Interpreting scores

Every ticket Vindex evaluates produces a story health score — a single number that tells you whether the story is ready for sprint planning, borderline, or one of the requirements that look ready but aren't — the kind that turn into rework and production defects.

Anatomy of a score

A story health score has three layers. Each layer answers a different question and is designed for a different reader.

Layer 1

Overall story health

A single whole percentage — 64%. The arithmetic mean of all six INVEST dimensions, stored on a 0–10 scale and always displayed as a percentage.

Layer 2

Plain-language finding

The problems Vindex found in plain English — “Acceptance criteria are missing or untestable.” No codes, no jargon.

Layer 3

INVEST dimension scores

The underlying six dimensions, each displayed as a percentage. For example T: 30% — testability is low.

Score ranges

The overall score and per-dimension scores display as whole percentages. The quality gate sits at 70%. Stories below it shouldn't enter a sprint without refinement.

Healthy

70 – 100%

Clears the quality gate. Ready for sprint planning with no blocking issues.

Needs review

50 – 69%

Borderline story health. Take it into refinement before committing to a sprint.

At risk

0 – 49%

Below the quality gate. Should not enter a sprint in this state.

Quality gate: a score threshold below which a story should not proceed to sprint. By default the gate is 70% for the overall score, but teams can tune it to match their own definition of ready.

Per-dimension fail threshold

Individual INVEST dimensions use a stricter threshold. A dimension score below 50%is counted as a “fail” — a clear signal the story doesn't satisfy that dimension well enough to move forward.

CodePlain-language labelWhat a low score means
I

Independence

Independent

The story depends on other work that is not yet done, so it cannot move into a sprint on its own.
N

Negotiability

Negotiable

Implementation is pre-decided in the ticket. There is no room for the team to choose the best approach.
V

Value clarity

Valuable

The user or business outcome is unclear. No one reading the ticket can tell who benefits and why.
E

Estimability

Estimable

There is not enough detail to size the work with any confidence.
S

Scope size

Small

The story is too large to complete inside a sprint as a working increment.
T

Testability

Testable

There are no acceptance criteria, or the criteria cannot be verified.

Where scores appear

Scores dashboard

The /portal/scores dashboard shows your company's all-time average, a time-series trend chart, and a paginated list of every scored ticket.

Score detail page

Click any scored ticket to see the full INVEST breakdown, recommendations, the context used during scoring, and the model that produced it.

Trend over time

Toggle individual INVEST dimensions on and off to see which areas are trending up and which are dragging down your overall story health.

Admin view

Vindex Admins see the same data globally at /admin/scores, with per-company averages surfaced directly on the companies list.

Acting on a score

A low score is a prompt to refine, not a judgement on the author. Open the story detail page and start with the top recommendations — they are ordered by the improvement that would raise the overall score the most.

Work down the INVEST breakdown to see which single dimension is pulling the score down. A 64% overall often hides one dimension at 30% while the rest sit at 70% or higher. Fixing that one dimension is usually enough to clear the quality gate.

Re-score after editing. In Jira the Issue Panel has a Re-score button; via the API, any update to the story body triggers a fresh evaluation automatically.

Continue reading