Skip to main content
ServicesData
Mid-market company (confidential)

25 people flying blind on manual reporting: time for dashboards that actually drive decisions

Before

½ day

lost per person per week on manual reporting and data compilation

After

½ day

saved per person per week

½ day

saved per person per week

25

employees impacted

4

business processes modeled with the teams

In this mid-market company, every team was flying blind. Activity data lived in scattered Excel files, data models were undocumented, and reports were built by hand every week. 25 employees spent a significant share of their time compiling and formatting data, without the resulting dashboards actually supporting decisions. Doing pragmatic business intelligence in a mid-market company means running into realities that tools alone do not solve: data of uneven quality, legacy models that nobody fully understands, and business teams that have neither the time nor the method to formalize their needs.

SITUATION BEFORE

Operational baseline
Activity data scattered across Excel files, team by team. Reports built by hand every week. The existing dashboards did not serve decisions.
Teams affected
25 employees across all teams (operations, financial control, HR, regulatory affairs).
Hidden costs
Time lost every week consolidating figures that nobody disputed but that everyone recalculated on their own, at the expense of the core business.
Why it was not solved earlier.
The difficulty of doing pragmatic BI in a mid-market company: undocumented data models, uneven data quality, business teams with neither the time nor the method to formalize their needs. Tools alone do not solve these problems.

THE STORY

In this mid-market company, the data existed; it was simply unusable. Each team had its own files, its own exports, its own weekly reporting rituals. The data models were undocumented. And the existing dashboards did not serve decisions; they served to produce reports. That is the reality of business intelligence in a mid-market company: the tools exist, but without a method to understand the data and connect it to decisions, they only produce noise.

Start from business processes, not from data. Nobori Partners worked closely with the 4 business teams (operations, financial control, HR, regulatory affairs) to model their processes and understand how they work day to day. Which decisions do they make? Which information do they actually need to make them? This hands-on work identified the KPIs that matter and tied them to concrete decisions, eliminating along the way the metrics that only fed reports nobody read.

A deep dive into undocumented data. Before building anything, Nobori dove deep into the existing data and models, largely undocumented, inherited from past choices that nobody fully understood anymore. Every source was combed through to analyze data quality: inconsistencies, duplicates, poorly filled fields, implicit business rules. In a mid-market company, doing pragmatic BI first means accepting that the data is not clean, and building a reliable foundation anyway.

Dashboards that are actually used. The data, cleaned and structured with dbt on Azure, now feeds a single source of truth. The Power BI dashboards were built from the modeled processes and the KPIs identified with the teams. Each view answers a specific decision, not a generic reporting need. All 25 employees now have a management tool they use every day.

The result is simple to measure: half a day saved per person per week across the 25 employees. Time given back to the core business, because the dashboards were built to serve decisions, not to fill slide decks.

OUR APPROACH

5 steps to get there

1

Working with the teams and modeling business processes

Working closely with the 4 business teams to understand their day-to-day work, model their processes and identify decision points. The goal was not to collect data but to understand how each team works, which decisions it makes, and which information it actually needs to make them.

2

Identifying KPIs and the decisions they drive

For each modeled process, identification of the key performance indicators and the decisions they must inform. This work separated the metrics that actually serve management from those that only feed reports nobody reads. Every KPI is tied to a concrete decision.

3

Deep dive into the data and quality analysis

An in-depth dive into the existing data and models, largely undocumented. Nobori combed through every source to analyze data quality: inconsistencies, duplicates, poorly filled fields, implicit business rules. In a mid-market company, that is the reality of pragmatic BI: before building dashboards, you have to understand what the data actually says and what it does not.

Azuredbt
4

Data foundation and transformation pipelines

Centralization of the data scattered across business tools, cleaning and structuring with dbt on Azure. The transformation pipelines embed the quality rules identified during the deep dive and feed a single source of truth, reliable and up to date, on which the dashboards rely.

Azuredbt
5

Building the management dashboards

Design and deployment of Power BI dashboards built from the business processes and KPIs identified with the teams. Each dashboard answers specific decisions, not a generic reporting need. All 25 employees have views that serve them every day.

Power BIAzuredbt

RESULTS

The measurable impact

½ daysaved per person per week
25employees impacted
4business processes modeled with the teams

The management dashboards built with the teams freed up half a day per person per week across the 25 employees. The 4 modeled business processes now have indicators tied to concrete decisions. Manual reporting has been eliminated in favor of real-time views on a single source of truth.

Book a discovery call