How Spar works

Each store in Spar goes through a pipeline. You don't need to understand every step to use the product, but it helps when you're explaining results to a stakeholder.

1. Store profile

Spar starts by reading the store as a visitor would: home page, key product pages, cart, and checkout. From that it infers:

• Brand tone and category.
• Likely customer segments and pricing tier.
• Page types present on the site (PDP, PLP, blog, etc.).
• Visual design language and primary CTAs.

The profile is stored on the store and feeds into every downstream step. You can review and edit it under Settings → Site profile.

2. Crawl

Spar crawls a representative slice of the site — typically the home page and the primary pages for each major page type it detects (PLP, PDP, cart, checkout, etc.). You can pin additional URLs under Settings → Crawl to make sure they're always included.

3. Rules engine

Each crawled page is evaluated against a large library of UX heuristics. Rules cover layout, copy, friction, trust, accessibility, mobile responsiveness, performance, and more. A gap is created when a page fails a rule.

4. Analytics analysis

If you've connected GA4 or Microsoft Clarity, Spar pulls your funnel and behavioral data and breaks it down by device, traffic channel, geography, funnel flow, page speed, and error rate. Gaps get cross-referenced against that data — a friction point that also shows a 30% drop-off in your real funnel is going to outrank one that doesn't.

5. Scoring

Every gap gets a revenue impact score based on the size of the affected segment, the funnel stage, and the typical lift seen for fixes of that class.

Every hypothesis gets three scores:

• ICE — impact × confidence × ease.
• PIE — potential × importance × ease.
• PXL — a binary checklist of high-signal factors (above the fold, noticeable in 5 seconds, analytics-backed, etc.).

The default ranking in Triage and Opportunities is by revenue impact.

6. Hypotheses and ideas

Each high-impact gap becomes one or more hypotheses — "if we change X, then Y will improve, because Z". Each hypothesis comes with:

• A wireframe preview of the proposed change.
• A scoped success metric.
• A recommended treatment (test it, or just ship a fix).

7. Tests and fixes

A hypothesis you choose to act on becomes either:

• A Test — Spar generates a variant, you deploy it, and Spar watches for significance.
• A Fix — a direct change you ship without an A/B test (often used for obvious bug-class issues).

8. Reports and learnings

When a test ends or an audit refreshes, Spar produces a Report snapshot you can share. Completed tests roll into Learnings so the workspace remembers what's already been tried.