HubSpot Lost 80% of Its Traffic, CMO Kipp Bodnar Explains Why Revenue Kept Growing

For years, the gold standard of digital marketing was simple: more traffic equals more growth. We built entire careers around chasing organic sessions, treating the Google Search Console graph as the ultimate scoreboard for success. But when the scoreboard breaks, you are forced to ask if you were measuring the right thing in the first place. The same pattern also shows up in 000x Human Traffic in 5 Years, where the practical question is how the signal becomes visible.

HubSpot recently provided a stark case study in this shift. Their blog once peaked at over 20 million organic visits per month, only to see about 80% of that traffic vanish as AI generated answers began dominating the search landscape. Yet, despite this collapse in volume, their revenue continued to climb. This disconnect suggests that the relationship between search traffic and business growth has fundamentally changed. This connects with Organic Traffic Is Still Worth Tracking when the same signal needs a clearer operating decision. A useful companion note is What Ziff Davis’ Jonathan Berthold Tests First, because it looks at a nearby part of the same system.

The Era of 'Get Found Using Google' is Over

To understand where we are, it helps to look at where we started. HubSpot essentially codified the modern inbound playbook in their 2009 book, which explicitly told marketers to get found using Google, social media, and blogs. The logic was linear: create the best answer to a query, Google rewards you with traffic, and that traffic converts into customers.

AI has effectively severed that chain. When a user asks Claude or ChatGPT a question, they often receive a tailored, complete answer without ever needing to click through to a website. Kipp Bodnar, HubSpot's CMO, noted that the rollout of Google's AI Overviews alone wiped out nearly 5 million visits in a single 30 day window.

The realization here is that the mechanism of discovery has changed, even if the need for information hasn't. If your primary KPI is still "sessions," a drop in traffic looks like a failure. However, if the users who do click through are more qualified, the "failure" is actually a filter.

Expert Interpretation: The tradeoff here is between reach and intent. High volume search traffic often includes "window shoppers" who are just looking for a quick fact. When AI handles the facts, only the users with high intent, those who actually need a deep dive or a product, visit your site. The decision you need to inspect is whether you are reporting "vanity volume" to your leadership or actual commercial intent.

Evaluating the True Value of Lost Traffic

Not all traffic is created equal. Bodnar pointed out that a significant portion of HubSpot's lost 80% was essentially "junk" traffic. They had posts targeting high volume, low intent keywords, such as how to type a specific emoji or lists of famous quotes. These visitors were never likely to purchase enterprise software.

When AI Overviews began answering these trivial questions directly, that traffic disappeared, but the revenue didn't. In fact, revenue grew nearly 20% in 2025. This happened because HubSpot had spent the previous five years diversifying away from a total reliance on Google. They invested in a podcast network, a YouTube presence, a creator program, and the acquisition of The Hustle newsletter.

By the time the search cliff arrived, about 90% of their content leads were coming from these diversified channels, with the blog contributing only about 10%. They also took the aggressive step of pruning over 3,000 blog pages that weren't actually reaching their target customers.

Expert Interpretation: This highlights the danger of "content for the sake of content." Many companies create "top of funnel" content to inflate traffic numbers, but this creates a fragile ecosystem. The lesson is to prioritize audience ownership (newsletters, podcasts) over audience rental (search engines). If you don't own the relationship, you are at the mercy of an algorithm update.

Understanding the 'Loop' Marketing Framework

In response to these shifts, Bodnar and Kieran Flanagan introduced "Loop Marketing." The traditional marketing funnel is a straight line from awareness to conversion, but the modern buyer's journey is fragmented. Awareness now happens in Discord, Reddit, YouTube, and AI chat windows. Buyers often arrive at a website already informed by an AI engine.

The Loop consists of four interconnected stages:

Express: Defining the brand's unique point of view and "taste" before ever writing a prompt. Tailor: Using real customer data to personalize messages for specific buyers. Amplify: Distributing the brand across all places where humans and AI look for answers. Evolve: Feeding results back into the system to improve the next cycle.

The goal is to create a system where the team that learns and iterates the fastest wins, rather than the team that simply publishes the most pages.

Expert Interpretation: The "Taste Profile" mentioned in the Express stage is the most critical part of this framework. In an AI world, the "average" answer is a commodity. The only way to avoid being commoditized is to inject a strong, specific point of view that AI cannot synthesize from a general dataset. The decision here is to stop focusing on "best practices" and start focusing on "unique perspectives."

Testing the Claims: AI Personalization in Practice

It is easy to promote a new framework, but the real value is in the data. HubSpot tested the Loop system against their previous segment based email programs. The old way relied on personalization tokens and incremental tests on subject lines or button colors.

They replaced this with AI written emails that were deeply personalized to each recipient. These weren't simple prompts; they were pages long, incorporating the brand's Taste Profile, actual language used in sales conversations, tone guidelines, and a list of banned words. This rigorous approach led to an 82% improvement in conversions for those specific emails and a 50% increase across the broader email program.

Expert Interpretation: This proves that AI is a force multiplier for quality, not a replacement for it. The massive lift didn't come from the AI itself, but from the "pages long" prompts and the months of iteration. The tradeoff is time: you can have "fast and generic" AI content or "slow and high converting" AI content. Most companies choose the former and wonder why it doesn't work.

The New Metrics: Mentions and Citations

Since users are now interacting with AI before they reach a website, HubSpot has shifted how it tracks visibility. They distinguish between two types of AI presence:

1. Mentions: When an AI answer names HubSpot.
2. Citations: When an AI answer provides a direct link to HubSpot.

They track these across the most critical prompts for their business. Interestingly, they also monitor which third party sources AI engines lean on, such as G2, Reddit, or LinkedIn. This tells them exactly where they need to build a presence outside of their own domain to influence the AI's output.

The results are significant. Qualified leads from AI answers grew by 1,850% between Q1 2025 and Q1 2026. These leads convert at three times the rate of traditional search, and the deal cycles are roughly half as long because the buyer has already been "pre sold" by the AI conversation.

Expert Interpretation: This is a shift from SEO (Search Engine Optimization) to AEO (Answer Engine Optimization). The key takeaway is that your brand's reputation on third party sites (Reddit, G2, etc.) is now a direct driver of your AI visibility. You can no longer control the narrative solely through your own blog; you must influence the ecosystem that the AI uses as its source of truth.

The Risk of AI Sameness

A common critique of AI driven marketing is that it leads to a "sea of sameness." If everyone uses the same tools and the same prompts, every brand starts to sound identical. Bodnar agrees that this is the biggest risk of the AI era. Because AI is trained on the "average" of the internet, it naturally produces average output.

The solution is to do the hard work that software cannot do: defining your customer, your product story, and your convictions. Software can distribute a message, but it cannot create a unique identity. If a competitor could put their logo on your marketing and the message would still make sense, your marketing is adding no value.

Expert Interpretation: This is a reminder that strategy is a human function. The tool is the amplifier, but the "Taste Profile" is the signal. The decision you must make is whether to invest in "prompt engineering" or "strategy engineering." The latter is what prevents your brand from becoming a commodity.

The Human Cost of Reorganization

While the Loop framework suggests that AI expands the marketer's role, the reality of corporate restructuring is often more complex. In October, HubSpot announced layoffs affecting nearly 660 employees, or about 7% of its workforce.

While these cuts weren't explicitly attributed to AI efficiencies, the company's CEO, Yamini Rangan, noted a shift in focus. HubSpot is moving away from simply building features and "Hubs" toward delivering specific customer outcomes. This suggests a broader reorganization of how the company views value delivery, moving from a product centric model to an outcome centric one.

Expert Interpretation: This highlights the tension between theoretical "AI empowerment" and operational reality. Even in a company that is winning with AI, the shift in strategy often requires a shift in headcount and structure. It serves as a warning that adapting to AI isn't just about new tools, but about rethinking the entire organizational structure around how value is delivered.

Adapting Your SEO Strategy for the AI Era

You do not need a HubSpot sized budget to implement these changes. The goal is to stop defending traffic that doesn't monetize and start measuring actual AI visibility.

First, identify your "shrug emoji" pages, the high volume, low intent content that AI is likely to cannibalize. Sort your organic landing pages by whether they have ever produced a lead or a sale. If a page has massive traffic but zero commercial intent, it is a liability. You should decide whether to prune, consolidate, or rewrite these pages to target a specific buyer persona.

Second, start tracking mentions and citations. Identify the 25 to 50 most important prompts for your business and manually check them across ChatGPT, Claude, and Google. Note when you are mentioned and when you are linked. More importantly, note which third party sites the AI is citing. If the AI consistently points to Reddit or a specific review site, that is where your marketing effort should be directed.

Expert Interpretation: The hardest part of this transition is the psychological shift of seeing traffic drop. Most marketers are conditioned to panic when the line goes down. The decision you need to make is to change your reporting. Stop reporting "sessions" as a primary KPI and start reporting "revenue per visit" and "AI share of voice."

Changing the Scoreboard

The most important lesson from HubSpot's experience is not the loss of traffic, but the admission that they were watching the wrong scoreboard for 15 years. They only realized the metric was flawed when the scoreboard broke.

The old inbound playbook was about being found. The new playbook is about being chosen. Being found is a technical challenge that can be solved with keywords and backlinks. Being chosen is a strategic challenge that requires a unique point of view, a strong reputation in the ecosystem, and a deep understanding of the buyer's journey.

The second job is significantly harder, but as HubSpot's revenue growth suggests, it is far more profitable. The first step is simply to stop pretending that the old scoreboard still works.