What is a dashboard and how do you build one? Choosing KPIs for managers

In a management meeting someone says, “We should have a dashboard,” and the job is handed to IT or a developer. A few weeks later you get a page lined with dozens of charts: colourful, but not the screen anyone opens first thing in the morning. The problem is not the charts; it is that nobody asked, “Which decision will this screen help me with?” The practical answer to “what is a dashboard?” starts here: building a dashboard means making a list of decisions before drawing anything.

A dashboard is a visual summary screen that brings together the most important indicators needed to understand the state of a business or company. It works like the instrument panel in a car: it shows speed, fuel and warnings without you having to take the engine apart. In this article we explain what a dashboard is for, how it differs from an admin panel, how to choose indicators and how to build one.

What is a dashboard for?

• Showing the situation at a glance: without waiting for a report, you see where the work stands the moment you open it.
• Flagging deviations early: an indicator below target catches the eye long before the month-end report.
• Reminding you of pending work: requests awaiting approval and upcoming tasks are gathered on the same screen.
• Creating a shared truth: when everyone looks at the same figure, arguments over “which file is up to date?” become rarer.

These four functions are linked. Without a shared figure, deviations are not spotted early; and without visible pending work, what to do after a deviation remains unclear. What makes a dashboard valuable is not how good the charts look but how much it shortens the path from a number to an action. If the screen doesn’t just say “you are below target” but takes you to the relevant process in one click, it really does its job.

What types of dashboards are there?

Dashboards come in three main types, depending on their purpose. Knowing which one you need saves you from unnecessary charts:

• Operational dashboard: tracks day-to-day work and shows current information such as pending approvals, overdue tasks and today’s workload.
• Management dashboard: shows status and deviation against targets; topics such as sales, budget, stock and team capacity are gathered here.
• Analytical dashboard: used to dig into the cause of a problem; it needs filtering and drill-down and is meant for investigation rather than a daily glance.

For most companies the starting point is the management and operational layers. The analytical layer makes sense once these two layers run smoothly.

What is an admin panel? How it differs from a dashboard

“Management panel” is used in two senses. The first is the indicator panel a manager looks at, in other words the dashboard. The second is the administrator interface of a piece of software used to manage users, permissions and settings (the admin panel). The two get mixed up in search results; here is the difference:

How do you build a dashboard? Step by step

Building a dashboard does not start with opening software and adding charts. The order below applies whatever tool you use; the same six steps are followed in Excel and in dedicated software.

1. Write down the decisions. Which decisions will this screen help with? For example, “Which order should we bring forward?” or “Which approval is waiting?”
2. Link indicators to decisions. Next to each indicator, write which decision it affects; remove any that cannot be linked.
3. Set definitions and sources. The same indicator must be calculated the same way for everyone and fed from a single data source.
4. Separate by role. Senior management, department managers and employees do not see the same screen; each sees their own indicators.
5. Add thresholds and alerts. Highlight “below target” or “overdue” with colour and notifications.
6. Track usage. If nobody opens it after a week, simplify the screen; don’t add more.

How do you choose indicators (KPIs) for managers?

Choosing indicators is the hardest and most valuable part of a dashboard. Showing all the data you have is easy; putting only the indicators that drive decisions on the screen takes discipline. Filter a candidate indicator with these questions:

• If this number changes, will I do something?
• Can I get the number reliably and regularly?
• Does everyone agree on how this number is defined?
• Is there a target or previous period to compare the number with?
• Which role should see this number, and which should not?
• When the number gets worse, what task falls to whom?

After this filtering, a few indicators remain for each role. The table below gives an example of how indicators can be split by role; your own indicators will depend on your industry and processes.

A dashboard that tries to show everything is a dashboard where nobody can see anything.

Consider a fictitious example: every morning, the general manager of a manufacturing company first looks at three things: sales against target, materials at critical stock levels and the tasks waiting for their approval. In the same system, the warehouse supervisor sees only their pending shipments and stock alerts, and the salesperson sees their own quotes and upcoming meetings. All three are fed from the same data, but none of them sees the others’ screens. The screen stays simple, sensitive data does not circulate unnecessarily and everyone can work out their first task of the morning from the screen.

Common mistakes when setting up a dashboard are usually these: starting with the design and settling definitions later, showing the same screen to all roles, showing only numbers without alerts and not writing down who looks at what.

Do you need dashboard software, or is Excel enough?

A summary sheet and a few charts in Excel make a good dashboard for work where one person looks at it and the data sits in a single file. Almost everyone starts this way at some point. The limit appears when more than one person starts using the screen: the data has to be refreshed by hand every time, different versions appear as the file circulates by e-mail, separating sensitive data becomes impossible because everyone sees everything, and it does not open comfortably on a phone. A “daily summary” message sent on WhatsApp, meanwhile, is a static image; you cannot drill into the details or go back and compare with the past.

This is where dashboard software comes in: it reads data from the connected source, shows different screens by role and delivers alerts to the right person. Software does not replace choosing indicators, however; putting the wrong indicators into an empty tool is as harmful as a wrong report.

How does Dashboard Management work in KurumsalPortal?

The Dashboard Management module is KurumsalPortal’s role-based dashboard layer. It publishes personal dashboards that show employees, department managers and senior management only the relevant data. A dashboard contains the KPIs defined for the role, pending and upcoming tasks, alerts from the notification centre, performance indicators and process shortcuts; users can go from a deviation they see straight to the relevant process and turn the decision into action.

The set-up order matches the steps above: data sources are connected, roles and KPIs are defined, personal dashboards are published and deviations are tracked. Which indicators go on the dashboard is agreed with you at this stage. Bringing data from another system such as an ERP into the dashboard is integration work; its scope is clarified during the project. KurumsalPortal runs on the customer’s own server and employees use it from their phone’s browser, so there is no app to install to see the dashboard.

We recommend starting the steps in this article with a list of decisions. To see how role-based dashboards are set up, visit the Dashboard Management page; to see the whole 33-module platform, explore KurumsalPortal.