that number matters, how to measure it correctly, and what to do if you don’t have enough conversion data yet.

How Automated Bidding Works

Automated bidding uses machine learning to set your bid for each auction based on the likelihood that a click will lead to a conversion. The system analyzes historical data—past clicks, conversions, device, location, time of day, and search query—to predict which auctions are most valuable. With enough conversion history, the algorithm can identify patterns and adjust bids in real time to maximize your chosen goal, whether that’s conversions, conversion value, or target CPA/ROAS.

Without sufficient data, the model has little to learn from. It may make random or overly cautious bids, leading to poor performance or wasted spend. Google’s own documentation suggests that for most campaigns, you need at least 15 conversions in the last 30 days for Target CPA, and 30 conversions for Target ROAS. For Maximize Conversions, the threshold is similar—around 15 to 30 conversions per month—though it can work with fewer if you set a target CPA.

Why 15 or 30 Conversions?

The numbers come from Google’s internal testing and the practical realities of machine learning. A model needs enough positive examples to distinguish high-performing from low-performing auctions. With fewer than 15 conversions, the signal-to-noise ratio is too low: random fluctuations in conversion rate can be mistaken for real patterns. At 30 conversions, the model has a more stable baseline and can start optimizing with confidence.

These are not hard rules. Some campaigns work with fewer conversions, especially if the conversion action is easy to predict (like a newsletter signup) or if the account has strong historical data from other campaigns. Conversely, campaigns with very high-value, rare conversions (like B2B software purchases) may need more than 30 to achieve stability. The key is to monitor performance and give the algorithm enough time—usually two to three weeks—to learn before making judgments.

How to Count Conversions Correctly

Before you worry about hitting a threshold, make sure you’re counting conversions the right way. A conversion is any action you’ve defined as valuable in Google Ads, such as a purchase, form submission, phone call, or sign-up. You can set up conversion tracking via Google Tag Manager, a global site tag, or Google Analytics 4 (GA4) import.

• Use primary actions only. Secondary actions don’t count toward automated bidding. If you’ve marked some conversions as “secondary,” they won’t help the algorithm.
• Include all relevant conversion actions. If you have multiple goals (e.g., purchases and newsletter sign-ups), decide which ones you want the algorithm to optimize for. You can combine them into one primary action or use separate campaigns.
• Count unique conversions. By default, Google counts each conversion once per click. If a user converts twice from the same click, only the first is counted. This is usually what you want.
• Attribution window matters. A conversion that happens 30 days after a click still counts toward that click’s data. If you shorten the window, you’ll see fewer conversions, which could drop you below the threshold.

What If You Have Fewer Than 15 Conversions?

If you’re starting a new campaign or have low volume, don’t force automated bidding immediately. Here are practical steps to build up data:

1. Start with manual CPC or enhanced CPC. These give you control while still using some automation. Enhanced CPC adjusts your bid up or down based on conversion likelihood, but it doesn’t require a minimum data threshold.
2. Use a conversion action with more volume. If purchases are rare, consider optimizing for a higher-funnel action like add-to-cart or email sign-up, then gradually shift to more valuable actions.
3. Consolidate campaigns. Combine campaigns with similar products or services to pool conversion data. A single campaign with 30 conversions will outperform two campaigns with 15 each.
4. Lower your target CPA temporarily. A more aggressive target can generate more conversions faster, but be careful not to overspend. Start with a CPA that’s 20-30% above your current average.
5. Use portfolio bidding. If you have multiple campaigns, you can apply a shared bidding strategy across them, which allows the algorithm to learn from all conversions collectively.

How to Choose the Right Bidding Strategy

Google Ads offers several automated strategies, each with its own data requirements:

• Maximize Conversions: Automatically sets bids to get the most conversions within your budget. Works with as few as 15 conversions/month, but more is better. Good for new campaigns or when you don’t have a target CPA.
• Target CPA: Aims to get conversions at or below your target cost. Requires 15+ conversions/month. Best when you have a specific cost per acquisition goal.
• Target ROAS: Optimizes for conversion value (revenue) rather than raw conversions. Requires 30+ conversions/month and accurate conversion values. Best for e-commerce.
• Maximize Conversion Value: Aims to get the highest total conversion value within your budget. Similar data needs to Target ROAS but without a target.
• Enhanced CPC: A semi-automated strategy that adjusts your manual bids. Can work with very little data, making it a good starting point.

Common Problems and Misconceptions

One common misconception is that more conversions always mean better performance. While more data helps the algorithm, quality matters too. If your conversion tracking is broken or you’re counting low-quality actions, the algorithm will optimize for the wrong thing. Always verify that your conversion tags are firing correctly and that your conversion values are accurate.

Another issue is switching strategies too frequently. The algorithm needs time to learn—typically 2-3 weeks. If you change your bid strategy every few days, the model never gets a stable baseline, and performance will suffer. Stick with one strategy for at least a month before evaluating.

Also, be aware of seasonality. If you launch automated bidding during a peak season, the algorithm may learn patterns that don’t apply later. Conversely, if you start during a lull, it may become too conservative. Try to start during a “normal” period or adjust your expectations accordingly.

Frequently Asked Questions

Can automated bidding work with zero conversions?

Technically, you can enable it, but it won’t have anything to learn from. Google will use general market data and your campaign settings, but performance is likely to be erratic. It’s better to start with manual bidding or enhanced CPC until you have at least a few conversions.

Does the 30-conversion rule apply to all campaign types?

Google’s guidance is general, but Search and Shopping campaigns typically need fewer conversions than Display or Video. Display campaigns often have lower conversion rates, so they may need more data. In all cases, the more conversions you have, the better the algorithm performs.

How long should I wait before judging automated bidding?

Give it at least 2-3 weeks. The algorithm needs time to collect data and adjust. If you see no improvement after 4-6 weeks, consider changing your target CPA/ROAS or switching strategies.

Can I use automated bidding with a low budget?

Yes, but your budget must be high enough to generate the required conversions. If you only have $10/day and your product costs $500, you may not get 30 conversions a month. In that case, consider a longer conversion window or a different conversion action.

What if I have 30 conversions but they’re spread across many campaigns?

Each campaign learns independently unless you use portfolio bidding. If you have 30 conversions total but spread across 10 campaigns, each campaign only has 3, which is insufficient. Consolidate campaigns or use a shared bidding strategy.

Does the conversion delay affect the required data?

Yes. If your sales cycle is long (e.g., 30-day consideration), the algorithm needs more historical data to account for the delay. Google uses conversion lag time in its models, but with fewer than 30 conversions, it may not have enough examples to predict future conversions accurately.

Related

• Expandable Housing & Faster Project Delivery
• Duplicate File Removal: What to Remember
• Convert FAT32 to NTFS Without Losing Data? Yes, Here’s How

in t | Conversions per month | conversions | 30 | 1 | 100000
in c | Current cost per conversion | currency | 50 | 0.01 | 100000
in d | Days in month | days | 30 | 1 | 31
out daily | Daily conversions | conversions/day | t / d | 2
out daily_budget | Suggested daily budget (to maintain pace) | currency | c * (t / d) | 2