Example Study

Decision-making behavior in a product research task

15 participants chose between three laptops while answering survey questions in LaBrowser's experiment pane. The browsing data reveals something surveys alone never could: what people say matters to them and what they actually research are often very different.

Study Overview

15

Participants

2,610

Total events

52

Search queries

11.4 min

Avg session

60

Survey responses

3

Laptops compared

Participants browsed freely while an experiment pane on the left showed survey questions. The three laptops: MacBook Pro M3 ($1,599), ThinkPad X1 Carbon ($1,299), and Dell XPS 15 ($1,199).

What Participants Said

Survey responses came through the experiment pane as INPUT_SUBMIT events — captured alongside browsing events in the same timeline.

Which laptop would you buy?

MacBook Pro was the most popular choice.

MacBook Pro M36ThinkPad X1 Ca...5Dell XPS 154

Most important factor?

Price was the most commonly cited reason.

Build quality6Price5Brand reputation3Performance1

Confidence in choice (1-10)

Most participants were fairly confident (median: 8).

02356678910

Did you consider all three equally?

Most said yes. The browsing data tells a different story.

Yes8Mostly4No3

What Participants Actually Did

While participants answered surveys, LaBrowser captured their full browsing behavior — every search query, every page visited, every second spent reading reviews.

Minutes spent researching each laptop

Participants who chose a laptop spent 2-3x more time researching it.

Chose this laptop
Did not choose
MacBook Pro M34.41.1ThinkPad X1 Ca..20.2Dell XPS 152.40.4

Most visited review sources

Notebookcheck and Rtings were the go-to technical review sites.

amazon.com29tomsguide.com22theverge.com19pcmag.com18notebookcheck....17rtings.com13reddit.com13bestbuy.com4

The Reveal: Stated vs. Actual

This is where it gets interesting. When we compare what participants said was important to them with where they actually spent their browsing time, clear mismatches emerge.

Participants who said "price was most important" spent 51% of their browsing time reading performance reviews — and only 15% comparing prices.

"Price is most important" group (6 participants)

What they said vs. where they actually spent their time (%).

Price
Performance
Build quality
General
Stated: "Price..100%Actual browsin..50%21%

"Performance is most important" group (4 participants)

This group's browsing better matched their stated preference, but price still consumed 30% of their research time.

Price
Performance
Build quality
General
Stated: "Perfo..100%Actual browsin..38%48%

Per-participant browsing breakdown

Five participants who chose different laptops with different stated reasons. Note how P01, P08, and P12 all said "Price" but spent most time on performance reviews.

Performance reviews
Price comparison
Build quality
General browsing
P01 — chose Ma..70%17%P02 — chose Ma..75%25%P03 — chose Th..27%39%34%P04 — chose De..100%P05 — chose Th..100%P06 — chose Ma..38%48%P07 — chose De..51%49%P08 — chose Th..100%P09 — chose Ma..100%P10 — chose Ma..81%19%P11 — chose De..100%P12 — chose De..81%19%P13 — chose Th..18%68%P14 — chose Th..71%29%P15 — chose Ma..100%

"Considered all options equally" — did they?

Participants who said they considered all three laptops equally still spent 58% of their time on the one they ultimately chose.

Time on chosen laptop
Time on other laptops
Said "Yes"8119Said "Mostly"7921Said "No"3565

Scroll Depth & Engagement

How deeply participants read correlates with their decision confidence. Those who scrolled further through reviews reported higher confidence in their final choice.

Avg scroll depth by page type (%)

Reddit threads were scrolled deepest; video review pages least.

Reddit threads62Full reviews57Spec comparisons57Price pages57

Confidence vs. avg scroll depth

More reading correlates with higher confidence.

MacBook Pro M3
ThinkPad X1 Carbon
Dell XPS 15
4953576165678910Stated confidence (1-10)Avg scroll depth (%)

What Only LaBrowser Shows

This study required two data streams synchronized in real time: explicit responses from the survey and implicit behavior from the browser. No other tool gives you both.

Survey tools

Capture what participants say, but not what they do. You'd see "price matters most" but never know they spent an hour reading performance benchmarks.

Browser extensions

Capture browsing history but can't embed a survey alongside it. Participants answer questions in a separate tab — and the context switch changes their behavior.

LaBrowser

The experiment pane shows the survey while participants browse. Both streams arrive as timestamped events in the same timeline. The mismatch analysis above is a direct result.

Download & Reproduce

The sample dataset and full analysis code are on GitHub. The data files match the exact format you'd export from the LaBrowser Study Console.

technologylab-ai/labrowser-example-product-decision

Sample data files (events.json, google_search_v1.json, study_config.json) plus a Jupyter notebook and Python script reproducing every chart on this page.

View on GitHub

Or clone locally

git clone https://github.com/technologylab-ai/labrowser-example-product-decision
cd labrowser-example-product-decision
pip install -r requirements.txt
jupyter lab analysis.ipynb
Want the product and analysis context behind this task? Read the Getting Started guide or compare this task with the health search example .