Custom business dashboard: BI tool or built into your software?

A business dashboard exists to support decisions, not to display charts. For most small and mid-sized companies, an off-the-shelf BI tool (Power BI, Looker Studio, Metabase) connected to existing software is enough and costs little. A custom dashboard built into your own business application is worth it in three specific situations: when people need to act on the numbers and not just read them, when the dashboard is opened to customers or partners with strict permissions, or when your calculation rules exist nowhere except in your team's heads and spreadsheets.

Most pages on this topic are agency service menus or vendor product tours. This one does not rank BI tools: they are good products and we often recommend them. It helps you choose between three paths (spreadsheet, BI tool, built-in dashboard) and understand what each one really costs.

What a business dashboard is, and what it is not

A business dashboard brings together, on one screen, the handful of numbers that trigger a decision. Each metric carries four pieces of information: its current value, its target, the gap between them, and its trend. A number without a target is trivia. A number with a target is a warning light.

It should not be confused with:

• reporting, which details everything that happened and is read on demand;
• exploratory analysis, where an analyst slices data to answer a new question;
• a work queue, which lists the cases someone has to handle.

All three are useful. A dashboard that tries to be all three ends up with forty charts that nobody reads.

Metrics: fewer, and tied to an action

One audience, one dashboard

Leadership, sales and operations do not look at the same numbers or at the same pace. Some examples to adapt to your business:

Leading indicators, not just results

Monthly revenue tells you what already happened. The number of quotes sent this week tells you what next quarter will look like. A good dashboard mixes both: a few outcome metrics, and at least two or three leading metrics that leave time to react.

A written definition for every number

"Revenue", "active customer" and "margin" mean different things to finance, sales and operations. Before choosing any tool, write down for each metric: its formula, its source, how often it refreshes and who owns it. This one-page metric glossary is the most valuable document of the whole project, whatever technology you end up with.

Three paths, three uses

Spreadsheets: the perfect first draft

An Excel dashboard is the best prototype you can get: it forces you to pick metrics and find their sources. It hits its limits when several people feed it, when someone rebuilds it by hand every Monday, or when a broken formula becomes expensive. Our guide on replacing Excel with a custom application lists the warning signs.

BI tools: what they do very well

BI tools are designed to read data that already lives elsewhere and make it readable. Their public prices, checked on 6 November 2026 on the official pages, excluding tax:

They bring what a new build cannot offer on day one: dozens of ready-made connectors, rich charts, and free exploration by anyone comfortable with data. If your numbers live in an accounting package, a CRM and a database these tools can read, and your need is to view them, start there. The real cost is rarely the licence. It is the time spent modelling data and fixing reports every time a source system changes.

The built-in dashboard: when the screen has to do something

A custom dashboard is not a hand-made copy of a BI tool. It is the home screen of your business application: it shows the numbers where the work happens and lets people act without switching tools.

Four situations where custom pays off

1. The number has to trigger an action

A manager sees that twelve quotes have had no follow-up for fifteen days. In a BI tool, she writes down the list and goes elsewhere to chase them. In a built-in dashboard, she clicks the metric, sees the twelve deals, reassigns them or sends the reminder. The gap between "I see it" and "I fix it" disappears, and that is usually what decides whether people adopt the tool.

2. Your teams already enter the data in the application

When your staff record jobs, work orders or site visits in a business application, the dashboard is its natural extension: no sync, no lag, numbers that are current to the second. No pipeline to maintain between two systems, and no "which figure is right" meetings.

3. You open it to customers or partners

Showing each customer their orders, lead times or usage, or each franchisee their own results, requires strict permissions: everyone sees only their own rows, and nothing may leak between accounts. BI tools can do this, but external embedding usually needs a specific licence or capacity that should be priced first. In a custom client portal, those permissions are the application's own, designed and tested once.

4. Your calculation rules are your business

Job margin including change orders, commissions across a reseller network, service levels under customer-specific contract terms: when a calculation is complex and unique to you, copying it into a BI tool means maintaining it twice. Writing it once in the application guarantees that the number on screen is the number you invoice from.

When not to go custom

Custom is the wrong choice if:

• your data lives in mainstream software that Power BI or Looker Studio already reads;
• the main need is ad hoc analysis by someone comfortable with data;
• you have not settled on your metrics yet. Start with a spreadsheet and you will learn what is worth building;
• nobody is ready to make decisions from the numbers. No tool fixes that.

Our build vs buy guide applies the same reasoning to software in general.

The hybrid route, often the most sensible

You do not have to choose. A common setup: the business application handles action screens and the few metrics people need every day, while a BI tool reads its database for monthly leadership analysis. Metabase, for instance, lets you embed its charts in another application; note that, according to its pricing page, people who view embedded charts count as users on paid plans.

The condition for this to work: one definition per metric, and one system of record for each piece of data.

The real work is the data, not the charts

Whatever the tool, the display is rarely the expensive part. The work is in:

• sources: where each number lives, which system owns it, how to read it (API, export, direct database access);
• quality: duplicate customers, half-filled records, product codes that changed mid-year;
• freshness: yesterday's cash figure is fine, a late delivery must be visible within the hour;
• edge cases: credit notes, cancellations, partial shipments, fiscal year changes.

Plan a parallel run: the old and new dashboards live side by side for a few weeks, and every gap is explained. That is what builds trust, and a dashboard nobody trusts is a dashboard nobody opens.

Permissions and personal data

Sales and HR dashboards show personal data: customer names, individual performance, absences. In the EU and UK, GDPR requires that each role sees only what it needs, and similar principles apply under many other privacy laws. In practice:

• permissions by role and, where needed, by scope (branch, team, customer);
• exports restricted to people who genuinely need them;
• an access log for sensitive views;
• aggregated metrics whenever named detail adds nothing to the decision.

What each path costs

There is no reliable public price for "a custom dashboard", and we will not invent one. Cost depends on four factors:

1. how many data sources you connect, and how clean they are;
2. how complex your calculation rules are;
3. permissions: a single internal role, or segregated external customers;
4. actions available from the screen.

To compare fairly, look at three years: BI licences multiplied by users, plus the time spent maintaining reports, on one side; development, hosting and maintenance on the other. When the dashboard is part of a business application you already plan to build, its marginal cost is usually small. Our guide to custom business application costs breaks down the line items, and our article on subscription vs upfront build explains how to spread the cost like a licence.

Common mistakes

• Starting with the tool instead of the decisions to make.
• Showing everything measurable instead of what drives action.
• Letting two definitions coexist for the same metric.
• Forgetting the target: a number without a threshold triggers nothing.
• Naming no owner for each metric.
• Ignoring the shop floor: a site supervisor will not open a desktop report; give them a tablet and three numbers.

Where to start

1. List the five decisions you make every week and the number you are missing for each.
2. Write the definition of each metric, with its source.
3. Run it in a spreadsheet for a month: anything nobody looks at comes off the list.
4. Pick the path using the grid above: BI tool if you view, built-in if you act, hybrid if both.

If built-in wins, see the screen before you pay for it. We build a free visual prototype, presented over video within 5 working days: your team sees its dashboard, with its real metrics, before a line of code is written. Our article on prototypes, MVPs and proofs of concept explains what a prototype can and cannot prove.