CodeveilStudio

Engineering digital products with design precision.

All work
Case studywebdesign

IQify Website

A browser-delivered cognitive assessment with a presentation layer engineered to earn trust

Visit live project
Presentation layer
80%
Install or signup
Zero
Scoring
Instant
Delivery
Static
IQify Website

Overview

IQify is a browser-delivered IQ assessment: a structured question sequence, deterministic scoring and an interpreted result, with no install and no account. The engineering weight sits deliberately in the presentation layer — roughly 80% of the codebase is markup and stylesheet — because in an assessment product the interface is the credibility.

The challenge

An online assessment is judged on how it looks before it is judged on how it works

Assessment products fail at the top of the funnel, not at the algorithm. A test that looks improvised gets abandoned around question three, and a result that arrives as a bare number gets dismissed regardless of how carefully it was calculated. The credibility problem is a design problem.

Two hard constraints shaped the build: it had to run anywhere with no install and no account, and it had to feel trustworthy enough that people finish it and believe the outcome.

  • Perceived credibility gates completion — abandonment happens in the first three questions
  • Scoring has to be deterministic and explainable, never a black box
  • A raw number is not a result; users need interpretation to act on it
  • Most sessions happen on a phone, so the mobile surface is the primary one

What we built

Presentation-heavy front end, deterministic scoring, static delivery

The codebase is around 80% HTML and CSS by volume — unusual for a web application and entirely intentional. Every unit of effort spent on typography, spacing, contrast and a calm single-question layout buys completion rate, which for an assessment is the only metric upstream of the score.

The test itself runs in a single container driven by state rather than page loads: a current index, a set of captured answers and a phase. Every screen is derived from that state, so there are no half-rendered transitions, no back-button traps and no route where a user can submit a partially answered test. Scoring compares captured answers against a fixed key and maps the total into interpreted bands.

The whole thing ships as static assets. There is no runtime to keep alive, no cold start, and no production failure mode beyond the CDN itself.

What we delivered

Single-container question flow

One container, driven by state, no page reloads. Progress is preserved through the sequence and the back button cannot desynchronise the test from what the user sees.

Deterministic scoring

Answers are compared against a fixed key with transparent totalling — no hidden weighting and no per-session variance. The same answers always produce the same score.

Interpreted results

The result page maps the score into readable bands with interpretation, rather than presenting a number and leaving the user to guess what it means.

Progress tracking and submission validation

Progression state is tracked through the sequence and incomplete submissions are blocked, so a partial run cannot produce a misleading result.

Responsive test surface

The same layout holds from phone to desktop, with question text and answer targets sized for touch first.

Static delivery

Served as static assets from GitHub Pages. Instant load, nothing to operate, and no server-side failure mode.

Engineering challenges

The problems that took real work to solve. Tap any one to read how.

Problem

The biggest risk in an assessment product is not a scoring bug — it is abandonment. Users leave a test that feels improvised long before they reach a result they could doubt.

Solution

The effort budget went into the presentation layer: type hierarchy, spacing, contrast and a single-question layout with nothing competing for attention. The 80/20 split between presentation and logic in the codebase is the direct measurement of that decision.

Results

80 / 20

Presentation to Logic

Deliberate weighting for a trust-critical interface

No signup

Zero-Friction Entry

Open the link and start the test

Banded

Result Interpretation

Score plus readable interpretation, not a bare number

Static

Delivery

Served from GitHub Pages with nothing to operate

Outcomes

  • A publicly accessible assessment that runs in any browser with no install, no account and no backend dependency.
  • A presentation layer that carries the product’s credibility, backed by deterministic and explainable scoring.
  • A result page that interprets rather than merely reports.
  • A defined hardening path: timed sections, randomised question order, submission locking, keyboard navigation and ARIA labelling for screen-reader support.

Related work

Want something like IQify Website?

Share your goals and we will come back with a clear plan and estimate.

Start a project