Every Monday, someone in your company exports data from two or three systems, pastes it into a spreadsheet, fixes the columns that came out wrong and emails a report. By the time managers read it, the numbers are a week old, and nobody can click through to the customer behind a figure.
A live dashboard replaces that routine. Done well, it shows the same numbers every day without anyone compiling them, and each number links to the records behind it. Done badly, it is a page of charts nobody opens after the first month.
This post is about the difference: which reports to replace, how to design the dashboard, and what it costs.
Which reports to replace first
Not every report should become a dashboard. Start with the ones that tick these boxes:
Compiled by hand on a schedule. Weekly or daily, by the same person, from the same sources.
Read by people who act on it. A manager moves staff, chases a payment or calls a customer because of it.
Built from systems you control, or that have an API you can read.
Same questions every time. "What's overdue? What's new? What's stuck?"
Keep the rest as they are. A quarterly board report with commentary is still a document. A one-off analysis is better done in a spreadsheet or a BI tool. For how to choose which numbers matter in the first place, see using your own data for decisions.
Design it around actions, not charts
The dashboards that get used share a few traits:
| Manual report habit | Dashboard that gets used |
|---|---|
| Totals for the whole company | Totals per team or person, filtered to the viewer's role |
| A chart of last month | A list of what needs action today: overdue, stuck, new |
| Numbers with no source | Every number clicks through to the records behind it |
| Emailed to everyone | Each role sees its own page; alerts only when a threshold is crossed |
| Rebuilt each week | Defined once, tested, and updated with the data |
Four to six numbers per page is plenty. If a number never changes anyone's day, remove it.
An example: results without a results report
Marafiq runs exam-prep courses for facilities-management professionals, with more than 1,700 learner accounts, 25 published courses and 300+ active enrolments. Quizzes are timed and graded automatically, and certificates are issued as PDFs with a serial number and a QR code when a learner passes.
Because grading happens in the platform, a learner's score exists the moment they submit. Nobody marks papers and types results into a sheet, and nobody compiles a weekly list of who passed. Staff see enrolments, progress and results in the admin, from the same records that drive the courses. The build is in the Marafiq case study.
That is the general lesson. The best dashboards are not bolted onto a reporting process. They sit on top of software that already records the work as it happens.
Where the data comes from
This decides most of the effort:
Your own platform. The cheapest case. Dashboards read the same database, with the same permissions, often as pages in the admin you already use. On Laravel projects we usually build them with Filament widgets.
Third-party tools with APIs. Accounting, CRM, ads, analytics. A scheduled job pulls the numbers into your database once a day. More work, and each API has its quirks and limits.
Spreadsheets and email. The hard case. The process itself needs to move into a system first, or the dashboard inherits the copying. We cover that step in when to replace a spreadsheet.
What moves the cost
Number of sources, and whether they have usable APIs.
Data cleanup: duplicates, missing links between records, inconsistent statuses.
Roles: one shared page is simpler than per-team views with different permissions.
Alerts: email or Slack when a threshold is crossed adds rules and tests.
History: trends need data stored over time, not just current values.
A dashboard inside a platform we already maintain is often a small piece of work. A dashboard that pulls from several outside tools and replaces a spreadsheet process is a project, scoped in a short discovery. If an off-the-shelf BI tool on top of your existing data does the job, we will tell you, and it will cost less.
Next step
Send us the last weekly report your team compiled by hand, with sensitive figures removed, and the list of systems it comes from. We will reply with what a live version would look like and how we would scope it. Request a quote.