Skip to content

Features
Pricing
Changelog
Status

Tag: Analytics

Discover the Best Open-Source MMM Tools for Your Needs
Marketing mix modeling (MMM) has become essential in today’s business landscape.

I’ve noticed how companies like Google, Meta, and Uber have opened doors with their free open-source MMM frameworks.

The challenge lies in figuring out which tool suits your needs without needing a PhD in statistics.

Understanding the Different Roles of Open-Source MMM Tools

These tools often get mentioned together but serve uniquely different purposes despite their collective fame.

Google’s Meridian and Meta’s Robyn are designed to deliver actionable insights by turning your marketing data into strategic budget recommendations.

These include valuable features like:
- Data transformations that capture advertising decay.
- Saturation curves to visualize diminishing returns.
- Dashboards and budget optimizers for spend allocation guidance.
Meanwhile, Uber’s Orbit and Facebook’s Prophet serve different needs.

Orbit is more about time-series forecasting and requires significant development to transform into an MMM tool.

Prophet acts as a forecasting component within other frameworks.

Think of it like transportation:

$```json { "alt": "Budget allocation analysis for Model ID 1_143_2 showing total optimization results, per channel allocation, and simulated response curves.", "caption": "Explore the strategic budget allocation insights for Model ID 1_143_2, highlighting total optimization results and per channel performance with detailed simulation data.", "description": "This image provides an in-depth analysis of budget allocation for Model ID 1_143_2. It presents total optimization results, showing initial and bounded allocations, and highlights ROAS and mROAS metrics. The image details budget allocation percentages across channels like Facebook, OOH, print, search, and TV. Simulated response curves illustrate the impact of spending on total responses, reflecting the effect of different budget scenarios. Designed with readability and analytical precision, this one-pager is crucial for strategic marketing decisions." } ```$
- Meridian and Robyn are ready-to-drive vehicles.
- Orbit is a high-performance engine requiring further builds.
- Prophet serves as the GPS navigation system within your car.
Dig deeper: Marketing attribution models: The pros and cons

Robyn: Making MMM Accessible and Powerful

I admire Meta for creating Robyn, a tool that breaks down barriers with its automation and ease of use.

Robyn turns weeks of model tuning into a quick data upload process, rapidly exploring thousands of configurations.

It stands out by providing multiple high-quality model solutions, catering to various business needs and risk levels.

Its capability to integrate real-world experimental outcomes boosts its credibility with decision-makers.

However, keep in mind that it assumes consistent marketing performance, which might not align with real-world dynamics.

Meridian: The Analytical Powerhouse

Meridian, representing Google’s advanced Bayesian approach, models the intricate mechanisms behind advertising effects.

Unlike Robyn’s practical strategies, Meridian focuses on answering what could happen with budget reallocation based on theoretical models.

This approach, especially its geo-level modeling, provides insights that are crucial for market-specific decisions.

$```json { "alt": "Graphs displaying channel contribution, spend and revenue by marketing channel, and a pie chart of contribution percentage.", "caption": "Analyze the impact of different marketing channels on revenue with these insightful charts. Discover which channels drive the most revenue and how spend correlates with returns.", "description": "This image features three main graphical representations: a horizontal bar chart showing channel contribution to revenue, a vertical bar chart detailing spend and revenue contribution by channel, and a pie chart illustrating contribution percentages. The bar chart highlights the 'Baseline' as the largest contributor with 79.3% of revenue, followed by Channel_4 and Channel_5. The pie chart indicates 79% revenue from the baseline and 21% from all channels. These visualizations provide a comprehensive overview of revenue distribution across various marketing initiatives, crucial for optimizing channel strategies. Keywords: revenue, marketing channels, pie chart, bar graph, channel contribution." } ```$

Although its methodology is robust, Meridian’s technical demands are high and require statistical expertise not common in most marketing teams.

Dig deeper: How Bayesian testing lets Google measure incrementality with $5,000

Uber Orbit: Expert Time-Series Forecasting

Orbit shines in time-series forecasting with its Bayesian time-varying coefficients, offering flexibility many traditional MMM tools lack.

It’s an advanced tool, best suited for teams that have the capacity for custom framework development.

Prophet: Unraveling Temporal Patterns

Prophet stands out in its primary role of time-series forecasting, effectively disentangling trends and seasonal influences from data.

Remember, while it can support MMM processes, it won’t serve as a standalone solution for attribution or optimization.

Dig deeper: MTA vs. MMM: Which marketing attribution model is right for you?

Making Informed Choices for Your Team

Your choice depends on your team’s statistical comfort and resource availability.
- Robyn suits most teams, offering rigorous insights with minimal setup time.
- Meridian is for those with the technical expertise to leverage its deeper capabilities.
- Orbit is ideal for custom framework developers.
- Prophet helps in preprocessing but isn’t a complete MMM solution on its own.
Choose a tool that your team can realistically implement and maintain, maximizing the benefit from its insights.

Dig deeper: How to avoid marketing mix modeling mistakes that derail results

Inspired by this post on Search Engine Land.
January 12, 2026
Crafting Year-End PPC Reports that Captivate Leadership

As the new year arrives, it’s my job to present an end-of-year (EOY) PPC report that truly reflects our performance.

EOY reports are not merely extended versions of our monthly check-ins. Instead, they cater to a different audience—mainly the leadership team, who need a broader narrative.

Executed well, these reports set the stage for the upcoming strategies, garnering buy-in and positioning me as a strategic ally rather than just a campaign overseer.

Here’s my approach for creating an impactful EOY PPC report that engages leadership and sets us on a successful path for the new year.

1. Understanding My Audience’s Priorities

Launching a new campaign without defined goals and target audiences is unheard of, and the same goes for my EOY report.

This year, my clients include diverse leadership teams—from those new to me wanting concise summary reports, to detail-oriented CEOs desiring a rich narrative.

Instead of a generic template, I tailor each report to fit the unique needs of each audience, avoiding confusion and maximizing satisfaction.

If you’re unsure of your audience, engage your primary contact to better understand the report’s recipients, their focal points, and decision-making goals.

2. Building a Clear Executive Summary

My executive summary’s role is to quickly provide leadership with an understanding of our PPC performance.

It’s the gateway that frames everything that follows, and though taught to write it last, I start with it to shape the report’s flow.

Focusing on Key KPIs

I prioritize metrics vital to my audience—be it revenue, leads, or conversions—ensuring these are front and center in my summary.

Providing Context with Benchmarks

By leveraging year-over-year performance, target achievements, and industry benchmarks, I ensure leadership comprehends our standing without needing to guess.

These benchmarks provide busy executives with an immediate grasp of our performance, priming them for deeper insights and actions to follow.

3. Diving into Performance Details

Here, I delve into the ‘why’ behind our performance, illuminating the strategies and decisions driving key outcomes.

Whether limited to pivotal insights or an in-depth analysis, my focus remains on information supporting the summary and informing our future direction.

Highlighting Best Performers and Resource Allocation

By showcasing top-performing assets and how we distributed efforts, I help leadership see where we’ve excelled and intelligently invested resources.

Reflecting on Tests and Trends

Sharing tests and trends that have shaped our year helps leadership understand the evolution of our strategy and sets the stage for potential opportunities.

4. Considering External Influences

It’s crucial to frame our performance within the wider environment, highlighting external factors that influenced results either positively or negatively.

An Analysis of Digital and Economic Factors

From shifts in digital marketing channels to broader macroeconomic trends, I contextualize performance against external events, explaining both impacts and non-impacts.

5. Planning for What’s Next

Looking ahead, I focus not on pre-determined paths, but on our decision-making framework, assuring leadership of a structured plan for adapting to future changes.

Outlining Next Steps and Innovations

By sharing strategic moves tied to last year’s data, as well as exploratory initiatives and adaptation strategies, I foster confidence and excitement for the year to come.

Finalizing with a Leadership Lens

Before submitting, I ensure all data is clearly sourced, negatives are addressed up front, and all stakeholder queries have been thoroughly answered.

This reflective practice not only strengthens my relationship with stakeholders but also lays the foundation for seamless reporting in the years ahead.

Inspired by this post on Search Engine Land.

January 6, 2026
Unlocking Incrementality with Bayesian Tests at a $5K Budget
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.

$```json { "alt": "Mathematical formula for Z-score involving proportions and sample sizes.", "caption": "Dive into statistics with this formula for calculating the Z-score from sample proportions. A fascinating glimpse into the world of data analysis!", "description": "This image displays a mathematical formula for calculating the Z-score based on the difference between two proportions, p2 and p1, over the standard error of the sample sizes, n1 and n2. This statistical formula is essential in hypothesis testing and helps determine how far apart proportions are in terms of standard deviation. Key elements include the square root, fraction, and parentheses, crucial in advanced statistics and data analysis." } ```$

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.

Inspired by this post on Search Engine Land.
December 18, 2025
Unlock the Power of Google’s PMax Channel Report

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.

Inspired by this post on Search Engine Land.

December 17, 2025
Rethink Your Dashboards: Beyond Click-Based Attribution
As someone deeply involved in marketing, I’ve seen how the explosion of marketing channels and touchpoints has made measuring success a truly strategic endeavor.

I’ve noticed that click-based attribution models—such as last-click and first-click—are still widely used as standard. Yet, as I delve deeper into these metrics, I realize they’re becoming less effective as standalone measures.

These models dominate executive dashboards, giving me pause because this reliance can impose significant limitations.

In my experience, click-based metrics can indeed be valuable for understanding digital interactions. However, it’s risky for executives to center major strategies and budget allocations solely around clicks, as this can lead to neglecting vital parts of the customer journey—parts that truly count.

In this article, I want to explore:
- What click-based attribution really captures.
- How it falls short in a complex, multi-channel world.
- The risks of over-relying on click metrics for business decisions.
- Alternative measurement approaches that better align marketing with actual business results.
- Ways marketing leaders, like myself, can guide executives toward more comprehensive outcome-focused frameworks.
My goal isn’t to dismiss clicks; they have their place. They should, however, provide context rather than serve as the core measure of success.

What Does Click-Based Attribution Actually Measure?

Click-based attribution tracks ad clicks and assigns conversion credit to the responsible marketing touchpoints. In my role, I observe that models vary—first-click, last-click, linear, time-decay, to name a few—but fundamentally, they all divide credit along the user journey differently.

Platforms tend to default to click-based models because clicks are straightforward to capture and report. However, their clarity can often mislead.

I’ve learned that click-based attribution hinges entirely on user interaction with tracking links. Without a click, or with delayed decisions, important touchpoints might be misattributed or entirely overlooked.

While this approach might work in simplistic funnels, today’s customer journeys are multi-device and multi-channel, quickly diminishing the value of clicks in context.

Dig deeper: The end of easy PPC attribution – and what to do next

The Problems with Solely Relying on Click-Based Attribution

When I examine today’s buyers, I see that they rarely follow neat, linear paths—an assumption made by click-based models.

Instead, buyers interact across many devices, channels, and may even engage through offline touchpoints. Consider social media, AI like ChatGPT, or brand recognition from videos, influencers, or website content.

Many valuable interactions go untracked by clicks, though they meaningfully influence buyer perception and conversion readiness.

Imagine a buyer: they watch a video on LinkedIn, then research your product through third-party reviews and your case studies on your website. Days later, they directly Google your brand and make a purchase.

In click-based systems, only the final branded search click would be credited, overlooking all previous touchpoints that educated and persuaded the customer.

Such blind spots aren’t trivial; they form a canyon between reality and measurement.
… (content continues in the same format) …

Inspired by this post on Search Engine Land.
December 15, 2025
Discover GPT-5.2 in ChatGPT: Unleash New AI Power in Your Dashboard

On December 11th, I was excited to learn about OpenAI’s announcement of GPT-5.2, their most advanced frontier model yet. Knowing how transformative AI can be, I was thrilled to see that Profound is now tracking GPT-5.2 responses in ChatGPT. This upgrade is integrated across our entire product suite, including Answer Engine Insights, Prompt Volumes, and Agent Analytics.

Starting today, every ChatGPT response on my dashboards reflects this cutting-edge model. It’s exhilarating to think about the enhanced capabilities and strategic insights that I can now access thanks to this update.

Inspired by this post on Try Profound Blog.

December 14, 2025
How Dark Themes Outperformed in Our B2B Landing Page Test

I have to admit, I was surprised when our dark-themed landing page outperformed the light one.

Everything I believed about conversion optimization suggested the light background would dominate.

Light themes are generally the norm for B2B lead generation due to their readability and clean look, aligning perfectly with accessibility standards.

Unbounce’s study of 41,000 landing pages backs up this trend for light backgrounds. Betting on the light theme seemed like a safe decision.

However, after dividing our paid traffic equally between a dark and a light landing page for our industrial fleet repair SaaS, despite a 16.62% higher CTR for the light variant, it resulted in 42% fewer conversions.

This isn’t a call for adopting dark themes universally.

Rather, it’s a case study showing how audience context and industry-specific associations can outweigh best practices drawn from broader samples.

We cater to a niche in the B2B SaaS market, particularly serving the transportation industry—businesses maintaining commercial vehicles and equipment.

Our intended audience includes shop owners and operators engaged in industrial settings, managing technicians, equipment, and demanding commercial clients.

Going into this test, my expectations were clear.

I anticipated light backgrounds would be more effective for our text-heavy lead generation pages, given their emphasis on whitespace and visual hierarchy. Our 7-field form aimed at busy shop operators seemed poised for success with light mode.

I also assumed blue CTAs would yield better results, with blue being associated with trust and security crucial for B2B software purchases. Thus, we used a blue CTA button.

I was incorrect on both fronts.

We conducted this test by isolating the visual design, directing traffic through Google Ads and Meta to two vastly different landing pages with identical copy.

The control page sported a dark theme with a black background, white text, high-contrast form fields, and a subtly outlined black CTA button. The header lacked a brand logo, intensifying the focus on the content.

Conversely, the treatment page featured a light theme, employing white and light gray elements, dark text, and a blue CTA button. Here, our brand logo was prominently placed in the header.

All other variables remained the same, emphasizing the importance of isolating design as the sole differentiating factor.

This test spanned three to four weeks, with Google Ads search campaigns topping $8,205.97, yielding 767 clicks and 30 conversions.

The light theme’s seemingly advantageous CTR masked the truth—it attracted less qualified traffic, converting at a similar or worse rate than expected.

A consistent preference for the dark theme also emerged in Meta tests, reinforcing the role of audience preference rather than algorithmic anomaly.

Understanding why the dark theme won lies in recognizing how it aligns with the psychological and environmental cues of our target audience in the industrial sector.

The dark theme resonated well with the familiarity of industrial aesthetics—functional, robust environments characterized by dark, metallic tones.

The contrast provided by white form fields on a dark background was unmistakable, drawing eye attention naturally.

Dark themes carry a tone of seriousness and value, fitting for the weighty decision-making expected in B2B software acquisitions.

Moreover, embracing familiar industry conventions, the dark interface enhanced trustworthiness and familiarity.

This test taught me that testing design psychology is just as crucial as testing visual elements themselves. Before embarking on similar experiments, consider what your design communicates to your audience rather than just aesthetic appeal.

Finally, ensure your experiments include significant contrast between variations while keeping other elements constant to draw accurate conclusions.

Audience context should guide optimization efforts more than generalized best practices. By focusing on specific audience needs and signals, I’ve learned that real, lasting optimization success can be achieved.

Inspired by this post on Search Engine Land.

December 11, 2025
AI Shopping Insights: Google vs. ChatGPT Citation Trends
I recently came across some eye-opening data highlighting the distinct approaches Google AI and ChatGPT take in citing sources when it comes to retail information. While Google mentions retailers only 4% of the time, ChatGPT cites them 36% of the time. This significant gap of nearly nine times suggests that each platform guides shoppers in noticeably different directions, and this insight comes from the latest BrightEdge data.

Why is this important to us? Nowadays, millions of shoppers are relying on AI to discover deals and gift ideas. However, the process differs greatly between the top AI search platforms. Google tends to focus on what users are saying, while ChatGPT zeroes in on where you can actually purchase items.

Regarding what each AI prioritizes, Google AI Overviews are inclined to reference YouTube reviews, Reddit discussions, and various editorial sites. In contrast, ChatGPT frequently cites retail giants such as Amazon, Walmart, Target, and Best Buy.

Let’s break down the priorities further. Google AI Overviews tend to cite:
- YouTube reviewers and unboxings.
- Reddit threads and community consensus.
- Editorial reviews and category experts.
Meanwhile, ChatGPT emphasizes:
- Major retailer listings.
- Brand and manufacturer product pages.
- Editorial sources (secondary).
This citation divide is quite telling. On Google, retailers show up only about 4% of the time, as it leans more towards user-generated content and expert reviews—acting more as a research tool rather than a purchase assistant. Top reference sources include:
- YouTube
- Reddit
- Quora
- Editorial sites like CNET, The Spruce Eats, and Wirecutter
Conversely, ChatGPT features retailers about 36% of the time, functioning as both an explainer and a shopping assistant, hence why retailer links are far more prevalent. Key sources often cited include:
- Amazon
- Target
- Walmart
- Home Depot
- Best Buy
About the data: BrightEdge scrutinized tens of thousands of e-commerce prompts across Google AI Overviews and ChatGPT during the 2025 holiday season, identifying and categorizing citation sources. Domains were sorted by type—retailer, UGC/social, editorial, and brand—and directly compared using identical prompts.

The detailed report is available here: Who Does AI Trust When You Search for Deals? Google vs. ChatGPT Citation Patterns Reveal Different Shopping Philosophies

Inspired by this post on Search Engine Land.
December 9, 2025
Unlock Google Data Manager API for Enhanced Ad Performance
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.

Inspired by this post on Search Engine Land.
December 9, 2025
Unlock Insight with Google Search Console’s New Reports

I’m thrilled about the latest update from Google Search Console, which now offers both weekly and monthly views in their performance reports. This change allows me to dive deeper into the data, providing a more detailed analysis beyond the usual 24-hour view.

What it looks like. At the recent Google Search Central event in Zurich, I snapped some photos of the announcement. It was a fantastic experience to see these changes unveiled in real-time.

Why we care. These updates, though small, are invaluable for SEOs, publishers, and site owners like me. The granular data now available helps me investigate changes in performance more effectively, whether it’s over a specific month, week, or day.

Inspired by this post on Search Engine Land.

December 9, 2025

←Previous Page

1 … 4 5 6 7 8