KPI definitions
A written definition for every metric on the dashboard, with the owner who agreed it. No two reports disagreeing about “revenue”.
Analytics and business intelligence
We build the pipeline, the metric definitions and the dashboards together — so the answer to “how are we doing?” is the same in every meeting.
Written scope and a fixed estimate within two business days. NDA on request.
Activation
Definition agreed
event-based, not page views
Retention
Weekly cohort
by signup month
Revenue
Single source
billing, not the spreadsheet
Freshness
Scheduled
with failure alerts
Illustrative layout — your metrics, your definitions.
The build
Dashboards are the last step, not the first. Building them earlier just makes unreliable numbers look official.
Operational databases, billing, CRM, ad platforms, spreadsheets.
Definitions agreed once: what “active user” or “churned” actually means.
A modelled store that queries fast and keeps history.
One page per audience, with the questions they actually ask.
Deliverables
A written definition for every metric on the dashboard, with the owner who agreed it. No two reports disagreeing about “revenue”.
One screen for the handful of numbers that drive decisions, each traceable to its source query.
Daily and weekly views for the teams who act on the data, delivered where they already work.
A modelled layer business users can filter without writing SQL and without breaking the numbers.
Threshold and anomaly alerts for the metrics where a quiet day is a warning sign.
Documentation, a walkthrough recording, and a session with the people who will own it.
Reporting vs analytics
| Consideration | Manual reporting | Modelled analytics |
|---|---|---|
| Main question | What happened last month? | Why did it happen, and what next? |
| Refresh | Monthly, by hand | Scheduled, automated |
| Trust | Depends who built the sheet | One definition, versioned |
| Effort to answer a new question | A new spreadsheet | A filter or a new view |
| Access | Whoever has the file | Role-based, audited |
Sources
Process
We list the decisions the dashboard has to support, then work backwards. Charts without a decision attached do not get built.
Where each number comes from, what is missing, and where two systems disagree. This is usually where the real work is.
Definitions, joins and history stored in a warehouse layer so every report uses the same arithmetic.
Dashboards per audience, with filters that cannot produce a wrong number.
Documentation, alerting and a training session, so the team can add a metric without calling us.
FAQ
Yes. Looker, Power BI, Metabase, Tableau and Superset are all fine. The tool is rarely the problem — the metric definitions and the data model behind it usually are.
A focused first dashboard on clean sources is typically three to five weeks. If source systems need work first, we will say so in discovery rather than promising a date we cannot hold.
That is the normal starting condition. We reconcile the definitions, document which system is authoritative for each field, and build the model from that decision.
Yes, once the modelled layer exists. Self-serve on top of raw tables is how organisations end up with five different versions of the same number.
We hand over with documentation and can stay on a light retainer for new metrics, source changes and pipeline failures. It is optional, not a condition of the build.
Request a quote
List the metrics, the systems they come from and who needs to see them. We will scope the pipeline, the definitions and the first dashboard.
No obligation. We will tell you if we are not the right team for the job.
Tell us the goal, platforms and must-have features.
We reply within two business days with scope, approach and a fixed estimate.
Sign an NDA first if you prefer — nothing about your idea is public.