SEO has traditionally been built around a familiar set of measurements: rankings (positions), impressions, clicks, click-through rate, traffic, conversions and, increasingly, visibility across AI search.
But search data may tell us more than whether a website is being discovered.
It may also contain signals about how people are thinking while they search.
That is the hypothesis behind a research project exploring whether the language and structure of search behaviour can reveal cognitive load, choice overload and decision fatigue during complex consumer decision journeys.
For the SEO industry, this raises a much bigger question:
Are search queries simply indicators of demand, or are they also behavioural traces of the decision-making process itself?
The hypothesis: search behaviour leaves a cognitive trail
When someone searches for something relatively simple, the journey may be short and predictable.
A query leads to a result. The person evaluates the information, perhaps clicks a page, and moves on.
Complex decisions are different.
The search journey can become iterative:
broad question → more specific query → comparison → clarification → reformulation → alternative → another clarification
From an SEO perspective, this is normally interpreted as search intent changing, keyword refinement or a user moving further down the funnel.
The hypothesis is that the pattern itself may contain additional information.
As the search journey becomes more cognitively demanding, the language, specificity, semantic relationships and repetition within queries may change in systematic ways.
That means search data could potentially be analysed not only as a representation of what consumers want, but as observable behavioural evidence of how they are processing information.
Why this matters to SEO
This has important implications for how SEO professionals interpret search behaviour.
Most SEO reporting asks questions such as:
- Which queries generate impressions?
- Which queries generate clicks?
- Which pages rank?
- Where is CTR increasing or decreasing?
- Which topics are growing?
- Which keywords are commercially valuable?
Those remain important.
But they predominantly describe search visibility and performance.
They do not necessarily explain the cognitive journey happening behind the query.
Consider two clusters of search demand with similar impression volume.
The first cluster might show increasingly specific queries, strong semantic convergence and high engagement with search results.
The second might show high query diversity, repeated reformulations, semantic movement between related topics and relatively low CTR despite substantial impressions.
Traditional SEO reporting may treat both as demand.
The hypothesis suggests they may represent very different search experiences.
One could represent efficient information seeking.
The other could be evidence of increasing difficulty, uncertainty or choice complexity.
That distinction could eventually influence how SEOs approach information architecture, content design, internal linking, SERP strategy and AI-search optimisation.
Choice overload may be visible in the query space
One of the central hypotheses concerns choice-set breadth.
In consumer psychology, choice overload describes situations where the complexity of available alternatives can exceed a person’s ability to process them effectively.
Translated into SEO terms, the question becomes:
What happens to search behaviour when the information environment becomes too complex?
A search topic may contain dozens of subtopics, alternatives, conditions and competing interpretations.
Rather than moving progressively towards a more precise answer, users may begin to move laterally.
They may compare.
Then reconsider.
Then search for something adjacent.
Then return to a previous theme.
The proposed research therefore treats the breadth and diversity of query variants within a topic cluster as a possible observable proxy for choice-set complexity.
The hypothesis is:
Hypothesis 1: Topic clusters exhibiting greater choice-set breadth will show higher reformulation density and lower aggregate CTR, consistent with choice overload.
Importantly, this is a hypothesis rather than a claim that the relationship has already been proven.
That distinction matters.
The research is designed to test whether these patterns hold in naturally occurring search data.
Search reformulation could become an SEO behavioural metric
SEOs already understand query reformulation.
A user searches one thing, changes the wording, adds a modifier, removes a term or approaches the topic from another angle.
Usually, this is interpreted through search intent.
But reformulation could potentially become something more measurable.
Instead of simply counting keyword variants, we can examine how the meaning of the query changes.
For example:
- Lexical reformulation
The wording changes while the underlying intent remains relatively stable.
- Semantic reformulation
The vocabulary changes more substantially, but the query remains within the same conceptual space.
- Specialisation
The query becomes more precise or conditional.
- Re-abstraction
The search becomes broader again after previously becoming specific.
That last behaviour is especially interesting.
If a decision journey is progressing efficiently, query specificity might reasonably be expected to increase.
But if users encounter increasing complexity, the journey may not converge.
It may oscillate.
Hypothesis 2: The proposed hypothesis is that specificity will generally increase as a decision progresses, but high-complexity topics may instead demonstrate oscillation or re-abstraction.
For SEO, that creates a potentially useful distinction between normal refinement and cognitive friction.
AI search changes the problem
Generative search introduces another layer.
Traditional search often requires users to formulate a query, inspect a results page, choose a source and continue the process themselves.
AI-mediated search changes that interaction.
A user may receive a generated answer and then ask a follow-up question designed to challenge, clarify or extend the previous answer.
This creates a different form of query reformulation.
The query is no longer necessarily an isolated information request.
It may be part of a conversational sequence.
That leads to another hypothesis:
Hypothesis 3: AI-search follow-up queries will exhibit distinct linguistic patterns, including greater specificity and more conditional or clarifying language, than comparable traditional-search queries.
From an SEO perspective, this is significant.
The emergence of AI-mediated search means that optimisation may increasingly involve understanding not just:
What query retrieves my content?
but:
What question does the user ask next?
That shifts the unit of analysis from the keyword to the search journey.
Decision fatigue may have a temporal signature
The third major idea concerns time.
Decision fatigue proposes that sustained deliberate effort can progressively reduce the capacity for complex decision-making.
Search data gives SEO professionals something particularly interesting in this context:
time-series information.
If cognitive effort changes across a decision cycle, it may leave measurable traces.
The research therefore proposes testing whether periods of greater decision pressure are associated with:
- lower query complexity,
- reduced specificity,
- increased repetition,
- and potentially more simplified search behaviour.
The corresponding hypothesis is:
Hypothesis 4: Under increased decision pressure, aggregate query complexity will decline while repetition rises, consistent with a temporal signature of decision fatigue.
Again, the important point is not that a shorter query automatically means fatigue.
A short query could simply be efficient.
The research therefore proposes looking for convergence across multiple independent signals, rather than treating one metric as proof of a psychological state.
What this could mean for SEO measurement
This research points towards a different way of thinking about SEO analytics.
Imagine moving beyond a dashboard that simply reports:
Impressions → Clicks → CTR → Rankings → Conversions
towards one that also analyses:
Topic breadth → Query reformulation → Semantic movement → Specificity → Repetition → Temporal change
That would not replace conventional SEO metrics.
It would add another layer of interpretation.
Search data could potentially help identify where users encounter information friction before they convert.
That could have practical implications for:
- Content strategy
Where are users repeatedly reformulating because existing content does not adequately address the underlying information need?
- Information architecture
Are users moving between disconnected topic areas because the website structure does not reflect the way the decision is actually being made?
- Internal linking
Could stronger contextual connections reduce unnecessary search effort?
- SERP strategy
Are there topic clusters where high visibility is not translating into engagement because users are still uncertain about which information to trust?
- AI SEO
Are AI-mediated follow-up queries revealing new forms of search intent that keyword research does not capture?
The SEO industry may need to rethink the unit of analysis
For years, the dominant unit of SEO analysis has been the keyword.
Then came the topic.
Then search intent.
Increasingly, we are moving towards the entity, the journey, the conversation and the search ecosystem.
The hypothesis discussed here pushes that progression one step further.
What happens when the search journey itself becomes a behavioural dataset?
Anonymised aggregate Google Search Console data can already provide substantial behavioural dimensions, including queries, landing pages, geography, device, dates, impressions, clicks, CTR and position. The research framework uses approximately half a million rows of such data and deliberately works at an aggregate rather than individual-user level.
That limitation is important.
It means individual users cannot be tracked from one query to another.
Instead, the research reconstructs behavioural structure through query relationships, topic and intent clusters, and population-level temporal patterns.
In other words, the objective is not to claim:
“This individual user became cognitively fatigued.”
It is to ask whether aggregate search behaviour contains reproducible patterns consistent with cognitive load, choice overload or decision fatigue.
That is a much more defensible proposition.
From SEO reporting to behavioural intelligence
The broader idea is potentially bigger than SEO.
Search engines have always been one of the largest behavioural datasets available to marketers.
Yet much of the industry still treats search data primarily as a mechanism for measuring demand.
The alternative hypothesis is that search data can also help us understand decision processes.
That would reposition SEO closer to behavioural analytics.
Instead of only asking which content receives visibility, we could begin asking:
- How difficult was the information environment?
- Where did the search journey converge?
- Where did it fragment?
- Which topics generated repeated reformulation?
- Where did users become more specific?
- Where did they move backwards?
- Did AI-mediated search reduce the amount of cognitive work required, or create new forms of information complexity?
These are not yet established answers.
They are research questions.
And that is precisely what makes the hypothesis interesting for the SEO industry.
The future of search measurement may not be limited to understanding what people search for.
It may involve understanding what their search behaviour reveals about the decision journey behind the query.
That could move SEO measurement beyond visibility and towards something much more ambitious:
behavioural intelligence derived from search.
