Skip to content

Odduel

We put the odds in your favour

Insights

Custom dashboard development for real-time business data

A business owner's guide to custom dashboard development: when Power BI is enough, what the work involves, where to start, best practices, cost drivers and how to measure success.

An operations manager in a warehouse office looking at a large monitor of live business indicators, with a tablet in hand

Custom dashboard development is the design and build of a screen, made for one company, that pulls live data from its own systems and shows the few figures each person needs to act on today. It makes sense when off-the-shelf reporting tools cannot reach your data cleanly, cannot show it in the way decisions are actually made, or have to be rebuilt by hand every week. This guide explains what the work involves, when a tool such as Power BI is enough, where to start and how to tell whether the dashboard is working.

Last updated: 21 November 2026.

What is a custom business dashboard?

A custom business dashboard is a web application that collects data from several sources, such as your sales system, online store, accounting software, booking tool or machines, and turns it into a small set of live indicators, alerts and drill-downs designed around your own processes. Unlike a generic report, it is built for specific people and specific decisions: the warehouse lead sees today’s backlog, the managing director sees margin by line, the sales team sees the pipeline that needs action.

“Real-time” means different things in different businesses. For a logistics operation it may mean seconds. For a management team it often means “updated every hour, and never older than this morning”. Deciding which figures truly need to be live is one of the first and most useful design decisions.

At Odduel, a creative and technology agency in Barcelona and São Paulo, dashboards are part of our application development service: we treat them as products with users, not as charts added at the end.

In short: a custom dashboard is worth building when your data lives in several systems, your key indicators are specific to your business, and people currently spend hours assembling reports that are out of date by the time they are read. Start with the decisions, not the data.

Power BI or a custom dashboard?

Business intelligence tools such as Power BI, Looker Studio or Tableau are excellent for analysts exploring data. A custom dashboard is better when the screen must be part of daily operations, used by people who are not analysts, and connected to actions in your own systems.

QuestionBI tool (Power BI, Looker Studio, Tableau)Custom dashboard
Who uses itAnalysts and managers comfortable with reportsAnyone in the company, including the shop floor or field staff
Data sourcesStrong connectors to common systemsAny system, including old databases and internal APIs
Real-timeOften scheduled refreshesCan be truly live where it matters
ActionsMostly read-onlyCan trigger tasks, approvals or messages
Look and feelThe tool’s interfaceYour brand, your workflow, your devices
Access rulesThe tool’s licence and permission modelRules modelled on your organisation and clients
Time to first versionFast if data is cleanLonger, because it is software
Ongoing cost driverLicences per userHosting and maintenance
BI tools compared with custom dashboard development

A practical rule: if a BI tool can show the figures cleanly and the users are happy working inside it, use it. Build custom when people need the information inside their working day, on their devices, with permissions and actions a BI tool cannot model. Many companies end up with both: BI for analysis, a custom dashboard for operations.

What does custom dashboard development involve in practice?

A dashboard project has five layers. Each one needs attention, and the visible screen is only the last.

  1. Decisions and indicators. List the decisions each role makes, then the indicator that informs each one, with an exact definition. “Revenue” is not a definition; “invoiced revenue excluding VAT, by invoice date, after credit notes” is.
  2. Data sources and access. Identify where each figure lives, how it can be read (API, database, export, webhook) and who owns that system.
  3. Data pipeline. Build the connectors that read, clean, join and store the data, with a record of when each figure was last updated.
  4. Business logic. Calculate the indicators once, in one place, so every screen and export shows the same number.
  5. Interface. Design screens for each role, on the devices they use, with drill-downs from a figure to the records behind it.

Our guide on how to build internal tools covers the wider family of software a dashboard often belongs to, and our list of business app examples shows where dashboards sit next to booking, portals and field apps.

Where should a company start with custom dashboard development?

Start small, with one role and one decision that costs money when it is made late or with the wrong figures.

  1. Pick the first user. For example, the operations manager who rebuilds a spreadsheet every Monday.
  2. Write down five to eight indicators that person needs, with definitions and the target or threshold for each.
  3. Audit the data. Check that each figure can actually be read from a system, how fresh it is and how clean it is. This step often reveals missing fields or duplicate records that must be fixed first.
  4. Sketch the screen on paper with the user before any design work. Which figure is read first? What happens when something turns red?
  5. Build a first working version connected to real data, used daily for a few weeks.
  6. Extend by role, reusing the same pipeline and calculations.

This order keeps the project grounded. The most common failure is to start from “all the data we have” and design a screen of thirty charts that nobody opens.

What are the best practices for a real-time management dashboard?

The best dashboards are boring to look at and fast to act on. These practices make the difference:

  • One screen, one question. Each view answers a question such as “Are we on track this week?” rather than displaying everything.
  • Context on every figure. A number alone means little; show it against a target, last period or a threshold.
  • Freshness visible. Show when each figure was last updated, so nobody acts on stale data without knowing.
  • Drill-down to records. A click on an indicator should lead to the orders, tickets or jobs behind it.
  • Alerts over staring. Send a message when a threshold is crossed rather than expecting people to watch a screen.
  • Accessible design. Do not rely on colour alone for status, keep text readable and make the dashboard usable by keyboard; the WCAG 2.2 guidelines are the reference.
  • Mobile where people move. Field and warehouse staff need a phone layout, not a shrunken desktop.
  • Permissions by role. People see only the data they should, which matters under the GDPR when dashboards show customer or employee records.

Which technologies are typically used?

The technology should follow your existing systems, not fashion. A typical modern stack has three parts:

LayerCommon choicesWhat to decide
Data pipelineScheduled jobs, webhooks, message queuesHow fresh each figure must be
Storage and logicA relational database such as PostgreSQL, sometimes a data warehouseWhere indicators are calculated once
InterfaceA web application built with a framework such as ReactDevices, roles and hosting in the EU
Typical technology layers in a custom dashboard

Hosting matters in Europe. If the dashboard holds personal data, keep it on servers and with providers that meet your obligations under the GDPR, log who accessed what, and agree a data processing agreement with your developer.

How long does it take and what drives the cost?

A first version for one role, connected to two or three systems with good APIs, is usually a matter of weeks. A dashboard that joins many sources, cleans messy data and serves several roles with different permissions is a project of months. Our article on the app development timeline breaks down each phase.

The cost is driven by:

  • The number and quality of data sources. Clean APIs are quick; old databases and spreadsheets take longer.
  • Freshness. Truly live data costs more to build and run than hourly updates.
  • Roles and permissions. Each role with its own views and rules adds design and testing.
  • Actions. Read-only screens are simpler than screens that trigger approvals or updates.
  • Data cleaning. Often the largest hidden share of the work.

A fixed quote follows a short brief listing the users, indicators and systems involved.

How do you measure whether a dashboard works?

A dashboard works when it changes behaviour. Measure it like any product:

  1. Use. Do the intended people open it, and how often? A dashboard opened once a month has failed.
  2. Time saved. Compare the hours spent assembling reports before and after.
  3. Decision speed. Track how quickly problems are spotted and acted on, such as late orders or low stock.
  4. Trust. Ask users whether they believe the numbers. If they keep their own spreadsheet “just in case”, something is wrong.
  5. Outcome indicators. Watch the business figures the dashboard was built to improve.

Review these a month after launch and again each quarter, removing views nobody uses and adding the ones people ask for.

What are the most common mistakes?

  • Designing screens before agreeing the definition of each indicator.
  • Showing every available metric instead of the few that drive decisions.
  • Ignoring data quality until the numbers on screen are wrong.
  • Calling everything “real-time” and paying for freshness nobody needs.
  • Forgetting mobile users and permissions.
  • Treating launch as the end instead of the start of regular review.

Questions people ask

When is a custom dashboard better than Power BI?

A custom dashboard is better when the information must sit inside daily operations, be used by people who are not analysts, combine systems that BI connectors cannot reach cleanly, or trigger actions such as approvals. If a BI tool already shows the right figures and users work happily inside it, there is little reason to build custom.

What data sources can a custom dashboard connect to?

Almost any source that can be read: sales and CRM systems, online stores, accounting and ERP software, booking tools, spreadsheets, internal databases, machines and sensors. Systems with modern APIs are quickest to connect. Older systems may need exports or direct database access, which adds time and should be checked during the data audit at the start.

How long does custom dashboard development take?

A first version for one role, connected to two or three systems with good APIs, usually takes a few weeks. Dashboards that join many messy sources and serve several roles with different permissions take a few months. Starting small with one user and extending by role keeps the timeline predictable and the dashboard useful early.

Does a business dashboard have to be real-time?

No. Only figures that drive immediate action, such as stock, queues or machine status, usually need to be live. Many management indicators are perfectly useful when updated hourly or daily. Deciding freshness figure by figure keeps the build simpler and the running cost lower, without making the dashboard less useful.

Who owns the dashboard once it is built?

You should. Agree in the contract that the source code, data and documentation belong to your company, and that the dashboard runs on hosting you control or can move. A good developer also documents the indicator definitions and data pipeline so another team could maintain it later.

Write down the one decision in your business that is made too late because the figures take too long to assemble: that is your first dashboard. When you want it built around your own systems and people, see our application development service or tell us about the project.

More articles

More from our Insights.

Explore all insights

Get in touch

Let’s change your odds.

Odds do not change on their own. If you have something to build, launch or fix, tell us about it and we will show you how we would approach it.

    Need an NDA before you share details? Start a project

    Chat with us