Why this exists

AI is changing how we build in OutSystems. OutSystems has announced it. Every partner is preparing for it. Every factory will soon be asked the same question: is AI making us faster? Better? Delivering more?

Today, no one can answer that honestly, because no one has a baseline.

Factories measure productivity the same way they always have: number of deploys, tickets closed, maybe velocity points. None of those tell you what is actually being delivered. And without that answer today, there will be no way to prove what changes tomorrow.

Impact Benchmark is that baseline. Install it, and the dashboard populates with up to a year of your factory's history from day one. The level of detail depends on what the platform still retains, but you see the shape immediately. From there, every new publish adds precision.

What it measures

Build. The real effort behind each publish, based on what was actually built: screens, server actions, client actions, aggregates, entities, integrations. Not saved files, not lines of code, not deploy counts. A label change scores differently from a new screen with logic and data access, and the scoring model reflects that.

Quality. Anti-patterns introduced versus anti-patterns resolved, for every publish. The component detects patterns such as N+1 queries, unbounded result sets, dead code, missing descriptions on public elements, and long undocumented flows. It also surfaces cross-module architecture findings: circular dependencies, monolithic concentrations, public elements nobody consumes.

Waste. Dead Application Objects accumulating in your factory. Screens, REST endpoints, entities, server actions, and other elements that nothing in the portfolio references. The stuff that quietly grows over the years and that no one has time to clean.

What you see

A single dashboard, embedded in the environment where you install it:

- Team Impact heatmap over 12 months, with three views

- Focused heatmaps for Build, Quality, and Waste

- Impact per Dev-Day, with percentile against peers

- Quality Net over the last 90 days

- Dead AO count and delta, with percentile against peers

- Monthly Impact trend, with a P25 to P75 peer band

- Quality Delta Flow, introduced vs. resolved, month by month

- Effort Distribution, creation vs. maintenance, by element type

- Where to Focus, top issues with occurrence counts

- Quality Patterns, with counts and peer percentiles

- Dead Code Inventory, by element type, with peer percentiles

How it works

Install the component on your environment. A background Timer reads your module history on the schedule you choose, and for each new publish it scores the delta across the three dimensions. Tuning is done through Site Properties in Service Center.

What you share by installing

Impact Benchmark is a peer benchmark. Most of what the dashboard shows, including percentile comparisons, pattern analysis, waste breakdown, and where to focus, is computed by a scoring service from the data your environment sends. This happens from the moment the component runs for the first time, without any additional setup or toggle. Installing means accepting it.

What leaves your environment: opaque identifiers and aggregated metrics. No source code, no business logic, no module, entity or action names. Human-readable names stay client-side and are resolved only in the dashboard you see.

Your environment is identified by a cryptographic hash of the activation code. The scoring service cannot reverse the hash back to your environment. Personal Environments are kept separate from Enterprise environments, so you are always compared against factories that look like yours.

If your organisation's policies do not allow this, do not install the component.

Who should install it

Delivery managers, team leads, architects, and CTOs running OutSystems factories who want visibility over what is actually being produced, how quality is moving, and where waste is accumulating.

Built by

Afonso Metello.