Dashboard design & build
Designed around the decisions somebody makes each week, not around the columns the database happens to have.
Business Intelligence
The numbers, in front of the people who act on them.
A report nobody opens is not reporting. We build the dashboards, the metrics layer underneath them and the self-serve access that stops every question becoming a ticket for your data team.
A worked example
One question,
followed all the way through.
Somebody asks a question in a Monday meeting. Here is everything that has to be true for it to get a straight answer — and for the person in the next meeting to get the same one.
- The question — asked in the words people actually use, not in the columns the database happens to have.
- The definition — “margin” means one thing, written down once as code, so the answer cannot depend on who ran it.
- The permission check — a branch manager opening the same report sees their branch and nobody else’s.
- The answer — one chart, with the outlier already marked rather than left to be spotted.
- The follow-through — that same number in the board pack, the manager dashboard and a 7am message, without anyone re-typing it.
“Which branch is losing margin, and since when?”
Defined once
margin = (net_sales − cost_of_sales) / net_sales
Agreed with finance. Every report below reads this one line.
Opened by a regional head — five branches. A branch manager opening the
same report sees one.
Branch
Sales volumeMargin
Branch 1
18.9%
Branch 2
19.4%
Branch 3
11.2%
Branch 4
20.1%
Branch 5
18.2%
Branch 3 sells the most and keeps the least. Margin fell from 19% to 11% over two months while volume rose — the discount policy, not the cost of stock.
Monday board packthe same figure, already written in
Branch manager dashboardtheir own branch only
WhatsApp digest, 7amthree numbers, no login
Sample branches and sample figures — deliberately numbered, so nothing here reads as a real book of business. The definition line, the permission check and the scheduled sends are the real mechanics.
Capabilities
What exactly we do
Six things we build so a report gets opened on a Monday instead of ignored.
Semantic & metrics layer
Revenue defined once, in code. Every report reads from it, so meetings stop opening with an argument.
Self-serve enablement
The people who ask the questions able to answer most of them, with training and a governed model.
Governed access
Row and column-level permissions matching your org chart, enforced at the data layer rather than by convention.
Embedded analytics
Reporting inside your own product or customer portal, multi-tenant and rate-limited.
Open-source BI
Metabase, Superset or Grafana, self-hosted, with no per-seat licence that grows every time you hire.
How we deliver
What you get, every time
A dashboard earns its place by being opened. Everything below exists to make that happen.
Talk to an engineer- Built around a weekly decision, not a data dump
- One metric definition, shared by every report
- Loads in under two seconds on the real dataset
- Access rules that match your org chart
- Your team trained to build the next one themselves
- Self-hosted and yours, if you would rather not rent it
Next door
Is BI what you need?
Reporting only works if the data underneath is sound and somebody has already asked the hard questions of it.
Our process
How we work
From a spreadsheet somebody maintains by hand to reporting the business runs on.
Proposal sign-up
Scope, timeline and a fixed number in writing — agreed and signed before any build starts, so nothing moves later without a conversation.
Strategy & planning
We sit with the people who will use it and find the decision it has to support.
Design & development
The metrics defined and agreed with finance before a single chart is drawn.
Testing & deployment
Dashboards built and reviewed with real users, iterated while changing them is still cheap.
Support & maintenance
Training, documentation and the ability to build the next one without us.
Before you ask
Questions we get asked
Power BI or an open-source tool?
Power BI if your organisation already lives in Microsoft 365 and per-seat licensing is not a constraint. Metabase, Superset or Grafana if you want to self-host, avoid a licence that grows with headcount, or embed reporting inside your own product. We will recommend against our own build if a licence genuinely works out cheaper.
Why do we need a metrics layer?
Because otherwise 'revenue' is defined slightly differently in four dashboards and every meeting starts with an argument about whose number is right. A metrics layer defines it once, in code, and every report reads from it.
Ready to talk about Business Intelligence?
Thirty minutes with an engineer, not a salesperson. You will get a straight answer on scope, timeline and cost.