SEO & Search

Forrester: B2B Marketers Need to Rethink Their Approach to Search

Forrester forecasts a 50–75% drop in B2B traffic within 12–18 months. LLM traffic converts 3–4 times better, but, according to analysts, brand visibility has already been cut in half.

3 min read

‘Quantity down. Quality up’: B2B marketers urged to rethink search
‘Quantity down. Quality up’: B2B marketers urged to rethink searchhomethods / Openverse

B2B website traffic could drop by 50–75% in the next 12–18 months — that is the forecast from Forrester analysts. The reason: buyers are shifting from classic search engines to "answer engines" built on large language models. The topic came up at the Forrester B2B Summit.

In 2024, users performed two billion B2B search queries on a weekday, and 40% of them ended with a click-through, according to Forrester data. Today, search volume is growing, but engagement is falling: buyers get answers directly from AI and never visit websites.

At the same time, traffic driven by LLMs converts three to four times better. "Volume down, quality up. Maybe it balances out; maybe we muddle through," said Nicky Briggs, vice president and principal analyst at Forrester. "But we're not in the business of muddling through. The fact remains: visibility has been cut in half."

Briggs described the situation as a "visibility vacuum" caused by a paradigm shift: instead of search engines, buyers are turning to AI answer generators.

Search Is No Longer a Separate Discipline

Gerald Murphy, a specialist in search, AI, and consumer behavior who holds the title of "evangelist" at Semrush, cited his employer's data: 90% of people now use LLMs alongside traditional search engines like Google, and one in three ChatGPT queries relies on SEO data. From this he concluded that AI search cannot be treated as a discipline separate from SEO, content, paid search, and PR — marketers need a unified strategy.

According to Murphy, it is time for SEO teams to stop focusing on individual keywords. "Your SEO team should no longer be targeting specific keywords. You need to start thinking about semantics," he noted.

Here is how it works: when a buyer enters a prompt, AI breaks it down into many related queries — so-called "query fanouts." Brands and competitors that were not in the original question can end up in the answers. Murphy urged analyzing these fanouts and passing the data to analytics, content, and paid search teams.

"A single AI query genuinely touches multiple teams, and we simply have to ensure regular information sharing between them," he said.

Old Metrics No Longer Work

Traditional metrics — rankings, clicks, referral traffic — lose their meaning when AI answers the buyer's question without a website visit. "The other elephant in the room is that we're applying outdated metrics to a completely new system," Murphy emphasized. "Ranking number one is no longer the goal, especially when it comes to AI."

Instead, marketers need to track: whether the brand appears in the AI answer, where exactly, which competitors are mentioned, and which sources are cited.

Authority Beyond Your Own Website

AI systems assemble answers from many sources, so brands need to understand which third-party platforms are being cited and how the brand is represented across the internet as a whole. "Don't chase yesterday's citations. You need to optimize for relevant articles," Murphy advised.

He cited Reddit as an example of a platform B2B marketers should pay attention to. It increasingly appears in searches for tool selection, strategy research, and choosing agencies and vendors. "If you look at Google trends in the US for Reddit in a B2B context — the growth is huge. It already ranks number one for thousands of keywords," Murphy explained.

Briggs added that sources of authority should also be sought within the company. "When we can identify the topics people talk about with genuine excitement that align with what we should be famous for — that's when authenticity comes through, and the content practically writes itself," she said.

Conclusion: A Unified Marketing Operating System

In Murphy's view, these changes require bringing all marketing functions together. "Content, PR, communities, and AI optimization — everything plays a role," he believes. AI may finally push businesses to implement the "holistic search strategies" marketers have been talking about for years.

"The machine doesn't respect the org chart. The north star is building an operating system," Murphy summed up. According to him, marketers need to account for both paths — how the machine sees the buyer's journey, and how a human does.

Source: Marketing Week

Share this article:

More from Сергей Мельник

Related articles

« Previous articleNext article »