---
title: "Reports and dashboards that drive decisions"
description: "Start from the decision A report is a product with customers. Before designing it, ask: who reads this, what decisions do they make, and what do they…"
url: https://optimizeall.com/learn/project-controls-with-ai/reports-that-drive-decisions
updated: 2026-10-05
---

Project Controls in the AI Era · Reporting, dashboards and data foundations · lesson 17 of 22 · 13 min

# Reports and dashboards that drive decisions

## Start from the decision

A report is a product with customers. Before designing it, ask: *who reads this, what decisions do they make, and what do they need to know to make them?* A sponsor needs a different view from a site manager.

| Audience | Key decisions | What they need |
|---|---|---|
| Board / sponsor | Funding, scope, risk appetite, escalations | Forecast vs approved, confidence range, top risks, decisions required |
| Project director | Resource allocation, recovery plans | Control-account performance, trends, critical path, change pipeline |
| Control account manager | Daily/weekly work priorities | Detailed variances, look-ahead, constraints |
| Client | Acceptance, change approvals | Progress, milestones, change status, issues needing client action |

## The one-page executive summary

```
PROJECT: ______   Period: ______   Overall status: [G / A / R with definition]
1. Headline (2–3 sentences): what changed and why it matters
2. Forecast: EAC ____ vs BAC ____ (+/- __%) | Finish ____ vs baseline ____ | P80 ____
3. Performance: CPI ____ (trend) | SPI(t) ____ (trend) | Contingency remaining ____
4. Top 3 variances: cause – impact – action – owner
5. Top 3 risks: change since last period – response
6. Decisions required from you: ______ by ______
```

"Decisions required" is the most important section and the most often omitted.

## RAG status with rules

RAG ratings are useful only if defined. Illustrative rule set:

| Status | Rule |
|---|---|
| Green | EAC within BAC; forecast finish within baseline; no unmitigated high risks |
| Amber | EAC within BAC + contingency, or forecast slip ≤ 4 weeks, recovery plan in place |
| Red | EAC exceeds BAC + contingency, or slip > 4 weeks, or no credible recovery plan |

Without rules, people report "green" until it becomes impossible, producing the famous "watermelon" project: green outside, red inside.

## Dashboard design principles

- **Show trends, not snapshots.** A CPI line over 12 periods beats a single gauge.
- **Compare with something.** Actual vs baseline, forecast vs approved, this period vs last.
- **Use the right chart.** S-curves for cumulative PV/EV/AC; bar charts for variance by control account; line charts for indices; tables for top risks.
- **Keep it sparse.** A few metrics that matter, with the ability to drill down.
- **Explain.** Every chart needs a sentence saying what it means.
- **Timestamp and source.** Show data date and data source on every page.

## The EVM S-curve chart

Plot cumulative PV (baseline), EV and AC over time, and extend the forecast (EAC) to completion. The gap between EV and PV shows schedule position in money; the gap between EV and AC shows cost efficiency. Add the BAC line and EAC point.

## Worked example

*Illustrative.* A fictional telecoms operator in Pakistan ran 30 tower-upgrade projects with 20-page reports each. Executives read none of them. The PMO replaced them with a portfolio dashboard: a scatter of CPI vs SPI(t) for all projects, a table of the ten worst EAC movements, and a one-page summary for any project in red. Review meetings shortened and focused on the handful of projects needing decisions.

## Narrative writing

Numbers need words. A strong narrative:

1. States the headline first.
2. Explains the cause, not only the symptom.
3. Quantifies impact on cost, schedule and risk.
4. Describes actions with owners and dates.
5. Asks explicitly for decisions.

## Common mistakes

- Reports designed around available data rather than decisions.
- Undefined RAG status.
- Dozens of KPIs with no prioritisation.
- No data date, so readers do not know how current it is.
- Charts without interpretation.

## AI in reporting

Generative AI can draft narratives from structured data in seconds. Use it to create a first draft, then have the controls lead verify every number against the source, check causes with the owners and add judgment. Label AI-assisted content according to your organisation's policy.

## Quick self-check

Hand your latest report to a colleague who does not know the project and give them two minutes. Ask them to tell you the forecast, the main problem and what decision is needed. If they cannot, the report is not doing its job, however accurate the numbers are. Rewrite the headline and the decisions section first; those two parts do most of the work.

## Hands-on: RAG by formula, not by feeling

```text
Named inputs: EAC, BAC, Contingency_remaining, Slip_weeks, Recovery_plan (Y/N), Unmitigated_high_risks (count)
RAG =IFS(
      OR(EAC>BAC+Contingency_remaining, Slip_weeks>4, AND(Slip_weeks>0, Recovery_plan="N")), "Red",
      OR(EAC>BAC, Slip_weeks>0, Unmitigated_high_risks>0), "Amber",
      TRUE, "Green")
```

Publish the rule on the report itself so readers know what each colour means. Thresholds are illustrative; set your own in the project's controls plan.

## Hands-on: two core visuals in Power BI (or Excel)

Model: one table `EVM` with `Period`, `Account`, `PV`, `EV`, `AC` (cumulative) and a `DataDate` table.

```text
DAX measures
PV Cum   = SUM ( EVM[PV] )
EV Cum   = SUM ( EVM[EV] )
AC Cum   = SUM ( EVM[AC] )
CPI      = DIVIDE ( [EV Cum], [AC Cum] )
SPI      = DIVIDE ( [EV Cum], [PV Cum] )
```

Visual 1: line chart, axis `Period`, values `PV Cum`, `EV Cum`, `AC Cum`, plus a constant line at BAC. Visual 2: line chart of `CPI` by `Period` with a constant line at 0.95. Add a card showing the data date. In Excel, the same visuals come from a pivot table and a line chart.

## Prompt template: executive narrative (approved enterprise AI tool)

```text
Role: project controls analyst. Write a 120–150 word executive headline for the one-page summary.
Use ONLY the table provided. Structure: (1) headline, (2) cost forecast with range, (3) schedule forecast,
(4) main driver, (5) decisions required with dates. Cite the table column for each figure in brackets.
If a cause is not in the table, write "cause to be confirmed". Use British English; no adjectives like "significant"
unless a threshold is stated.
```

Tools such as Microsoft 365 Copilot, ChatGPT Enterprise or Claude can run this template if your organisation has approved them; the controls lead verifies every number and approves the text.

## How to measure success

- The two-minute test passes with a reader outside the project.
- Every report carries a data date, a source and a decisions-required section.
- Time from data date to report issue (target falling without loss of accuracy).

## Video lecture: Reports and dashboards that drive decisions

Lecture coming soon · 9 chapters · about 8 minutes. Read the full transcript below.

1. Twenty pages nobody reads
2. Why it matters
3. The concept: start from the decision
4. The one-page executive summary
5. Worked example one: RAG with rules
6. Worked example two: a telecoms portfolio in Pakistan
7. Watch me do it: an S-curve and trend in Power BI or Excel
8. Dashboard principles and narrative writing
9. Common mistakes, recap and try this now

## Lecture transcript

### Twenty pages nobody reads

A telecoms operator I'll describe in a moment was producing twenty-page monthly reports for thirty projects. Six hundred pages a month. Executives read none of them. Not because they were lazy, but because the reports answered questions nobody was asking. In this lecture you'll learn to design reports the way you'd design a product: starting from the reader and the decision they need to make. We'll build a one-page executive summary, write RAG rules that stop the 'watermelon' effect, apply dashboard principles that show trends instead of snapshots, and write narratives that ask for decisions. By the end you'll be able to take any report you produce and redesign it so a busy sponsor can understand it in two minutes.

### Why it matters

Why does this matter? Because a controls system only creates value when someone acts on it, and the report is usually the moment of action. If the report doesn't make the decision obvious, the whole control cycle stalls at the last step. Second, status colours without rules are dangerous. People report green until it's impossible, and you get the famous watermelon project: green on the outside, red on the inside. And third, attention. A sponsor might oversee dozens of projects. They'll give yours a couple of minutes. If the headline, the forecast and the decision needed aren't clear in those two minutes, you've lost them.

### The concept: start from the decision

Here's the key idea. A report is a product with customers. Before you design it, ask three questions. Who reads this? What decisions do they make? And what do they need to know to make them? A board or sponsor decides on funding, scope, risk appetite and escalations, so they need forecast versus approved, a confidence range, the top risks and the decisions required. A project director allocates resources and approves recovery plans, so they need control-account performance, trends, the critical path and the change pipeline. A control account manager needs detailed variances and a look-ahead. And a client needs progress, milestones and anything waiting for their action. Think of it like a restaurant: the diner wants the menu, the chef wants the order tickets, the owner wants the margins. Same kitchen, different views.

### The one-page executive summary

Here's the template. Section one, the headline: two or three sentences saying what changed and why it matters. Section two, the forecast: EAC versus budget with the percentage, forecast finish versus baseline, and P80 if you have a risk analysis. Section three, performance: CPI with its trend, schedule performance in time with its trend, and remaining contingency. Section four, the top three variances: cause, impact, action and owner. Section five, the top three risks, with what's changed since last period. And section six: decisions required from you, and by when. That last section is the most important and the most often omitted. If a report doesn't ask for anything, don't be surprised when nothing happens.

### Worked example one: RAG with rules

Let's make RAG status objective, with an illustrative rule set. Green: EAC within budget, forecast finish within baseline, and no unmitigated high risks. Amber: EAC within budget plus contingency, or a forecast slip of up to four weeks, with a recovery plan in place. Red: EAC beyond budget plus contingency, or a slip of more than four weeks, or no credible recovery plan. Now try it. A project has an EAC three per cent over budget but within contingency, a two-week slip and a recovery plan. What colour? Amber. Not 'green because we'll recover'. Not 'red because the sponsor might panic'. Amber, because the rules say so. Once the rules exist, the colour stops being a negotiation and starts being information.

### Worked example two: a telecoms portfolio in Pakistan

Now the realistic example. A fictional telecoms operator in Pakistan ran thirty tower-upgrade projects, each with a twenty-page monthly report. The PMO replaced them with a portfolio dashboard. One scatter chart: every project plotted by CPI on one axis and schedule performance in time on the other. Projects in the bottom-left quadrant, over cost and late, jump out immediately. Next to it, a table of the ten biggest EAC movements since last month. And a one-page summary, in the template you've just seen, for any project in red. That's it. Review meetings got shorter and focused on the handful of projects that needed decisions. The data didn't change. The design did.

### Watch me do it: an S-curve and trend in Power BI or Excel

Let me show you the two charts I put on almost every controls dashboard, and you can build them in Excel or Power BI. First, the S-curve. Cumulative planned value, earned value and actual cost by period, from one table with a period column. Then I add the budget line and extend a dashed forecast from today's actual cost up to the EAC. The gap between earned value and planned value shows schedule position in money. The gap between earned value and actual cost shows cost efficiency. Second, the CPI trend over the last twelve periods, with a shaded band at nought point nine five, so a reader instantly sees whether it's inside tolerance and which way it's heading. And on every page, a footer with the data date and data source. Without it, nobody knows how current the numbers are.

### Dashboard principles and narrative writing

A few design principles. Show trends, not snapshots: a CPI line over twelve periods beats a single gauge. Always compare with something: actual versus baseline, forecast versus approved, this period versus last. Use the right chart: S-curves for cumulative values, bars for variances by account, lines for indices, tables for top risks. Keep it sparse, with drill-down for detail. And explain: every chart needs one sentence saying what it means. Then the narrative, because numbers need words. State the headline first. Explain the cause, not only the symptom. Quantify the impact on cost, schedule and risk. Describe actions with owners and dates. And ask explicitly for decisions. Generative AI can draft that narrative from structured data in seconds; the controls lead verifies every number and owns the words.

### Common mistakes, recap and try this now

The common mistakes: reports designed around whatever data is available rather than the decisions needed, undefined RAG status, dozens of KPIs with no priority, no data date, and charts without interpretation. So, to recap. Start from the reader and the decision. Use a one-page executive summary with a decisions-required section. Give RAG status explicit rules. Show trends and comparisons, with the data date on every page. And write narratives that go headline, cause, impact, action, decision. Here's a quick test: hand your latest report to a colleague who doesn't know the project, give them two minutes, and ask for the forecast, the main problem and the decision needed. Your try-this-now: redesign a report you produce or receive into the one-page template, with a decisions-required section and objective RAG rules.

## Key takeaways

- Design reports around audiences and the decisions they make.
- The executive summary must include forecasts, confidence, top variances and risks, and decisions required.
- Define RAG rules to avoid 'watermelon' reporting.
- Dashboards show trends and comparisons, cite the data date and explain each chart.

## Try it

Redesign a report you receive or produce into the one-page executive template. Add a 'decisions required' section and objective RAG rules.

- [Previous: Change control and baseline management](https://optimizeall.com/learn/project-controls-with-ai/change-control)
- [Next: Data foundations: integration, quality and tools](https://optimizeall.com/learn/project-controls-with-ai/data-foundations)
- [All lessons of Project Controls in the AI Era](https://optimizeall.com/learn/project-controls-with-ai)