For many businesses, their website serves as a crucial storefront, making a loss of 140 million visits in a single year a significant concern.
This is precisely what happened to HubSpot, and the culprit was artificial intelligence.
HubSpot provides a suite of sales, marketing, and customer service tools tailored for business-to-business operations.
Like numerous other firms, HubSpot has been impacted by a fundamental shift in how individuals conduct internet searches.
“I recall a time when searching [the web] yielded little to no valuable information,” states Kipp Bodnar, HubSpot’s chief marketing officer.
“Occasionally, something would surface, but it required sifting through 10, 20, or even 30 links.”
“Today, we have instantaneous access to the world’s collective intelligence. Consequently, the way people discover information and subsequently take action has undergone a profound transformation.”
For companies such as HubSpot, the drop in traffic stems from multiple factors. Search engines have refined their algorithms to combat AI-generated content, underscoring the importance of a website establishing credibility within its core subject matter.
Users are increasingly turning to AI tools rather than traditional search engines. Simultaneously, search engines are integrating AI-driven overviews at the top of their results pages, often providing users with answers directly, eliminating the need to navigate to external websites.
“The click-through rate for searches featuring AI overviews is approximately 60% to 70% lower,” Bodnar notes.
Consequently, companies are exploring strategies to gain prominence in the responses generated by AI.
Answer engine optimisation (AEO), sometimes referred to as generative engine optimisation (GEO), focuses on improving website rankings within AI tools, including AI overviews and platforms like ChatGPT.
These tools are powered by large language models (LLMs), a form of AI technology.
Many businesses are implementing AEO in conjunction with search engine optimisation (SEO), which aims to improve website rankings in search engine results.
“We’ve leveraged answer engine optimisation to enhance the conversion rate and the caliber of individuals engaging with our platform,” Bodnar explains.
“I struggle to envision a future where a business can remain competitive without demonstrating proficiency in this area.”
AEO necessitates an understanding of evolving search behaviours.
“Traditional Google searches might involve entering four to six words,” Bodnar states. “Conversely, AI search engine queries average between 40 and 60 words, reflecting a significant shift in specificity.”
He illustrates this with the example of a motorhome rental company in New Zealand. A user might task AI with formulating a comprehensive holiday itinerary for a family of five, incorporating opportunities to observe a specific animal.
To be referenced in the response, the motorhome company might need to publish an article detailing the most popular animals in New Zealand for children, written in natural language that aligns with potential user queries.
HubSpot has been actively reorganising its content.
The company previously published lengthy articles detailing its products and their interconnected features. However, Bodnar explains that this level of detail is now less critical, as AI can furnish such explanations.
The revised structure utilizes concise content modules that AI can readily extract. For instance, if a user inquires about a contact management feature, AI tools can quickly locate the relevant information.
Currently, AI contributes between 7% and 12% of HubSpot’s monthly website traffic, but Bodnar anticipates that it will become an even more critical channel for customer discovery.
“We anticipate increased direct traffic and traffic from other sources due to the influence of LLM responses,” he says.
“Adaptation is essential for survival,” asserts Ann Lowe, head of PR and communications at Spice Kitchen, a company specializing in spice gift sets.
To promote its latest product, Spice Kitchen is developing a content cluster focused on the history of the spice trade. This dedicated section of the company’s website aims to establish authority on the subject.
“We’re aiming to capture the attention of AI search bots with this content,” she states. “It won’t be a sales-oriented platform, but rather resemble a training course designed for research purposes, allowing users to discover us organically along the way.”
She has been collaborating closely with the agency Lumos Digital.
“Traditionally, optimisation efforts have centered on product pages, targeting individuals poised to make a purchase,” explains Nathan Pearson, co-founder of Lumos Digital.
“However, the focus now appears to be shifting towards the research and decision-making stages, with the goal of engaging potential customers earlier in the process.”
He recommends that companies publish comprehensive buying guides. “If you’re offering a guide to the best trainers for long-distance running, ensure that all products are listed with a clear winner identified. AI algorithms favor this approach.”
Research and media organisations seeking to rank prominently in AI results can glean insights from Spice Kitchen’s strategies.
Andy Lochtie, co-founder of Lumos Digital, emphasises the significance of expertise, authority, and trust indicators.
These indicators include securing numerous inbound links from other reputable websites, providing outbound links to high-quality sources, and implementing content policies and author biographies to enhance credibility.
Andy Pickup serves as digital director at MKM Building Supplies, an independent builders’ merchants that also caters directly to the public.
“We’ve observed a decline in website traffic as users obtain answers from AI models,” he states. “They no longer need to visit our website to consult a blog post on installing artificial grass, for example.”
“If this trend persists, website traffic could potentially dwindle to near zero.”
Pickup recognised the importance of being cited in AI results. “We must ensure that AI models reference us, rather than our competitors, when users seek answers related to building projects.”
He hopes this will drive foot traffic to physical stores, where customers can receive personalised assistance from staff.
While Google remains the dominant search engine, ChatGPT generates more website visitors than Google’s integrated AI.
“This represents a seismic shift in user preferences regarding app usage,” he observes. “Customers are consciously choosing to bypass Google, despite its built-in AI, and are instead opting for ChatGPT.”
He embarked on what he terms a “defensive strategy,” creating blog posts about best-selling products for AI tools to reference.
“Similar to SEO, the goal was to establish ourselves as experts in these areas and provide LLMs with the information they need to furnish comprehensive and conclusive answers,” he explains.
“The content has evolved beyond simply describing a product; it now focuses on how the product can help users solve a problem.”
While search engines once prioritised keywords, AI engines require content that facilitates easy processing of the page’s meaning. As a result, MKM’s new pages feature summaries, bulleted lists to segment information, and frequently asked questions (FAQ) sections.
“The key is to ensure that your content is clear, concise, and easily understood,” he emphasises.
Behind the scenes, a site map helps AI bots navigate the website.
While many users will simply read the AI-generated answer, some will click through to the source. Over the past year, MKM’s traffic from AI has risen from negligible levels to a “low double-digit percentage” and continues to increase.
Pickup notes that AI visitors are significantly more likely to make a purchase than visitors from search engines. “My hypothesis is that customers gain the information they need from the LLM response, instilling confidence in their purchasing decision.”