Hitting the sweet spot; being in the zone; being right on target — call it what you will, but when done correctly, proper bid optimization is the best way to ensure that your online ads get served the most number of times, to the right users, and at the best price. Here’s how it works.
What Is Bidding Optimization?
Bidding optimization in online advertising involves adjusting how much you bid for ad placements (and in turn pay for placements when your bid wins). This helps you get the best possible results, such as clicks, conversions, or impressions, all while staying within your predetermined budget.
Once set up, most bid optimization happens automatically, without the need for constant human intervention on the part of the marketer. But, it only works effectively if a person or a team of people have carefully put it in place.
Where Is It Used?
Programmatic Advertising
Programmatic advertising is the automated, real-time buying and selling of online ads. A demand-side platform (DSP) is the software platform advertisers use to manage these automated ad purchases. DSPs connect advertisers with available publisher ad inventory through real-time bidding (RTB) auctions, allowing them to target specific audiences, place ads across various channels, optimize campaigns in real-time, and manage their advertising spend more efficiently. The ads you see on a site you’re visiting that appeal directly to your interests are an example of programmatic advertising.
See how Minor Hotels achieves 5X ROAS using Maximize Conversion bidding strategy. Realize.
Search Engine Marketing
You’ve no doubt seen the ads served atop search engine results pages (SERPs). Bid optimization in search engine marketing (SEM) involves strategically adjusting bids for ad placements to maximize return on investment (ROI) and achieve specific marketing goals. Those goals may include generating conversions or increasing brand awareness by using data and algorithms to determine the most effective bid amounts for keywords or other ad targets, ensuring ads appear in prime positions without you overpaying for the placements.
E-commerce Platforms
Placing ads on platforms where people are already primed to spend money can be a great way to use your ad dollars in an impactful way. Bid optimization on e-commerce platforms like Amazon involves strategically adjusting your advertising bids — or letting them be adjusted automatically — to achieve specific goals. This process involves dynamically changing how much you’re willing to pay for clicks or ad placements based on data such as product performance, audience behavior, and real-time competition within ad auctions.
Pros and Cons of Bidding Optimization
As with all aspects of marketing, there are upsides and downsides to the process of trying to squeeze every last bit of value out of your online ad spend.
| Pros | Cons |
| Hands-off real-time optimization. | It might take you a while to notice if your optimization tool is not performing as desired. |
| Ideal return on ad spend and budget management. | The learning period still costs time and money while you find that sweet spot. |
| Detailed data collection. | Harder to change course quickly. |
Advantages
Hands-Off Real-Time Adjustments
The single biggest benefit of automated bid optimization in advertising is arguably the fact that it frees advertisers up to focus on other work. When ads are being placed based on calculations run by machines, humans can be developing new ad campaigns, or they can work on myriad other aspects of their business.
Ideal Return on Ad Spend
When you know exactly what you’re going to spend on a given ad campaign, the element of surprise is removed and you can plan your budget and manage your expenses. If you’re happy with your ROAS (return on ad spend), you can let things ride; if not, you can make changes.
Detailed Data Collection
Savvy advertisers collect much more than just revenue off the ads they place — they also collect information. The more you know about how much ads are costing you, the population to whom they are being served, and what’s working well (or not so well), the better you can plan for future ad campaigns.
Disadvantages
Less Immediate Visibility
The self-driving car or automated train is the pinnacle of safety and convenience — until there’s a glitch, followed by a crash. While issues in automated bid optimization may not be as dramatic as a multi-car pileup, it also means you may not notice at first that things have gone awry, leading to greater losses.
Money and Effort Spent During Learning Period
One of the most amazing things about our computing machines these days is their capacity to learn about and improve how they handle the tasks we give them. But, peak algorithmic performance doesn’t emerge right away, and as with human testing, there’s still a price to pay during that learning period.
Difficulty in Changing Course
If you detect issues with an automated bidding optimization process, you can, of course, get things back on track, but it may not be as fast and/or straightforward as when doing things manually.
Bidding Optimization Best Practices
Setting a Core Strategy
Building a core bid optimization strategy involves aligning your bidding approach with specific marketing goals, using data to improve campaign performance, and maximizing ROI. To decide which bidding strategy you should use for a campaign, align your strategy with your primary campaign goal. You can select goals like targeting cost per action (CPA) or maximizing conversions.
Deciding which bidding strategy to use for a campaign can seem overwhelming, but if you focus on target CPA, target ROAS, maximized conversions, or manual cost-per-click (CPC), you can get down to the granular work.
To determine the right target CPA or target ROAS, first ensure you are working with accurate conversion tracking and data. For determining the best target CPA, set a target that aligns with your overall profit margin and business goals — Google Ads can provide a recommended starting point based on historical data. For target ROAS, which focuses on conversion value (revenue, e.g.), calculate your target by analyzing your historical conversion value and setting a target at (or below) that historical rate to ensure safe profitability. Start with a less aggressive target and gradually adjust based on performance data, especially if you’re starting off without much data. You can take a more aggressive approach later.
As for choosing a bid strategy — whether for clicks, impressions, conversions, or something else — you need to think in both the short- and long-term: Will it be more valuable to spend money and effort on building brand awareness now, then creating conversion opportunities later, or do you need to start generating sales right away, even if the initial pool of potential customers will be smaller?
Realize analyzes historical campaign data (impressions, clicks, conversions, cost per conversion, etc.) to identify which bidding strategies have performed best in similar contexts, and can help with these processes and decisions.
Make smart, data-driven decisions on your marketing campaigns. Realize.
Leveraging Data for Optimization
Leveraging data is essential for optimizing ad bidding, moving your efforts beyond manual adjustments toward creating more precise, efficient, and profitable campaigns. This approach relies on collecting and analyzing vast amounts of data — both historical and real-time information — to better predict user behavior, refine audience targeting, and inform automated bidding algorithms.
To determine the right target CPA or ideal ROAS to grow your revenues (once you have established things are working!) again analyze your historical campaign data to understand your average CPA or ROAS, then set the target slightly above the average, in order to maintain profitability and account for likely fluctuations, and increase cash flows. Choose a target ROAS for campaigns with varied conversion values, focusing on revenue generation. Realize uses both past campaign data and industry benchmarks to suggest realistic and competitive CPA or ROAS targets.
To improve your bidding strategy by leveraging historical data, ensure accurate conversion tracking and assign differentiated conversion values to specific actions customers have taken — and can take again. Then, use this data to set bid strategy targets based on past performance, leveraging value-based strategies like target ROAS to guide your bidding and, ideally, to achieve your business goals.
To use bid adjustments effectively, first you need to analyze performance data for each factor (device, location, time of day, audience, and more) and identify trends — and spot areas ripe for improvement. Then apply percentage-based bid increases or decreases to improve performance for high-converting segments and reduce investment in underperforming ones. Sometimes, you may need to do deep dives, even studying the performance of specific keywords, changing them slightly, and seeing if your ROI improves. You can analyze a search term report and identify new opportunities for bidding on new or altered keywords, or you can add negative keywords that you don’t want affecting performance.
Automated Bidding
At some point, you have to take your hand off the proverbial tiller and let the algorithms have a go at ad bidding optimization. This can be nerve-racking in the early phases of a campaign: For a new Google Ads campaign with little conversion data, for example, the best bidding strategy is to start with manual cost-per-click inputs to gather initial data and control costs, then transition to a smart bidding strategy only once you have enough conversions to train an algorithm. (Realize often recommends manual CPC for new campaigns until sufficient conversion data is collected.)
While “smart” bidding can use even minimal available data, manual (AKA human) CPC provides the necessary control to establish baseline performance and an understanding of your audience before you allow the artificial intelligence (AI) to take over and optimize for conversions. To “feed” your smart bidding strategy with high-quality data, ensure accurate and robust conversion tracking, make sure you have sufficient conversion data and budget, and implement value-based bidding to achieve a complete picture of your business goals.
Then, refine your audience targeting, use portfolio bid strategies to group campaigns, and consider adding ad extensions to increase interactions. Finally, allow sufficient time for the algorithms to learn, and always monitor performance and make adjustments as needed.
If you did all that and you’re still having issues with automated bidding, that’s not unique to you! Common reasons for poor smart bidding performance include insufficient data or budget, excessive campaign changes during the machine learning phase, unrealistic targets (like a CPA or ROAS set too high for realistic expectations), and inaccurate targeting or settings. To fix these issues, allow sufficient time for the bidding algorithm to learn, provide a larger budget to gather more data or narrow your focus, avoid frequent adjustments to settings, set realistic performance goals, and try to verify that your audience targeting and language settings are authentic to your products, platform, or services.
To balance “broad match keywords” (which allows for variations on wording and phrasing) with automated bidding, focus on negative keywords to maintain efficiency and accuracy and monitor search query reports to identify opportunities. Start new broad match campaigns with a smaller budget and leverage signals like audience lists and contextual information. Use automated bidding to optimize bids in real time, seeking out new, higher-performing search terms while you refine your negative keyword list to help ensure campaign relevance.
If you’re interested in setting up an ad experiment for testing a new bidding strategy, navigate to the “Experiments” section in your Google Ads account, select “Custom experiment,” and choose your base campaign to create a duplicate. In the Experiment settings, modify the campaign to test your chosen bidding strategy, then schedule the experiment with a traffic split, define your success goals, and set the duration. After the experiment runs, review the results in Google Ads to decide whether to apply the new bidding strategy to a base ad campaign.
Budget and Scalability
Scaling your marketing budget without hurting your CPA or ROAS requires a gradual, data-driven approach. A sudden, large budget increase can force advertising algorithms to target less qualified audiences, which raises costs and diminishes returns. Don’t plan for major scaling until you have enough data to account for outliers.
To balance your ad spend for consistent performance, determine your overall monthly budget. Then, calculate your ideal targeted daily spend by dividing the monthly total by the number of days in the month — or, the number of days on which your ads will run. Leverage platform features for overspending on days with higher traffic and monitor key performance metrics (KPIs) like clicks, conversions, and ROI in real time to make informed budget adjustments. Finally, use historical data (and take into consideration holidays and seasonality) to adjust budgets for periods of increased or decreased consumer activity, to ensure steady performance across the month.
Key Takeaways
Bidding optimization in online advertising is an ongoing process of adjusting how much you are willing to spend for ad placements to get the best possible results. Those results come while you stay within the budget you have determined, and largely without any hands-on effort once you dial the ads in via algorithm, thanks to machine learning.
Frequently Asked Questions (FAQs)
How do I leverage first-party data, like website visitors, email lists, and CRM data, to create targeted audience segments for real-time bidding?
Advertisers best leverage first-party data for real-time bidding (RTB) by segmenting audiences based on behavior and demographics, thereby creating customized, personalized ad campaigns. Marketers use customer relationship management (CRM) data for targeting, website visitor data for remarketing and lookalike audiences, and email lists to build detailed audience profiles. Integrating this data into ad platforms allows for more precise targeting, optimizing bids, and improving overall campaign performance and ROI. Realize can help greatly here thanks to its Predictive Audience Targeting, which was built for performance predicated on behavior rather than just identity.
What is the best strategy for A/B testing different bidding models and creative assets on a massive scale without compromising campaign performance?
To perform larger, effective-scale A/B testing without compromising campaign performance, adopt a tiered experimentation framework that isolates variables, prioritizes incrementality testing, and leverages automation. This approach mitigates risks of inaccurate or inactionable data by testing significant changes on a smaller and then ever-increasing scale before broader implementation, ensuring that optimization is continuous and data-driven, and not a waste of time, effort, and money. Regular check-ins and strategic adjustments will allow you to maintain consistency throughout the month. Realize can detect when performance could be improved with a different bidding strategy and may recommend setting up a test campaign with a different bid type (e.g., manual CPC vs. target CPA) or a different budget or CPA target.
Beyond standard post-click conversions, how can marketers best measure the incremental impact and true return on ad spend (ROAS) of open web campaigns?
To measure the incremental impact and true ROAS of an open web campaign beyond those standard post-click conversions, marketers can use controlled experiments and advanced modeling techniques. These methods isolate the causal effect of advertising by comparing outcomes in exposed and unexposed groups, offering a more accurate view of true ad effectiveness. To scale your advertising budget without negatively impacting your CPA or ROAS, you must take a gradual, data-driven approach: Abruptly and aggressively increasing your spending can disrupt the ad platform’s algorithm, forcing it to find lower-quality audiences and causing your CPA to rise and your ROAS to fall. Realize helps marketers use specific and even customized tools for analysis rather than relying on more “generalist” tools, thereby helping produce more granular information and driving better results.
