How to Optimize Bidding Strategies in Real-Time Bidding
Real-time bidding (RTB) has transformed how advertisers buy digital media. Instead of negotiating fixed rates for ad placements, RTB lets you bid on individual impressions in milliseconds. But with speed comes complexity – and the potential to waste budget if your bidding strategy isn't optimized.
In this guide, we'll walk you through the essentials of optimizing your RTB bidding strategies to improve campaign performance, reduce costs, and achieve better returns on investment.
Understanding Real-Time Bidding Basics
Before optimizing, it's worth refreshing what RTB actually does.
When a user visits a website, their impression becomes available in an ad exchange. Your demand-side platform (DSP) receives data about that impression in real time – including user information, page context, device type, and more. Your DSP then decides whether to bid, and if so, how much to bid.
This happens in under 100 milliseconds. Your bidding strategy is the algorithm that makes these rapid-fire decisions.
Why Bidding Strategy Matters
A poorly optimized bidding strategy can lead to: - Overpaying for low-quality or irrelevant impressions - Underbidding and losing valuable inventory - Inefficient budget spend across poor-performing placements - Wasted time manually adjusting bids instead of focusing on strategy
The right approach helps you bid confidently, allocate budget intelligently, and scale campaigns profitably.
Step 1: Define Clear Campaign Goals and KPIs
Your bidding strategy should align directly with what you're trying to achieve.
Identify Your Primary Objective
Common RTB goals include: - Cost per click (CPC): You want clicks at a target price - Cost per acquisition (CPA): You want conversions at a target cost - Return on ad spend (ROAS): You want revenue relative to spending - Cost per mille (CPM): You want impressions at a target price per thousand - Viewability: You want ads seen by users, not just served
Example: An e-commerce brand might optimize for CPA of £25, while a brand awareness campaign optimizes for CPM at £5.
Set Realistic Targets
Your targets should be based on: - Historical campaign data (if you have it) - Industry benchmarks for your sector - Competitive landscape research - Your business margins and profitability thresholds
Setting a CPA target lower than what your data supports will cause your bidding strategy to consistently underbid, losing good opportunities.
Step 2: Choose the Right Bidding Strategy Type
Most DSPs offer several bidding strategy options. Understanding each helps you pick what's best for your goal.
Manual CPM Bidding
You set a fixed maximum price per thousand impressions. Your DSP bids that amount on all eligible impressions.
Best for: - Campaigns focused on reach and frequency - Brand awareness campaigns - Testing new audiences where you don't have conversion data yet
Pros: Simple, predictable spend Cons: Doesn't account for impression quality or likelihood to convert
Automated Bidding (Smart Bidding)
Your DSP uses machine learning to adjust bids based on predicted performance. Common types include:
Maximize Conversions: The platform bids to get as many conversions as possible within your budget.
Target CPA: The platform automatically adjusts bids to hit your target cost per acquisition.
Target ROAS: The platform optimizes bids to achieve your target return on ad spend.
Best for: - Campaigns with strong conversion tracking - Campaigns at scale (thousands of conversions monthly) - Advertisers wanting to hand off optimization to algorithms
Pros: Leverages machine learning, adapts to market changes, often outperforms manual bidding Cons: Requires historical data, less transparent control, needs proper tracking setup
Impression-Level Bid Optimization
Your DSP evaluates each impression's attributes and predicts its value before bidding.
Best for: Campaigns where audience quality varies significantly
Example: A B2B SaaS company might bid higher on impressions from high-intent keywords and lower on brand awareness placements.
Step 3: Set Up Accurate Conversion Tracking
Optimized bidding relies on reliable data. If your conversion tracking is broken, your strategy is flying blind.
Implement Tracking Pixels
Work with your DSP to place conversion pixels on: - Thank you/confirmation pages (post-purchase or form submission) - Account creation pages - Key engagement events (depending on your goal)
Use Server-Side Tracking When Possible
Server-side tracking (sending conversion data directly from your server to the DSP) is more reliable than pixel-based tracking, especially post-iOS privacy changes.
Validate Your Data
Regularly audit your conversion tracking: - Compare DSP-reported conversions to your website analytics - Check for pixel firing delays - Ensure you're capturing all relevant conversions
Common pitfall: A client discovered their conversion pixels weren't firing for 30% of transactions due to ad blocker interference. Once fixed, their CPA targets became realistic.
Step 4: Segment Audiences and Bid Accordingly
Not all users are equally valuable. Smarter segmentation allows smarter bidding.
Create Audience Tiers
Tier 1 - High Intent: Previous customers, website visitors in the last 30 days, users matching your ideal customer profile - Bid strategy: Higher bids, broader inventory - Rationale: Closer to conversion, higher lifetime value
Tier 2 - Medium Intent: Lookalike audiences, users interested in competitor brands, relevant contextual placements - Bid strategy: Moderate bids, quality-focused - Rationale: Potential to convert with proper messaging
Tier 3 - Prospecting: Cold audiences, broad contextual targeting - Bid strategy: Conservative bids, high reach - Rationale: Unknown conversion likelihood, test new markets
Use Bid Modifiers
Adjust bids based on impression attributes: - Device: Bid 20% higher on mobile if mobile converts better - Time of day: Bid 30% lower at 3 AM if users don't convert then - Placement quality: Bid 15% lower on low-viewability sites - Geography: Bid higher in high-performing regions
Step 5: Implement Frequency Capping
Showing the same user too many ads wastes budget and can harm brand perception.
Set Reasonable Frequency Caps
- Brand awareness campaigns: 3-5 impressions per user per day
- Conversion-focused: 5-10 impressions per user per day
- Retargeting: 10-15 impressions per user per week
These are starting points – test to find your optimal level.
Monitor Frequency
Check your campaign reports for average frequency. If it's below 1.5, many users see your ad just once. If it's above 8, you may be over-exposed.
Step 6: Optimize Bids Based on Performance Data
Once campaigns are running, you'll gather performance data. Use it to refine your strategy.
Monitor Key Metrics Weekly
- Conversion rate by placement: Which sites/apps convert best?
- Conversion rate by audience: Which segments perform well?
- CPA by bid amount: Is higher bidding yielding better conversions?
- CTR by creative: Which ads drive clicks leading to conversions?
Make Data-Driven Bid Adjustments
If your CPA is above target: - Lower bids on underperforming placements - Reduce bids on low-intent audiences - Pause non-converting creative variations - Increase bid modifiers for top-performing segments
If you're not spending your budget: - Raise bids gradually (10-15% increments) - Expand audience targeting - Relax placement restrictions - Increase daily budget limits if available
Example: A fashion retailer noticed their CPA was £45 (target: £30). Analysis showed that users from one publisher had a CPA of £65, while another had £22. They cut bids to the first by 30% and increased them to the second by 20%, bringing overall CPA to £32 within two weeks.
Use A/B Testing
Test bid changes on a portion of budget first: - Run two campaigns with identical setup but different bid amounts - Let them run for 2-4 weeks (enough volume to be statistically meaningful) - Analyze results before rolling out
Step 7: Leverage Dayparting
Bidding behavior can vary significantly by day and time.
Analyze Performance by Daypart
Review conversions and CPA by: - Hour of day - Day of week - Season/holiday
Adjust Bids by Daypart
If evening conversions have 40% lower CPA than midday: - Increase bids by 20-30% for evening hours - Decrease bids by 20-30% for midday hours
This ensures your budget flows to high-performing windows.
Step 8: Monitor Competitive Landscape
Your bidding happens in an auction. Understanding competition helps.
Watch for Market Changes
- Are CPMs rising? Other advertisers may be increasing budgets
- Did a competitor launch a campaign? They might be bidding aggressively
- Is there seasonal demand? Holiday periods often see higher competition
Adjust Your Strategy Accordingly
If CPMs spike: - Consider whether your target ROI is still achievable - Focus on higher-converting audiences where you have advantages - Test new placements or audience segments with lower competition
Common Pitfalls to Avoid
1. Over-Relying on Automated Bidding Without Monitoring
Machine learning is powerful, but not magical. Set it up correctly, but review performance weekly.
2. Setting Unrealistic Target CPA
If your historical CPA is £40 but you set a target of £20, your algorithm won't spend budget. Base targets on real data.
3. Ignoring Viewability
A cheap impression that no one sees wastes money. Prioritize quality publishers and viewability standards.
4. Not Accounting for Attribution Delays
Some conversions happen days after the initial ad impression. If you optimize daily, you'll make decisions with incomplete data. Allow at least 3-7 days before judging performance.
5. Changing Too Many Variables at Once
If you adjust bids, pause placements, and shift budgets simultaneously, you won't know what caused changes in performance. Make one change at a time, wait for data, then iterate.
Key Takeaways
- Start with clear goals: Your bidding strategy should directly support your campaign objective
- Use the right strategy type: Manual CPM for brand awareness, automated bidding for conversions
- Build on good data: Accurate conversion tracking is essential
- Segment smartly: Different audiences deserve different bids
- Monitor and adjust: Performance optimization is ongoing, not one-time
- Test incrementally: Make small changes, measure impact, scale what works
Optimized RTB bidding isn't about complicated formulas – it's about making smart decisions with good data. Start with these foundations, track what works, and continually refine your approach. Over time, you'll develop intuition for what drives results in your specific market.