On September 11, 2026, Jason Tabeling, a solutions lead at media agency Further, published a read of eighteen months of Google Ads account data on Search Engine Land: one- and two-word queries, which generated 62% of conversions in January 2025, generate only 52% in August 2026. Three- and four-word queries, meanwhile, went from 20% to 46% of conversions over the same period, more than double. For a B2B advertiser still bidding on short, generic head terms like “crm software” or “b2b marketing agency,” the signal is worth pausing on: this is a search behavior reshaping itself over time, beyond quarterly bidding noise.
This shift builds on something already tracked here: how a campaign converts depends increasingly on what happens once the click lands, a point covered in PPC Campaigns: ROI Happens After the Click. This time, it’s the shape of the query the customer types that’s changing, and it shows up directly in how impressions and conversions split by keyword length.
Eighteen months of movement in search term reports
Tabeling quantified the shift on both impressions and conversions. Between January 2025 and August 2026, the share of one- and two-word queries in impressions drops from 42% to 24%, eighteen points lost. Over the same window, three- and four-word queries, phrased as specific needs (“crm software for a 20-person company” rather than “crm software”), gain fifteen points, from 33% to 48%. The movement continues into longer phrasing: five- and six-word queries triple their share of conversions, from 3% to 9%, and queries of seven words or more quadruple theirs, from 1% to 4%, even though their absolute weight stays modest.
In his article on Search Engine Land, Tabeling notes he had already spotted the same pattern a year earlier, without giving further detail on that first read. His conclusion in September 2026: the shift hasn’t just continued, it has accelerated, driven by the rise of AI Mode and Gemini in the Google search experience.
The scale still needs the right caveat. This analysis comes from a practitioner, not from Google. Tabeling doesn’t publish the detail of his sample (accounts, industries, markets) or his measurement method. Nothing indicates a given B2B account follows this exact curve: the number works as a trend indicator worth checking on your own account, a market signal rather than an official statistic you can apply as is.
The usual shortcut, when cost per lead climbs on short keywords, is to look at the bid or the Quality Score. Tabeling’s numbers suggest a different read: these queries structurally carry less weight in conversions than eighteen months ago, independent of ad quality or bid amount. A rising CPL on short keywords can simply reflect this shift in weight, before the bid or the Quality Score even come into question.
Why the query changes nature with generative AI in search
The mechanism owes less to a bidding algorithm change than to an interface change. Since AI Mode and AI-generated summaries have taken up a growing share of Google results, typing a query increasingly looks like asking an assistant a question rather than stringing keywords together in a classic search engine. Someone who used to search “crm software” to compare options now more readily types a sentence that already describes their need, because the interface itself invites a full request.
The timeline helps gauge the scale of the shift. Google rolled out AI Mode broadly in the US on May 20, 2025, at its I/O conference, after an earlier open test phase in Labs; the tool then expanded to more than 180 countries in the following months. The window Tabeling observed, January 2025 to August 2026, brackets that rollout closely: short keywords lose ground at roughly the pace AI Mode spreads over the period.
On the organic side, the same shift is documented by AI Overviews’ arrival in France: the citations feeding generative summaries already favor long, contextualized phrasing. In Ads, the symptom differs (a split in impressions and conversions rather than a citation), but the cause traces back to the same transformed search interface.
What this changes for arbitrating a B2B Google Ads account
The shift counts double in B2B, where the buying cycle stretches over several months. A four-word query naming a specific need (team size, feature, budget) generally signals a more advanced stage of thinking than a generic one- or two-word query, which catches a real need and simple curiosity alike.
A campaign anchored on generic head terms is therefore capturing a share of traffic in structural decline, a shift a simple bid adjustment won’t fix. Match type is where the arbitration now plays out: broad or modified broad match on short keywords catches less and less of the intent that converts, while exact and phrase match on longer phrasing capture a growing share of conversions.
Measuring this shift properly requires clean conversion data, a condition already set out when Data Analysis: Start From the Decision, Not the Data made the case for B2B lead gen: without reliable tracking, there’s no way to tell a genuine drop in conversions on short keywords from a plain measurement artifact. The other side of the coin is negatives: a short head term that keeps pulling volume without converting deserves to be neutralized before it absorbs the budget meant for longer, more qualified phrasing.
Here are three concrete checks worth running on your own account before considering a broader strategy shift.
- Cross-reference, in the search terms report, the split of impressions and conversions by query length over the last eighteen months.
- Review the match types applied to high-volume short keywords.
- Revisit the negative keyword list on generic head terms that pull volume without converting.
Nothing says your account follows Further’s exact curve. But the question deserves to be asked with the real numbers in front of you, not with the assumption that short keywords have always performed.
And if this read makes you revisit an assumption on your own account, write to us. We’re curious.
Photo credit: Pixabay.
