I’ve recently been intrigued by how Bayesian testing allows Google to measure incrementality with just $5,000. It’s fascinating how this modern approach opens up new possibilities for advertisers.
Through these tests, advertisers like me can now explore lift measurement options without needing big enterprise budgets, as reported by Search Engine Land.
This change immediately raises an important question: How exactly does Google achieve accurate measurements of incrementality with significantly less data?
Previously, achieving reliable lift measurements demanded substantial budgets, lengthy test timelines, and the patience to handle inconclusive outcomes.
Given this context, Google’s claim of delivering precise results with merely $5,000 seems almost too good to be true. But it isn’t just marketing fluff; it’s a utilization of innovative mathematical models.
This transformation is powered by a testing methodology that emphasizes probability and learning, rather than aiming for absolute certainty.
Understanding this new approach is crucial for accurately interpreting these incremental results and for enhancing our PPC strategies.
Before we delve deeper, let’s quickly revisit some key Bayesian terms that marketers often encounter.
Glossary: Bayesian terms for search marketers
Prior: What we assume before the test begins.
Posterior: Updated belief after analyzing the data.
Credible interval: It shows the likely range of the result.
P-value: Frequency-based probability indication.
Traditional A/B testing, which most PPC advertisers know even if unknowingly, follows frequentist statistics.
These conventional A/B tests use metrics like p-values and fixed sample sizes to evaluate if changes reach statistical significance, often restricting smaller-budget tests.
In contrast, Bayesian testing veers away from this binary framework, instead asking, “Given all we know, how likely is this result to be true?”
Let’s see how Google legitimately manages to make $5,000 tests work effectively by embracing priors combined with its extensive data resources.
Google’s strategy rests on informed priors, hierarchically modeling, and probability assessments based on extensive campaign history.
This enables a competent analysis even with modest budgets, thus transforming limited data insights into actionable intelligence without averaging noise across campaigns.
Bayesian methods provide flexibility and adapt as more data is gathered, making them ideal for dynamic marketing environments, unlike their frequentist counterparts.
As more data rolls in, Bayesian tests evolve, relying increasingly on real results rather than priors, ensuring refined decision-making from smaller experiments to large-scale trials.
Using Bayesian inference, Google allows advertisers to derive directional insights without needing enormous budgets, effectively bridging gaps where frequentist testing falls short.
Takeaways for advertisers interested in Bayesian testing include understanding the diminishing role of priors as data accumulates, needing a discerning approach to interpreting outcomes.
To conclude, this mathematical ingenuity leverages Google’s vast data resources, offering a practical perspective over traditional methods, empowering PPC campaigns with more cerebral decision-making.
I recently discovered that Google has supercharged its Merchant Center with some noteworthy additions. If you’re like me, always on the lookout for ways to make your product listings pop, this update is exciting!
Google’s Product Studio is now equipped with three creative features that add flair to your product images. Previously, it was all about generating images, but now there’s so much more on offer.
What’s New: Imagine transforming your static product pictures into engaging short videos with just a few text prompts. Product Studio now makes it easy to do just that, perfect for creating eye-catching ads for social platforms.
Another cool feature is the one-click background removal. This tool is fantastic for making your product images look clean and professional, allowing products to stand out more vividly in Shopping visuals.
The third addition is really handy—enhancing image resolution. It lets us upscale older, lower-quality images to meet today’s high visual standards, ensuring our listings look their best.
Why We Care: High-quality images are crucial for boosting Shopping performance. However, creating and updating these assets has always required time and effort. These new features speed up the process and keep us from relying heavily on design teams.
The Big Picture: Google’s integration of AI-powered tools within Merchant Center is a game-changer. By making it easier to animate and enhance images, Google lowers the barriers to testing creative content—essential for maximizing campaigns.
What to Watch: For those of us with limited creative resources, these tools could be a massive time-saver. As Google pushes for more video-focused and visually enhanced ad formats, staying ahead with these updates will be vital.
First Seen: I came across this exciting update thanks to a post by Senior PPC Specialist, Vojtěch Audy.
For years, I’ve been fascinated by how PPC advertisers navigate the complexities of Google’s campaigns, especially Performance Max (PMax).
While the automation behind PMax is impressive, the lack of transparency has often been a source of frustration for me and many others.
Thankfully, Google has finally started to address some of these concerns with the introduction of the new Channel Performance report.
This guide is designed to help you understand the report, its benefits, and how you can leverage it effectively.
The Channel Performance report represents a major shift in how we can view and assess campaign performance.
Located under Campaigns > Insights and Reports > Channel Performance (beta), it’s a pre-built network report offering tabular and flow diagram data.
It’s currently exclusive to Performance Max campaigns but could potentially expand to other types in the future, hinting at a broader applicability.
Previously, getting insights into channel performance required tedious manual reports, or at best, third-party tools with limited capabilities.
Now, the Channel Performance report provides a direct, Google-native solution to this problem.
The report has two primary components: an account-level view and a campaign-level view, complete with a data table and a Sankey diagram.
The account-level view offers a new perspective with a convenient table displaying campaign and channel metrics, making it easier to analyze at a glance.
This view allows for sorting by different metrics, which is a handy way to compare and prioritize campaigns.
My favorite feature is the ability to switch segments, offering insights into ‘ads using product data’ versus ‘ads not using product data’, which was a significant challenge in understanding PMax campaigns.
Upon switching to the campaign-level view, you’ll notice a striking Sankey diagram that visualizes user interactions from impressions to conversions.
Though visually impressive, the data table below is more reliable for detailed analysis, showing performance metrics by channel and ad type.
For a deeper dive, I recommend exporting the data and using it in spreadsheets for comprehensive analysis.
However, the report has some drawbacks, like the misleading proportions in the Sankey diagram and lack of ratios in the data table.
Despite this, it offers valuable insights into which channels are genuinely delivering results, enabling you to maximize asset and traffic quality.
Utilizing placement data for quality control and customizing reports through Google Sheets can enhance your strategy.
Google has promised future features like API access, which will expand the report’s utility significantly.
As we continue to explore these insights, the challenge lies in accurately interpreting the data to make informed decisions.
I’ve been diving into some recent updates from Google regarding keyword match types, especially for those of us working with AI Overviews (AIO) and AI Mode ad placements. It’s crucial to understand these changes, particularly for those testing AI Max and using various match-type strategies. Let’s break it down so we can all optimize our ad reach effectively.
Why this matters to us. As the digital advertising landscape embraces AI-powered placements, it’s more important than ever to grasp which keywords are ready to serve ads and avoid unintentionally limiting our ad reach or misjudging performance metrics.
In May’s developments. When I followed the conversation between Marketing Director Yoav Eitani and Google’s Ads Liaison, Ginny Marvin, it was clarified that ads can serve either above or below an AI Overview—or appear within—but not in both placements simultaneously. Marvin stated, “Your ad could trigger to show either above/below AIO or within AIO, but not both at this time.”
When we talk about ad placements, it turns out both exact and broad match keywords can trigger ads above or below AIO. However, only broad match keywords (or those using keywordless targeting) have the privilege to appear within the AI Overviews.
What’s different now. In a later discussion with Paid Search specialist Toan Tran, Marvin provided further insight into Google’s updated eligibility criteria. Before this update, the presence of an exact match keyword could block a broad match keyword from filling AIO spots. But thanks to Google’s revisions, that’s no longer an issue.
Marvin detailed, “The presence of the same keyword in exact match will not prevent the broad match keyword from triggering an ad in an AI Overview, since the exact match keyword is not eligible to show Ads in AI Overviews and hence not competing with the broad match keyword.”
This adjustment means that with exact and phrase match keywords not qualifying for AI Overview placements, they won’t compete with broad match keywords in those auctions. So, a broad match can still trigger successfully even if its exact match counterpart is present.
The broader perspective. Google’s strategic update strengthens the distinction between traditional keyword matching and AI-powered intent matching. Ads in AI Overviews now depend on a keen understanding of both user queries and AI-generated content, requiring broader targeting signals.
The takeaway for us. If you, like me, are pushing into AI Max and AIO placements, it’s clear that broad match and keywordless strategies are key to tapping into Google’s AI-driven ad spaces. Exact and phrase match keywords might not appear in AI Overviews, but crucially, they won’t stop us from leveraging broad matches.
I’ve noticed that Google’s AI-powered bidding can truly be enticing. It promises to optimize my campaigns if I just feed it my conversion data and set a target, allowing me to focus on the bigger picture of strategy.
The idea is that machine learning will take care of everything else. But, what Google doesn’t really highlight is that its algorithms prioritize Google’s outcomes, which might not align with my goals.
As I delve into 2026, it’s clearer than ever that with Smart Bidding becoming more opaque and Performance Max absorbing more campaign types, discerning when to direct the algorithm—and when to take charge—has become an essential skill for exceptional PPC managers.
AI bidding can yield impressive results, but there’s also a risk of it undermining profitable campaigns by prioritizing volume over efficiency. The key isn’t in the technology itself but in knowing when the algorithm requires direction, tighter constraints, or a complete override.
This article will guide you through:
How AI bidding actually operates.
Recognizing the warning signs when it’s failing.
The intervention points where human judgment is crucial.
How AI Bidding Actually Works – And What Google Doesn’t Tell You
Smart Bidding offers various strategies, such as Target CPA, Target ROAS, Maximize Conversions, and Maximize Conversion Value. Each uses machine learning to predict conversion likelihood and adjusts bids in real time.
The algorithm evaluates numerous signals during auctions—device type, location, time of day, and more—to determine an optimal bid. During the “learning period” of typically seven to 14 days, the algorithm probes the bid landscape to understand the conversion probability curve.
Although Google advises patience during this phase, sometimes campaigns get stuck in perpetual learning and fail to stabilize.
Google’s Optimization Goals vs. Your Business Goals
The algorithm optimizes for metrics that increase Google’s revenue, which might not align with my profitability goals. For instance, setting a Target ROAS at 400% might prompt the system to maximize total conversion value, focusing on spending the full budget rather than understanding the varied nuances of my business.
My business goals might require a different approach, such as a specific volume threshold or maintaining varying margin requirements across products. The algorithm doesn’t account for these intricacies, like cash flow constraints.
Key Signals the Algorithm Can’t Understand
While AI bidding is effective, it has its limitations. Without intervention, several factors may go unaccounted for, like seasonal patterns, product margin differences, and changes in market conditions.
For example, the algorithm might not recognize the distinction between products with different profit margins. A $100 sale on Product A with a 60% margin is distinct from a sale on Product B with a 15% margin, yet the algorithm treats them equally, highlighting the need for profit tracking and margin-based segmentation.
Warning Signs Your AI Bidding Strategy Is Failing
The Perpetual Learning Phase
Extended learning periods are a major red flag. If my campaign’s “Learning” status persists for over two weeks, it indicates a problem. The causes could range from low conversion volume to frequent changes that reset the learning phase.
When to Intervene
Boost the budget to speed data collection.
Relax the target for higher conversions.
Switch to a less aggressive strategy, like Enhanced CPC.
Budget Pacing Issues
Healthy AI campaigns show smooth budget pacing. If I observe erratic patterns like front-loaded spending or consistent underspending, it signals a lack of algorithm confidence.
The Efficiency Cliff
This refers to when performance starts strong but then deteriorates. It’s usually visible in Target ROAS campaigns where, month after month, the ROAS declines as the algorithm exhausts efficient segments and expands into less qualified traffic.
Traffic Quality Deterioration
Even when metrics seem fine, qualitative signals might suggest otherwise. I might notice a drop in engagement or shifts in geographic targeting, indicating the algorithm is prioritizing cheaper clicks which don’t necessarily convert better.
The Search Terms Report Reveals the Truth
Regularly exporting the search terms report helps identify issues. I look for irrelevant expansions or low-intent queries that consume budget with little conversion value, such as a luxury retailer finding clicks for “free furniture donation pickup.”
Strategic Intervention Points: When and How to Take Control
Segmentation for Better Control
When it comes to AI bidding, a one-size-fits-all approach might not work for diverse business models. By segmenting my campaigns, I can tailor algorithms to meet specific goals, using separate campaigns for high- and low-margin products or different regional performances.
Bid Strategy Layering
Sometimes, a hybrid approach serves better. I might run a Target ROAS under normal conditions and adjust it manually during peak times to capture volume, or use Maximize Conversion Value with bid caps to honor unit cost constraints.
The Hybrid Approach
Pairing AI with manual campaigns can optimize effectiveness. Allocating a percentage of the budget to each allows for capturing valuable traffic through manual efforts while still leveraging AI for broader campaign management.
COGS and Cart Data Reporting (Plus Profit Optimization Beta)
Google now supports reporting cost of goods sold and cart data, allowing a clearer view of profitability within Ads reporting. Although still in testing, this feature could soon enable profit-focused bidding rather than revenue-focused, enhancing performance analysis.
AI bidding thrives under solid fundamentals, such as sufficient conversion volume and a stable business model with clear margins. In these contexts, AI often surpasses manual bidding by processing more variables than a human possibly could.
This tends to hold true for mature ecommerce accounts, stable lead generation programs, and SaaS models with predictable conversion paths.
Preparing for AI-First Advertising
As Google continues to simplify advertisement management through automation, my role has evolved from bid management to being an AI strategy director. My focus is now on setting clear goals, providing context, and intervening when needed.
Despite the reduction in advertiser control, certain strategic decisions remain human-driven, ones that require intelligence beyond what an algorithm alone can provide.
Master the Algorithm, Don’t Serve It
AI-powered bidding is a remarkable tool for optimization that delivers unparalleled results when conditions are optimal. However, the key lies in mastering it, ensuring that my business context informs the algorithm’s decisions, and knowing when to take control to align it with my strategic goals.
The strongest PPC leaders today are those who don’t just manage bids but helm the systems that manage them.
I’ve noticed a significant shift in Google Ads as they now allow us to optimize bidding for view-through conversions (VTC) in Android App campaigns. This change highlights a growing emphasis on video-driven performance.
In the past, VTC was a subtle, behind-the-scenes signal within Google’s system. Now, it’s a visible option that allows me to focus on conversions that occur after an ad is seen, rather than clicked.
The shift. It’s evident that Google is steering app advertising away from traditional click-focused strategies, encouraging an approach centered around influence and incremental results. This is particularly beneficial for platforms like YouTube and in-feed video advertising.
This update means our bidding strategies align more intuitively with the actual ways users discover and install apps today.
Why it matters to me. This flexibility allows me to go beyond mere clicks, enhancing measurement metrics for video-centric app campaigns. It’s an exciting validation for those of us invested in upper-funnel marketing activities.
Who benefits the most? Advertisers who prioritize video content and focus on creating awareness and engagement. This is a game-changer for teams oriented towards long-term growth, not just immediate installs.
What I’m keeping an eye on:
How Google’s attribution models affect campaign reliance
Potential shifts in Cost-Per-Acquisition expectations
The growing importance of creative quality over click-centric strategies
First seen by. I came across this update thanks to Rakshit Shetty, a Senior Performance Marketing Executive who first spotted this change.
Bottom line. Google is elevating view-based data for app campaigns to a priority status, marking a shift towards a performance marketing strategy led by AI and agnostic of sales funnels.
I recently discovered that uncontested ads might be silently eating away at my holiday budget. Even when I’m the sole bidder, my CPCs remain stubbornly high. Here’s how I began to reclaim those wasted dollars.
This holiday season, Google Search and Shopping Ads are projected to surpass a staggering $70 billion in spending. However, many advertisers, myself included, overlook a critical flaw in Google’s auction system that drains our funds—even in the absence of competitors.
The team at BrandPilot identifies this issue as the “Uncontested Google Ads Problem,” a significant yet often ignored source of wasted ad spend during peak times.
During SMX Next, I learned from John Beresford, the Chief Revenue Officer at BrandPilot, about a little-known quirk in Google’s auction logic. It’s fascinating how this can lead advertisers like me to overspend on our brand terms, shopping placements, and category keywords because Google doesn’t automatically lower our CPCs when no one else is bidding.
Instead of enjoying lower costs as the sole bidder, I found myself paying the same high rate as if competitors were still active. It’s a situation that unfolds thousands of times a day for major brands, and like me, many marketers don’t even realize it.
In John’s session, we explored:
Understanding why “competition gaps” are far more frequent than we think.
Discovering how uncontested moments can warp CPCs, even on brand keywords.
The potential of real-time auction visibility—and how AI is revolutionizing the field.
He also shared how advertisers are deftly reclaiming wasted spending and channeling it back into growth, without giving up impression share, traffic, or revenue.
Identify why CPCs are artificially high when competitors are missing.
Calculate the true financial impact of the Uncontested Ads Problem on your budget.
Execute AI-driven bidding and suppression strategies to avoid self-bidding and increase ROAS.
If you’re managing Google Search or Shopping campaigns this holiday season, this session is a must-see. Learn how to keep Google from sneaking off with your budget and start converting those savings into real performance improvements.
I’ve discovered that Google has quietly introduced a new feature in their Performance Max (PMax) campaigns, allowing advertisers like us to access video assets directly from the Merchant Center. This seemingly small adjustment is poised to make a significant impact on how we handle retail and e-commerce ads.
How it works. As part of this update, Google Ads now enables us to:
Auto-surface product-associated videos directly from Merchant Center during the PMax setup process.
Shorten creative workflows for our retail and e-commerce teams, saving us valuable time.
Improve product-to-creative alignment, thereby enhancing ad relevance.
Boost performance especially for those of us managing extensive SKU catalogs.
Why this matters. This update is a game-changer because it eliminates a key friction point within PMax: the challenge of integrating high-quality, product-relevant videos into our campaigns. By streamlining this process and pulling videos directly from the Merchant Center, Google is enhancing the connection between inventory and creative assets. This means higher ad relevance, greater engagement, and improved overall performance.
For brands like ours that have vast SKU inventories, this feature significantly accelerates the workflow and guarantees comprehensive video coverage — something we used to find challenging and resource-draining.
The bigger picture. It seems that Google is on a mission to expand PMax’s creative capabilities. From integrating social video imports to this new Merchant Center video feature, there’s a clear intention to make PMax more user-friendly for advertisers heavily involved in commerce.
First seen. This update caught my attention thanks to senior performance marketing executive, Rakshit Shetty, who shared his insights on LinkedIn.
The bottom line. Although it’s a subtle change, it’s undoubtedly a meaningful victory for brands operating at scale in the eCommerce space.
As someone who navigates the complexities of Google Ads, I know the mere mention of ‘Recommendations’ can send shivers down your spine. It’s like a pop-up that corners you on every platform screen—when you’re tweaking keywords, setting campaigns, or batching bids, even when you’re simply checking on things!
I’ve had countless emails from clients fretting over why their ‘Optimization Score’ has suddenly dipped. In this article, I want to demystify what Google Ads Recommendations really are, dispel some myths, and share some tactical advice on how to handle them.
Why do Google Ads Recommendations get such a bad rap?
So why this widespread disdain? To me, it’s plain: the expectations simply don’t align. While the system tailors Recommendations to our accounts, it often lacks the nuance needed for unique business goals.
The algorithm’s designed to spot patterns and suggests tweaks based on what’s working in other accounts. Say you only use Exact and Phrase match keywords—the system might suggest ‘Test Broad Match’ because, theoretically, it could broaden reach, but it may not align with budget constraints or niche specifics.
Bear in mind, Recommendations initially served as a tool for Google Ads sales reps to identify potential client improvements. In their hands, human insight ensured suggestions were relevant. Now, the human filter is absent, making Recommendations feel less tailored.
Is the Optimization Score really that important?
When Google tells you your Optimization Score is low, it’s tempting to perceive it as a failing report card. Many fall into the trap of blindly accepting every suggestion just to see that 100% score light up.
Let me be candid: resist the urge. This score doesn’t reflect performance but rather measures how actively you are reviewing recommendations. Dismissing a suggestion has the same impact on your score as applying it. So, keep your score at 100% if it’s crucial to your Google Partner status—but otherwise, let it slide.
Decoding Recommendations vs. Real Issues
Recommendations might pop up anywhere across the ad platform, not just the designated tab. You’ll see them during account setup, keyword addition, or bid adjustment. These prompts can set off alarms due to their visibility.
Remember, blue or yellow notifications are mere suggestions. Red or purple signals require immediate attention, potentially indicating a billing error or disapproved ad. Maintain a calm head, and only adopt changes that align with your objectives.
Are Recommendations just Google’s strategy to boost spending?
An argument often made is that Recommendations aim to skyrocket spending, subtly capturing more dollars. And sure, Google is profit-driven, but they understand you’ll curb spending if returns don’t justify the expense.
Suggestions are twofold: some aim at increasing reach and expenditure, while others focus on ROI and account refinements that might not increase costs but enhance efficiency.
Turn Off Auto-Apply Recommendations
It’s crucial to mention Auto-Apply Recommendations when discussing these aspects. It’s a feature Google previously championed, enabling automatic implementation of suggestions without checks. Thankfully, it’s losing focus now.
To take control, head to the Recommendations tab, switch to All Campaigns, click Auto-Apply Settings, and ensure all selections are unchecked. Keep the reins in your hands—Google doesn’t need unsupervised access to your budgets, bids, or keywords.
Recommendations aren’t inherently good or bad. They are mere prompts to evaluate and test. Listen to your instincts: review, test if promising, or move on if irrelevant.
This article is part of our ongoing Search Engine Land series, Everything you need to know about Google Ads in less than 3 minutes. In each edition, Jyll highlights a different Google Ads feature and how to maximize it efficiently.
I’ve just discovered a game-changer from Google that could simplify our advertising efforts significantly. Their new Data Manager API offers a streamlined way for us to feed our valuable first-party data directly into Google’s sophisticated AI systems.
As an advertiser, utilizing the Data Manager API means I can seamlessly connect our first-party data with Google’s AI-driven ad tools. This connection is poised to elevate our measurement, targeting, and overall performance, eliminating the hassle of managing multiple systems.
Why I care. By leveraging the Data Manager API, I’m able to inject high-quality data into Google’s AI, which optimizes targeting, measurement, and bidding processes. It replaces the need for various APIs, reducing our engineering workload and accelerating insights into our campaigns. With the decline of cookies, this API is crucial for maximizing the data we already have.
Driving the news. This API serves as a single integration point, unifying multiple Google platform APIs. It’s designed for advertisers, agencies, and developers, making our lives a lot easier.
Here’s what I can do with it:
Upload and refresh audience lists
Send offline conversions for improved measurement
Enhance bidding performance by providing Google AI with richer signals
This API expands upon Google’s existing codeless Data Manager tool, which is already in use by thousands of advertisers to activate first-party data.
Partnership push. To speed up adoption, Google is integrating with several partners, including AdSwerve, Customerlabs, Data Hash, and others.
State of play. Starting today, the API is available across Google Ads, Google Analytics, and Display & Video 360, with more integrations to follow.
The bottom line. Adopting the Data Manager API empowers us by enhancing Google’s AI capabilities, improving measurement, reducing technical complexities, and driving better ad performance, all while gearing up for a future that prioritizes privacy.
