Tāmaki Makaurau Auckland
Stop rebuilding the same spreadsheet every month.
DataHorizon replaces manual Excel reporting with governed Power BI and Microsoft Fabric platforms, then engineers the pipelines, predictive models and applications that put those numbers to work.
A senior Auckland team across BI, data engineering, governance and AI. It's the blend most organisations are trying, and failing, to hire one person at a time.
Prefer to look before you talk? The health check scores your reporting in the browser and names where you'd start — no email needed.
Connect the dots
Signal, not noise ↗
Positioning
Most consultancies hand over a report. We hand over the system the decision runs on.
Data engineering, semantic modelling, AI and the application layer are one team here — so there's no handoff gap between the people who model the numbers and the people who build the interface your team actually uses.
Where our credibility comes from
DataHorizon is a small firm. The delivery record behind it is not.
We won't show you a logo wall or quotes from unnamed executives. What we will do is walk you through the platforms, semantic models and applications our principals have shipped inside organisations, and put you in touch with named, contactable referees during scoping.
- 18+ years
- Combined principal experience
- 6 industries
- Sectors delivered into
- 4–6 weeks
- To first production release
- 2 principals
- On every engagement
Platforms, BI and AI delivery since 2016.
In-house and consulting roles across both principals' careers.
A production-grade slice in your hands before scale-out.
The consultants who design the system are the ones who build it.
Sectors we work in
Logistics & ports · Financial services & insurance · Agritech & primary export · Infrastructure & utilities · Healthcare · Professional services
Want the detail? Ask on the scoping call and we'll take you through the architecture we built, the decisions it changed, and the people who can confirm it. Scoping conversations are free and no NDA is needed to talk.
TALK TO A PRINCIPALWho you actually work with
MEET THE TEAM →No account layer between you and the build. The principals who scope the work are the ones who write the DAX, shape the Fabric pipelines and review the code.
Mohit Saini
Principal Consultant — Data, AI & BI Systems
End-to-end data systems that turn analytics and AI ambition into governed, operable infrastructure
- Delivered data platforms across port logistics, insurance and agritech
- Hands-on every build — DAX, Microsoft Fabric pipelines and TypeScript
- Takes engagements from diagnostic to a production-grade slice
Teo McArthur
Principal Consultant — BI Strategy, Finance & Transformation
Turning financial and operational complexity into one governed source of truth for leadership decisions
- Led organisation-wide Power BI rollouts
- Replaced fragile Excel workflows with SQL-backed reporting
- CFA Level I candidate — valuation and modelling grounding
Problems we solve
If one of these is being said out loud in your business, start there.
“Our reporting is a spreadsheet someone rebuilds every month.”
Excel is still the incumbent BI tool in most Auckland finance and ops teams. We move that estate onto a governed Power BI model with automated refresh, and the monthly rebuild goes away.
BI & SEMANTIC MODELING →“Two reports, two different numbers, same question.”
Metric sprawl is a definition problem, not a dashboard problem. We build one certified semantic layer in Power BI and Fabric so every report resolves to the same measure.
BI & SEMANTIC MODELING →“Nobody trusts the numbers behind the dashboard.”
Quality failures upstream of reporting are the most common problems we're called into. We fix ingestion, reconciliation, lineage and access control so the numbers survive scrutiny.
DATA ENGINEERING & GOVERNANCE →“We need to move off legacy — or off spreadsheets — onto Fabric.”
Assessment first, then a phased migration on Azure with Fabric Lakehouse, Snowflake or Databricks and dbt/ADF transforms. No big-bang cutover, no reporting blackout.
DATA ENGINEERING & GOVERNANCE →“The board wants AI but our reporting is broken.”
You cannot model on records you can't reconcile. We run an AI-readiness audit on your reporting estate, then ship the first predictive model where the underlying numbers are already trustworthy.
AI & PREDICTIVE MODELING →“We have the insight but no way to act on it.”
Dashboards end at the read. We build the portal, approval flow or field-capture app that writes the decision back into your systems. That last layer is the one BI-only firms leave out.
INTELLIGENT WEB APPLICATIONS →Who we work with
The same problem shows up at fifteen people and at fifteen hundred. The entry point differs.
10–150 staff
Growing NZ businesses
Your reporting lives in spreadsheets and one person rebuilds it every month. We turn that into an automated Power BI report set your team owns, without a platform programme to run.
Typically $12k–$20k + GST for the first release
Spreadsheet-to-Power-BI reporting rescue200+ staff
Enterprise data teams
You already have BI authors, Azure and a backlog. We add senior capacity where it hurts: certified metric definitions, governed pipelines, and Fabric migration without a reporting blackout.
Assessments from $9k + GST; build phases typically $25k–$60k + GST
Metric governance & semantic layerCouncils, agencies, CCOs
Public sector
Reporting that stands up to scrutiny: auditable lineage, documented definitions, accessibility designed in, and regional hosting options where your environment requires them.
Fixed-scope audits from $9k + GST; phased delivery quoted per stage
Fixed-scope reporting & AI-readiness auditPublic sector
Designed around the obligations your teams need to manage.
We deliver for councils, agencies and council-controlled organisations under the constraints that matter to them: personal information considered against Privacy Act 2020 expectations, regional hosting options where policy requires them, security controls mapped to NZISM guidance, and accessibility treated as a design target rather than a retrofit. Engagements can be run as fixed-scope stages so the spend fits an existing delegation, with procurement material prepared to support GETS responses where needed.
- Privacy Act 2020 considerations mapped into the information model and operating process
- New Zealand or Australian Azure regional deployment options, subject to your tenancy and policy
- WCAG 2.1 AA used as the accessibility target for dashboards and applications
- GETS-style response material prepared for procurement conversations when needed
Scoping conversations are free, and there's no NDA needed to talk. Book 30 minutes.
What we build
Intelligent Web Applications
Portals your operators actually open.
Internal portals and customer applications built on React, TypeScript and Vite, styled with Tailwind CSS and backed by Supabase. That includes installable Progressive Web Apps when a team needs app-like access on a phone without going through an app store.
- React
- TypeScript
- Vite
- Tailwind CSS
- Supabase
- PWA
AI & Predictive Modeling
AI in the engine room, not the pitch deck.
You cannot model on data you can't reconcile, so we treat data quality as the first AI deliverable. Then we apply AI where it compounds: entity resolution, predictive analytics on operational history, and intelligent process automation wired into the tools your teams already use.
- Feature stores
- Predictive models
- LLM automation
- Evaluation harnesses
Advanced BI & Semantic Modeling
One definition of every number.
Most Auckland reporting is still manual Excel. We move it onto governed Power BI semantic models, built with careful DAX and Power Query and orchestrated through Microsoft Fabric, so refresh runs on its own and every report resolves to the same certified measure.
- Power BI
- Microsoft Fabric
- DAX
- Power Query
- Semantic layers
- Excel migration
Data Engineering & Governance
Pipelines that hold under load.
Untrusted data is the reason most dashboards get ignored. Our data engineering and data integration services fix it upstream: ingestion and orchestration across Microsoft Fabric Lakehouse, Snowflake or Databricks on Azure, with dbt and Data Factory transforms, reconciliation tests, lineage and access control built in from the first commit.
- Microsoft Fabric
- Azure
- dbt
- Snowflake / Databricks
- Lineage
- Data contracts
See it working
Sample dashboards, built to show the craft.
Nothing from your systems — these are real reports we built on representative sample data, so you can judge the design, modelling and governance before we ever touch yours.
Sample dataView full →Logistics & freight
Freight operations control tower
On-time-in-full, cost per shipment and fleet use across a national network, with a live map of where the network is losing time and a ranked view of the delay causes behind it. Built to show how we shape an operations report around the decisions a control room makes.
- Operations analytics
- Semantic model
- Root-cause design
DEMONSTRATION BUILD · SAMPLE DATA
Sample dataView full →Media & broadcasting
Commercial performance: audience & yield
A two-view executive report tracking audience growth against falling value per viewer — LTV:CAC, ARPU, retention cohorts and the replacement gap between linear and digital revenue — ending in three costed moves. Built to show how we turn a tangle of commercial metrics into a decision.
- Executive reporting
- Metric modelling
- Narrative design
DEMONSTRATION BUILD · SAMPLE DATA
More demos in build: Financial services · Healthcare.
Method
- 01
Diagnose the decision
We start at the decision being made badly or late, then trace back to the systems feeding it.
- 02
Engineer the pipeline
Ingestion, modelling and governance built as code, with lineage and tests from the first commit.
- 03
Ship the interface
A certified model, a predictive score or a purpose-built portal — whatever puts the answer where the work happens.
FAQ
Questions we get before the first call.
How engagements start
ALL ENGAGEMENTS →The front door
Spreadsheet-to-Power-BI reporting rescue
Manual Excel reporting that won't scale, no automation, and a monthly close that eats a week of senior time.
4–6 WEEKS TO FIRST PRODUCTION RELEASE
The credibility fix
Metric governance & single-source semantic layer
Every team has its own definition of revenue, margin or utilisation, so leadership meetings argue about data instead of decisions.
6–8 WEEKS PER BUSINESS DOMAIN
The higher-value play
Fabric & modern-stack migration
A legacy BI platform, or a sprawling spreadsheet-and-SQL estate, that can't carry the next three years of reporting or AI.
ASSESSMENT IN 2–3 WEEKS, PHASED MIGRATION FROM THERE
The blended senior capability
Fractional data team — retainer pod
You are trying to hire one person who can do BI, engineering, governance, AI and build the app. That person is not on the market.
MONTHLY RETAINER, THREE-MONTH MINIMUM
The fixed-scope opener
AI-readiness reporting audit
There is board pressure to do something with AI, but the reporting layer underneath it can't be reconciled.
2 WEEKS, FIXED SCOPE
Where the insight gets acted on
Decision application build
The model or dashboard exists, but the decision still happens in email, on paper, or in a spreadsheet nobody governs.
4–6 WEEKS TO A USABLE ALPHA
Insights
Notes on the work we do.
How we think about governance, pipelines and the last mile of BI — written for the people who have to live with the system afterwards.
3 OCT 2026 · 8 MIN READ
Audience up 6%, value per viewer down 29%: measuring yield in a shifting media market
Audience up 6%, value per viewer down 29%. How a media dashboard tracks yield, LTV:CAC and the digital replacement ratio, and three moves to fix the gap.
- Media
- Commercial analytics
- LTV:CAC
READ →
3 OCT 2026 · 7 MIN READ
Where the network is losing time: the freight KPIs NZ operations teams should track
OTIF, cost per shipment, dwell time and fleet utilisation: see how a freight dashboard shows where an NZ network loses time and money, and what to do.
- Logistics
- Operations analytics
- KPIs
READ →
24 SEPT 2026 · 7 MIN READ
Why your Power BI numbers disagree — and the governance that fixes it
The same metric, three different answers across three reports. The cause is almost never the data; it's the missing semantic layer. Here's how we build one.
- Power BI
- Governance
- Semantic model
READ →
23 SEPT 2026 · 9 MIN READ
Architecting a Microsoft Fabric medallion lakehouse that survives audit
Bronze, silver, gold is the easy part. Lineage, data-quality gates and row-level security are what keep it standing when the auditor arrives.
- Microsoft Fabric
- Data engineering
- Lineage
READ →
22 SEPT 2026 · 6 MIN READ
The last mile of BI: writing the decision back into your systems
A dashboard ends at the read. The value is in the portal, approval flow or field app that captures the decision and acts on it — the layer BI-only firms leave out.
- Applications
- AI
- Operations
READ →
Low-risk by design
You're weighing up a small firm. Here's how we take the risk off the table.
- Free scoping, no NDA
- The first conversation costs nothing and needs no paperwork to begin.
- A fixed-scope diagnostic first
- We open with a defined, costed diagnostic — not an open-ended retainer.
- Production slice in 4–6 weeks
- Something real in your hands early, before any scale-out spend.
- No lock-in
- We document and hand over the system. It's yours to run — with us or without us.
Tell us the decision you can't make fast enough.
Thirty minutes with the engineers who would run the programme. No pre-sales layer.
Scoping conversations are free · No NDA needed to talk