Roll-up Scorecard

What the Roll-up Scorecard is

The Roll-up Scorecard is a single, at-a-glance view that brings together the engineering metrics your organization cares about most and presents them side by side across your teams and workspaces. Each row represents a team (or workspace), and each column represents a metric — for example PR cycle time, velocity, coding days, or PRs merged. Every cell shows the team's value for that metric over a chosen time window, color-coded so that strong, acceptable, and at-risk results are immediately visible.

Rather than asking leaders to open many dashboards and mentally stitch results together, the scorecard consolidates them into one structured table that can be reviewed in a regular cadence such as a monthly technology review.

Why it matters

Engineering leaders are usually accountable for outcomes across many teams at once, and the underlying metrics live in different places. The Roll-up Scorecard exists to close that gap by giving a consistent, comparable summary that supports a few common needs:

• A leadership-ready summary. A concise view that a director or executive can review and bring into reviews with their own leadership, without rebuilding it each time.

• Consistent comparison across teams. The same metrics, measured the same way over the same period, so teams can be looked at on a level footing.

• Fast identification of attention areas. Color-coded scoring surfaces which teams and which metrics are trending well versus which need a closer look.

• A repeatable cadence. Because the view is configured once and refreshes on its own, it supports recurring reviews instead of one-off reporting exercises.

How it is structured

The scorecard is organized as a grid of teams and metrics:

• Rows correspond to teams or workspaces. A row can represent a single team, or it can be a roll-up row that summarizes a group of teams (see How roll-up behaves below).

• Columns correspond to metrics. Each column is backed by a metric configuration that defines exactly what is being measured and how it is filtered.

• Cells show each team's value for a metric over the selected time window, with color-coding applied based on the scoring rules for that column.

Because each column is independently configured, different teams can apply different filters for the same metric where their workflows differ. This flexibility is powerful, but it also means that a metric is only as comparable across teams as its configurations allow.

Metrics and time windows

Columns are added from a set of metric templates that cover common engineering measures. Frequently used examples include contributors, velocity, normalized velocity, PRs created, PRs merged, PR cycle time, committed work, and completed work. Templates are intentionally generic; the specific scoring and any team-specific filtering are applied when the column is set up for your organization.

Each metric is evaluated over a defined time window — for instance a trailing period such as the last 30 days, or a comparison of the most recent full month against the current period. The window is part of the configuration so that every team in a column is measured over the same span.

How scoring works

Each metric column carries scoring rules that turn a raw value into a color and, where configured, into points:

• Thresholds and color. A metric's value is compared against thresholds to produce a red / yellow / green indicator, so a reviewer can read status without interpreting raw numbers.

• Points and weighting. Metrics can be assigned point values so that more important metrics contribute more to a team's overall score. The team's score reflects the points earned across its configured metrics relative to the points available.

• Unconfigured metrics. If a metric has not been configured for a team, the scorecard indicates that it is not configured rather than showing a blank or implying a failing result.

How roll-up behaves

A roll-up row is meant to summarize a group of teams at a parent or organization level. It is important to understand how this summary is produced:

There are a few practical reasons a parent row may not equal the arithmetic sum of its children:

• Shared people and tools. When contributors or tools are shared across teams, they cannot be cleanly divided among child rows, so a parent total will not always reconcile to the children added together.

• Per-team metric definitions. Teams sometimes define the same metric differently (for example, applying different state or component filters to velocity). When child definitions differ, a single parent configuration cannot simply combine them.

• Roster alignment. The set of people and tools on a parent view needs to be maintained so that it genuinely reflects its children; if they drift apart, the roll-up will reflect that drift.

For these reasons, a common and reliable approach is to define a roll-up row as an aggregate (such as an average) across the child team views rather than expecting an exact sum. The right approach depends on the metric and how your teams are organized, and is worth confirming with your Allstacks contact when configuring roll-up rows.

Permissions and visibility

Visibility of the scorecard can be aligned to workspace permissions, so that a given team or business unit sees the portions of the scorecard relevant to them rather than the entire organization.

Data freshness

Scorecard values reflect the data processed by your Allstacks pipeline, and results are cached so the view loads quickly. After new data is ingested, the scorecard updates on its normal refresh cycle, so a very recent change in a source system may take a short time to appear.

Getting help

The Roll-up Scorecard is configured collaboratively during onboarding and refined over time. To add or change metrics, adjust scoring thresholds, or set up roll-up rows for your team structure, reach out to your Allstacks Customer Success contact.

Draft for internal review. Sections marked "Confirm before publishing" contain product specifics that should be verified before this page is published to customers.