What is the New Target ROAS Conversion Bid Strategy and How Has it Improved?
Target ROAS (Return on Ad Spend) is a conversion-based bidding strategy designed to help advertisers scale as much as possible while maintaining a defined ROAS goal.
With the new Target ROAS, budget is no longer the primary lever for scaling. Instead, the system automatically adjusts spend based on performance and data availability, allowing advertisers to focus on setting a realistic ROAS target and letting the algorithm optimize delivery.
What’s new with Target ROAS?
The new Target ROAS introduces a smarter and more automated way to scale campaigns:
- You no longer need to constantly increase or decrease budgets based on performance
- The system dynamically adjusts the desired daily spend to balance scale and efficiency
- Bidding decisions are driven by predicted conversion value at the auction level
This allows campaigns to grow naturally when performance is strong and slow down when efficiency is at risk.
How does Target ROAS work?
Target ROAS optimizes campaigns using three core mechanisms:
1. Desired spend adjustment
Each day, the system recalculates how much the campaign should spend based on:
- How close actual ROAS is to the Target ROAS
- The amount of recent conversion data available
If performance is far from the target, desired spend is reduced.
If there is not enough data to make reliable predictions, desired spend may increase to gather more conversions.
The campaign will never spend more than the daily budget you set.
2. ROAS evaluation over time
Performance is evaluated over a rolling 7-day window, with more recent results weighted more heavily. This helps smooth short-term fluctuations while still reacting quickly to changes in performance.
3. Auction-level value prediction
Once sufficient conversion volume is reached, the system predicts the expected value of each auction using signals such as:
- Placement and context
- Platform and device
- Browser and operating system
If an auction is expected to deliver higher value, the system bids more aggressively.
If expected value is low, bids are reduced or skipped.
After the learning period
After the learning phase, Target ROAS becomes more stable and precise. With enough historical data, the algorithm:
- Makes more confident spend decisions
- Applies tighter bid adjustments
- Scales delivery while consistently protecting the ROAS goal
At this stage, advertisers typically see smoother scaling and fewer performance swings, with less need for manual intervention.
Budgeting and scaling behaviour
- Campaigns start with an initial spend level designed to generate enough conversions for learning
- Once stable, spend can increase gradually when performance meets or exceeds the ROAS goal
- If performance remains strong, daily spend may increase by up to ~50% day over day
- Expected spend is not shown, as it can change throughout the day based on performance signals
Best practices for Target ROAS
To get the best results:
- Set a high enough daily budget to allow the system to scale when performance is strong
- Avoid frequently changing budgets, as this can disrupt learning
- Ensure enough conversion volume to support prediction (ideally ~20 conversions per day)
- Avoid constantly changing your Target ROAS, as this resets spend calculations and may cause volatility
If you are happy with results but want more scale, lowering your Target ROAS can help unlock additional volume.
If efficiency is the priority and ROAS is not being met, a different bidding strategy may be more suitable.
What are the main benefits?
- Automated scaling without manual budget management
- Smarter bidding based on predicted conversion value
- More stable performance over time
- Reduced risk of overspending when performance drops
When should I use Target ROAS?
Target ROAS is best suited for advertisers who:
- Have consistent conversion volume
- Want to scale while protecting return
- Prefer automated, performance-driven optimization