Mango Signal App Overview

Mango Signal

1. What it is

Mango Signal is a workforce-engagement intelligence layer. It reads signals from apps your organization already uses — Surveys, Recognitions, Performance, News Feed, and Shift Feedback — blends them into a single engagement health score, clusters the recurring themes behind that number, and flags the teams or locations that have dropped below a risk threshold. Managers then create actions against those risks, and the app measures whether the score actually improved after the action was completed.

• Enablement: Mango Signal requires a license and is disabled by default. An administrator enables it from the Apps Marketplace.
• What it is not: Mango Signal is not a survey tool. It sits on top of Surveys (and several other apps) and interprets the cross-signal context. It does not replace any source app — it reads their data as inputs.

2. Standing it up

1. Go to Apps Marketplace and enable Mango Signal. The app appears in the sidebar once enabled.
2. Open Mango Signal → Settings and confirm the defaults are acceptable. The engagement score, risk alerts, and close-the-loop flow are all on by default. The weekly manager digest is off by default.
3. Run a backfill from Settings to populate the last 90 days of historical signals. Without this step the dashboard starts empty and only fills as new upstream events arrive.
4. Verify that at least one source app (Surveys, Recognitions, Performance, News Feed, or Shift Feedback) is also enabled. Mango Signal works with whatever subset you have, but with zero source apps there are no signals to score.
5. Confirm the result: open the Mango Signal dashboard and check that an engagement score and at least one theme appear. Managers should see risk segments and recommended actions if any segment is below the threshold.

Silent-until-paired: The backfill step is the most common miss. Without it, a freshly enabled Mango Signal shows no data until enough new upstream events accumulate — which can take days if survey or feedback volume is low. The backfill sweeps Surveys, Recognitions, Performance, Shift Feedback, Attendance, Leave, Ideas, and Ballots from the last 90 days and is safe to re-run (it deduplicates on the source record).

3. How it fits together

Signals

A signal is a single normalized event from an upstream app: a survey analysis, a recognition post, a performance review, a shift-feedback submission, a news-feed interaction. Each signal carries a type (sentiment, recognition, performance, engagement, or theme), a numeric score, and exactly one subject — a user, a team, a location, or a named segment. Mango Signal never creates signals itself; it only reads what the source apps produce.

Five source apps feed the weighted engagement score today: Surveys, Recognitions, Performance, News Feed, and Shift Feedback. Four more — Attendance, Leave Management, Ideas, and Ballots — ingest signals that appear in the signal browse and export but do not yet factor into the score.

Individual employee sentiment is never surfaced in dashboards, analytics, or the AI assistant. Every score and theme stays aggregated to team, segment, or location scope. The only place individual signals are visible is the admin-only Signals tab, which is restricted to administrators.

Engagement health score

A 0-to-1 number (higher is healthier) computed from the rolling window of signals. The score blends five signal types with fixed weights:

• Sentiment — 40%
• Recognition — 20%
• Performance — 15%
• Passive engagement — 15%
• Themes — 10%

The score is computed for the company as a whole and separately for every team, location, manager, and segment. Scores are recomputed nightly. If the nightly run has not yet produced a score (for example, on a freshly enabled tenant), the dashboard computes one on demand so it is never blank.

Example: A company with three locations — Warehouse A, Office B, and Field Crew C — gets four scores: one company-wide and one per location. If Warehouse A’s 30-day sentiment average drops while its recognition volume stays flat, Warehouse A’s score falls below the threshold and appears in the risk panel, while the other two stay healthy.

Themes

Recurring concepts extracted from signals by a nightly clustering pipeline. Each theme has a label, a polarity (positive, neutral, or negative), a signal count, and an average score. The dashboard surfaces the top five themes over the last 30 days so managers can see what is driving the number, not just the number itself.

Risk segments

Any team, location, manager scope, or named segment whose rolling engagement score falls below the configured threshold (default 0.40). The dashboard sorts them worst-first. Segments backed by fewer signals than the minimum group size (default 3) are excluded entirely — they are not shown anonymized, they are simply not reported.

Actions

A manager-owned work item created in response to a risk. Each action has an owner, an assignee, a due date, a status (Open, In Progress, Complete, or Dismissed), and optionally a set of linked source signals. Action titles are limited to 200 characters.

When a manager sees a risk segment on the dashboard, the app recommends one to three concrete next actions. On the initial page load these are heuristic (instant); clicking Refresh with AI replaces them with AI-drafted recommendations via the AI assistant.

Close-the-loop

When an action is marked complete, three things happen:

1. A baseline engagement score for the action’s scope is captured.
2. A company-wide “you said, we did” update posts via News Feed or Broadcast (chosen automatically by audience reach — Broadcast for scopes reaching 100 or more people).
3. The contributors behind the action receive a direct email.

The baseline triggers a follow-up measurement after 30 days to compute the real delta. The close-the-loop toggle controls only the outbound posts and emails — completed actions still appear in the What Changed archive even when the toggle is off.

Impact measurement

Each completed action gets a pre-action baseline score and a post-action re-measurement after the configured window (default 30 days). The Impact tab shows per-action delta, delta percentage, and a status: measured, positive, negative, or inconclusive. It also rolls up counts and average delta across all measured actions. A minimum of 5 post-action signals is required before a measurement is considered sufficient.

Role-based access

Mango Signal has three access tiers. What each role sees in the sidebar:

Role	Tabs visible
All users	Dashboard, What Changed
Manager or above (and app admins)	Dashboard, What Changed, Actions, Impact, Analytics
Admin or above (and app admins)	All of the above plus Signals and Settings

App admin status can be granted per user, giving them the same access as a full administrator within Mango Signal without elevating their platform-wide role.

4. Running it

Reviewing the dashboard

The dashboard is the landing page. It shows the current company-wide engagement score, the top five themes, and (for managers) the worst three risk segments with recommended actions. Below the risk panel, the current user’s open actions appear in due-date order (up to 5). A sample-data banner appears when demo signals are loaded.

Creating and managing actions

1. On the dashboard, click a recommended action or go to Actions → New.
2. The form opens in a drawer. Fill in the title, description, assignee, due date, and optionally the scope (team, location, manager, or segment) and linked signals.
3. Save the action. The assignee receives an email notification if they are not the owner.
4. When the work is done, click Complete. This fires the close-the-loop flow.

Overdue actions trigger reminder emails: a one-time due-soon reminder 3 days before the due date, then an overdue nag every 3 days for up to 60 days. These are automatic and require no admin action.

Viewing impact

Go to Impact. Each row shows a completed action, its baseline and post-action scores, and the delta. The summary bar at the top shows counts of measured, positive, and negative outcomes plus the average delta. Sample-data rows are labeled and excluded from the summary averages.

Reviewing what changed (all employees)

The What Changed tab is visible to all users, not just managers. It lists completed actions with their title, description, completion date, and impact direction — but omits the owner and assignee names. This is the employee-facing “you said, we did” archive.

Browsing raw signals (admin only)

Signals shows every individual ingested signal with its source app, type, score, and date. This tab is restricted to administrators because it exposes individual named sentiment. Filter by source app, signal type, segment, or time window.

Using analytics

Go to Analytics. The page shows the current engagement score, total signal volume, at-risk segment count, open action count, and a daily signal-volume trend chart. Use the window selector to switch between 7, 30, 90, or 180 days. Analytics is available to managers and above.

Exporting and importing data

Managers can export raw signals to CSV or XLSX from the Analytics page. Exports cover the same selectable windows (7, 30, 90, or 180 days). Admins can also import signals from a CSV file using a downloadable template — useful for backfilling from a retired feedback tool.

Notifications

Mango Signal sends six types of notifications. Each is delivered by email and in-app bell under the mango_signal notification category, which users can mute from their notification preferences.

Notification	Who receives it	Frequency
Risk alert	The manager whose team dropped below the threshold	Deduplicated per manager — at most once every 7 days, stops after 90 days
Action assigned	The assignee (when different from the owner)	One-time on creation
Action due soon	The assignee	One-time, 3 days before the due date
Action overdue	The assignee and owner	Every 3 days while overdue, stops after 60 days
Close-the-loop update	Contributors behind the completed action	One-time on completion
Weekly digest	All managers with signals or open actions	Weekly (opt-in; off by default)

Mobile

Mango Signal has a read-only mobile dashboard at /m/apps/mango-signal. It shows the engagement score, top themes, and open actions assigned to the current user. Desktop users on a phone are redirected automatically; the redirect is session-sticky and can be overridden with ?desktop=1.

5. Settings

Setting	Where	Default	What it changes
Compute engagement health score	Settings	On	Turns the nightly score recompute and dashboard score display on or off. When off, signals still ingest but no scores are computed.
Send risk and theme alerts to managers	Settings	On	Controls whether managers receive email alerts when their team drops below the risk threshold.
Close-the-loop “you said, we did” flow	Settings	On	Controls whether completing an action triggers the company-wide post and contributor emails. Completed actions still appear in What Changed even when this is off.
Mango Signal AI assistant in Ask AI	Settings	On	Enables or disables the Mango Signal assistant in Ask AI.
Weekly engagement digest email to managers	Settings	Off	Sends managers a weekly email with the current score, trend, top themes, risk count, and their open/overdue action counts.
Engagement score window (days)	Settings	30	The rolling window of signals used to compute the engagement score.
Risk alert threshold (0..1, lower = more sensitive)	Settings	0.40	Segments scoring below this value appear as at-risk.
Minimum group size for segment reporting (anonymity floor)	Settings	3	Segments backed by fewer signals than this number are excluded from risk reporting entirely.

6. More help

• Mango Signal FAQ — specific setup and operating questions, limits, and troubleshooting.
• REST API — a scoped read/write API is available at /api/v1/mango_signal for native mobile and external integrators, tenant-scoped and role-gated to match the web surface.
• Ask AI — the assistant answers questions about Mango Signal from these articles.