The Data

Page / 02 · Dataset / 001

IntelliLight sits at the intersection of neuroscience, physical environments, behavior and technology.

That creates difficult questions.

  1. 01Who is the product really for?
  2. 02Which features matter enough to build?
  3. 03How should people interact with it?
  4. 04Which environments deserve adaptation?
  5. 05What should the product cost?
  6. And eventually: what actually changes
    when the environment changes?

  1. 01Discovery
  2. 02Measurement
  3. 03Analysis
  4. 04Decision
  5. 05Validation
  6. 06Iteration

Our data function exists to turn those questions into measurable product decisions.

Data is not a reporting layer at IntelliLight. It is part of the product-development system.

The data function

Led by

Shayan Bhatti

Founding Technical Engineer

Portrait of Shayan BhattiFig. 02
SHAYAN BHATTI / Engineering / Data / Validation

Shayan leads much of the technical and analytical infrastructure behind IntelliLight’s evidence system, helping transform research, behavioral validation and product testing into structured information that can influence what gets built.

View Shayan’s LinkedIn

Why data

A neuroscience product
cannot be built
on intuition alone.

Human behavior is noisy. Preferences differ. Context changes. What people say they want can differ from what they use, and what looks compelling conceptually may not matter in a real environment.

LightSoundTimingTaskContextIndividual preference

Every additional variable introduces another hypothesis. Our response is to measure, narrowing the gap between what we think people need and what the evidence actually shows us.

The goal is to gather as much meaningful evidence as we can.
And to use every bit of it to make deeply informed decisions.

The IntelliLight evidence system

Six connected modules.
One operating system.

  1. 01

    Discover

    Start with the human problem.

    • Who experiences the problem
    • How often it occurs
    • How people respond today
    • Who uses the product
    • Who purchases the product
  2. 02

    Structure

    Turn interviews into variables.

    • Participant-level records
    • Standardized categories
    • Cleaning and segmentation
    • Comparable session signals
  3. 03

    Measure

    Observe behavior, not only opinion.

    • Usability observation
    • Mode comprehension
    • Feature-level response
    • Repeat intent
  4. 04

    Analyze

    Look for patterns that survive scrutiny.

    • Preference distributions
    • Audience differences
    • Purchase signals
    • Commercial modeling
  5. 05

    Decide

    Data becomes valuable when something changes because of it.

    • What should we build first?
    • Which features earn complexity?
    • What price can the economics support?
    • Which assumptions remain weak?
  6. 06

    Iterate

    Every answer creates the next experiment.

    • Discovery informs prototypes
    • Prototypes generate behavior
    • Behavior informs priorities
    • New questions return to discovery

The tools are not the point. The output is a clearer decision, that is the difference between analytics and decoration.

Case study 001 · Who is the customer?

We started
by thinking about students.

The data
changed the question.

Students represented an intuitive early audience. They study. They work late. They move between focus, learning, creativity and rest. But customer discovery revealed an important distinction.

User

Student

Buyer

Parent

The person using IntelliLight may not be the person purchasing IntelliLight.

42

Parent interviews

92%

Reported problem fit

80%

Stated purchase intent

74%

Gift appeal

These are early customer-discovery findings, not population-level market estimates.

Their importance is not that they prove the market. Their importance is that they changed how we think about it.

USER ≠ BUYER

  • Positioning

    Speak to the student experience while communicating value to the person making the purchase.

  • Distribution

    The acquisition channel may differ from the environment where the product is ultimately used.

  • Product design

    The experience needs to make immediate sense to the student while remaining legible in value to the buyer.

  • Pricing

    Willingness to pay should be examined at the purchasing level rather than inferred only from the end user.

  • Messaging

    “Study better.” is different from “Create a better environment for someone you care about.”

We did not change the data to fit the strategy.
We changed the strategy to fit the data.

Case study 002 · From interest to interaction

What people say matters.
What they do matters more.

Customer discovery helps us understand whether an idea resonates. Prototype testing begins to show what deserves to exist.

60+

Student opt-ins

11+

Completed validation sessions

Our internal validation framework standardizes signals across testing rather than treating each participant session as an isolated anecdote.

What we are measuring

  • 01

    Usability

    Can a participant understand the system without excessive instruction?

  • 02

    Mode comprehension

    Do Focus, Learn, Create and Reset make intuitive sense?

  • 03

    Feature value

    Which environmental controls appear meaningful enough to prioritize?

  • 04

    Adaptive lighting

    How do participants respond to changes in lighting conditions and configuration?

  • 05

    App integration

    Where does an interface improve the experience, and where does it add management?

  • 06

    Personalization

    Which preferences appear stable, and which change with task, timing or context?

  • 07

    Repeat intent

    Would participants realistically incorporate the product into an existing routine?

Every additional feature
has to earn its place.

Internal infrastructure

From raw behavior
to product decision.

  1. 01

    Raw input

    • Interviews
    • Surveys
    • Opt-ins
    • Testing sessions
    • Usability observations
    • Feature feedback
    • Pricing responses
  2. 02

    Data structure

    • Python
    • SQL
    • Standardized variables
    • Participant-level records
    • Categorization
    • Cleaning
    • Segmentation
  3. 03

    Analysis

    • Behavioral patterns
    • Feature relationships
    • Audience differences
    • Preference distributions
    • Purchase signals
    • Usage signals
  4. 04

    Visualization

    • Tableau validation dashboards
    • Cohort views
    • KPI monitoring
    • Feature comparison
    • Participant funnels
  5. 05

    Decision

    • Build
    • Test
    • Change
    • Remove
    • Prioritize
    • Reposition

The purpose of the stack is not technical sophistication for its own sake. It is reproducibility.

When a hypothesis changes, analysis can be revisited. As a dataset grows, the same framework can be rerun. New participants can eventually be interpreted against a growing body of evidence.

From signal to roadmap

Not every requested feature
should become a feature.

Early validation has already helped IntelliLight examine priorities including adaptive lighting and app integration before moving further toward manufacturing.

Hardware decisions are expensive. Software can often be revised rapidly after launch. Physical products create constraints much earlier.

The more we can learn before manufacturing, the more intentionally we can build.

High interest · High value

Build + validate

High interest · Low evidence

Investigate

Low interest · High complexity

Question

Low interest · Low value

Do not build

Measure
before
metal.

With hardware,
analytics moves upstream.

Many digital products can launch, observe behavior and iterate rapidly. Physical technology introduces a different cost structure.

MaterialsComponentsToolingAssemblyInventoryShippingReturnsManufacturing

IntelliLight’s data work therefore begins before mass production, not after it, reducing uncertainty around what deserves to become physical.

Case study 003 · Commercial modeling

A product can solve a real problem
and still fail the math.

Estimated unit cost

$35

under current assumptions

Willingness to pay

Derived from customer-discovery research

Competitive context

Benchmarked across 5 relevant products / alternatives

Channel

DTC versus B2B economics

Early model / DTC

$99.99

Modeled price

~65%

Modeled gross margin

Early model / B2B

$85

Modeled price

~59%

Modeled gross margin

These figures represent internal early-stage modeling assumptions and are not finalized IntelliLight pricing or financial guidance. Costs, pricing and margin expectations may change substantially as the product moves toward manufacturing.

Commercial modeling led by Aizaz Faisal ↗, Chief Financial Officer

  • Does the problem exist?
  • Does the product make sense?
  • Does someone want it?
  • Can we build it?
  • Can the economics support it?

The data function

An in-house system
for turning uncertainty
into decisions.

IntelliLight’s data function operates across customer discovery, product validation and commercial strategy. Its role is not confined to dashboards generated after decisions are made. It participates upstream.

PythonSQLTableauFinancial modeling
  • Discovery analytics

    Understanding the problem and market.

  • Behavioral analytics

    Understanding participant response.

  • Product analytics

    Understanding which features deserve investment.

  • Experimental design

    Developing repeatable validation frameworks.

  • Commercial analytics

    Understanding pricing, economics and market structure.

  • Decision intelligence

    Connecting evidence back to strategy.

Current workstreams

  • 01Customer discoveryConverting qualitative interviews into structured behavioral and commercial signals.
  • 02SegmentationSeparating user, buyer and influencer behavior instead of treating “the customer” as one category.
  • 03Validation infrastructureStandardizing usability and feature KPIs across participant sessions.
  • 04Product prioritizationUsing validation findings to examine adaptive lighting, interface and personalization priorities.
  • 05Commercial modelingConnecting willingness-to-pay research with projected costs and channel economics.
  • 06Experiment designDeveloping measurement architecture for larger environmental testing.
  • 07Environmental response dataBuilding toward a system capable of studying how preferences respond to different configurations over time.

Data philosophy

What we believe
about data.

  • A large dataset cannot rescue a bad question.
  • A dashboard is not evidence.
  • A correlation is not automatically a product feature.
  • What people say and what people do are different signals.
  • Sample size should be visible, not hidden.
  • Uncertainty is information.
  • A result that changes the roadmap can be more valuable than one that confirms it.

The purpose of analytics is not to prove IntelliLight right.

It is to make
IntelliLight
less wrong.

Science creates the hypothesis.
Data tests the assumption.
The product is what survives both.

IntelliLight is still learning. Follow the research as we build.

Early access

Follow IntelliLight
as we build.

Join the early-access community for prototype updates, research developments, testing opportunities and future product announcements.

IntelliLight is an emerging technology platform under development. Product concepts, specifications and functionality may evolve as research, validation and product development continue. IntelliLight is not intended to diagnose, treat, cure or prevent any medical condition.

We use your details only to share IntelliLight research and product updates. You can unsubscribe at any time.