Business Intelligence for Mid-Sized Businesses: Build or Buy – and Where to Start

Business Intelligence in mid-sized businesses: Make or Buy – and start with?

Short: Business Intelligence (BI) was long the terrain of large companies with dedicated data departments.

Business Intelligence (BI) was long the terrain of large companies with dedicated data departments. That changed.

Modern BI tools are affordable, cloud-based and so user-friendly that professionals can create meaningful dashboards without programming knowledge.

At the same time, medium-sized companies often have significant unused data potentials: sales data in ERP, customer data in CRM, production data in MES – but no structured evaluation.

This article explains when BI really makes sense and how to get started.

When does Business Intelligence make sense?

Short: Short answer: Business Intelligence sounds like big corporation – it is not.

Short answer: Business Intelligence sounds like big corporation – it is not.

As a basis for deciding Business Intelligence in mid-sized businesses: Make or Buy – and what to start? are suitable data analysis & business intelligence and digitalization in mid-sized businesses.

BI makes sense when decisions are made today on the basis of incomplete, outdated or unstructured data. Typical symptoms:

• Reporting by Excel with manual data compilation every week or month
• Several figures from different departments for the same question
• Felection visibility on operating performance indicators in real time (stock, open orders, utilisation)
• Decisions after abdominal feeling because the evaluation is too complex
• Slate response to problems because reports only come at the end of the month

BI makes no sense, however, if the data situation is basically bad (GIGO: Garbage In, Garbage Out), if there is no willingness to decide on data-based approaches, or if the company is so small that Excel is really sufficient.

What Business Intelligence does

BI comprises three core functions:

Reporting: Structured, automated reports – daily, weekly, monthly – without manual compilation. Example: Automatic sales report per region and product every Monday morning in the inbox.

Dashboards: Interactive visualizations of current KPIs in real time or almost real time. Example: Production dashboard with utilisation, reject rate and OEE per shift.

Ad hoc analysis: Users can ask questions themselves and explorate data without IT ticket. Example: sales staff analyses themselves which products in which channel in Q3 have performed above average.

Make-or-Buy decision

Short: **When a finished BI tool (Buy):**A pre-configured BI tool is the right entry for the majority of medium-sized companies.

**When a finished BI tool (Buy):**A pre-configured BI tool is the right entry for the majority of medium-sized companies.

Suppliers such as Power BI (Microsoft), Metabase, Grafana, Tableau or Looker offer different price models and strengths:

| Tool | Strength | Price (approx.) | |---------- **Power BI | Microsoft integration, broad connectors | 10–20 €/user/month | | Metabase | Simple, self-hostable, open source | free up to 500 €/month | | Grafana | Technical dashboards, time series | Open Source (Self-Hosted) | **Tableau | Very powerful, complex | 70–115 €/user/month | **Looker | Enterprise, Google integration | on request |

Power BI is the natural entry point for Microsoft-365 houses. Metabase is especially attractive for technical teams without enterprise budget.

When individual development (Make): If standard tools cannot map the specific data models or integration requirements if special data protection requirements (on-premise, no cloud sharing) apply or if BI is to be integrated into an existing system that customers or partners see. In these cases, individual software development is useful for BI components.

The data base: What really decides

Short: The best BI tool is worthless without clean, consolidated data.

The best BI tool is worthless without clean, consolidated data. Most common hurdles:

Data in different silos: ERP, CRM, commodities management, accounting – each system has its own truth. Before BI becomes productive, it needs system integration or a data warehouse as a central data source.

Quality: Missing values, duplicates, inconsistent names. Data quality projects are unspectacular and complex – but without them BI provides false answers.

No clear ownership: Who is responsible for data quality? Without Data Owner, any data landscape degenerates.

Entry strategy: How to get it right

Short: **Step 1: A question, a data source.

**Step 1: A question, a data source. ** Do not start with the complete company BIO.

Choose a specific, important question: "How does our margin evolve per product group?" – and answer a question well with a data source. Small scope, real added value, quick feedback.

**Step 2: Build data pipeline. ** Automated data transfer from the source database to the BI tool. No manual export. interface development or ETL tools (Airbyte, Fivetran, dbt) connect sources to the BI layer.

**Step 3: Insert specialist users. ** BI fails when it remains IT project.

Controlling, sales management, production must help shape from the outset – which key figures count, which granularity, which filter options. .**Step 4: Establish governance. ** Who cares for the data?

Who checks quality? Which reports are "official" and what are experiments? Without data governance, BI becomes wild growth.

Data Analysis and AI: What is possible today

Short: Modern BI tools integrate AI functions: automatic anomaly detection, trend forecasts, natural-language queries ("To me the turnover of the last three months by region").

Modern BI tools integrate AI functions: automatic anomaly detection, trend forecasts, natural-language queries ("To me the turnover of the last three months by region").

These functions significantly reduce the entry barriers for specialist users.

For deeper analysis – predictive analytics, demand forecasts, quality forecasts – data analysis & business intelligence is the next step as dedicated performance.

Conclusion

Short: Business Intelligence in mid-sized businesses is not a luxury and not a big project when it comes to real.

Business Intelligence in mid-sized businesses is not a luxury and not a big project when it comes to real.

With a clearly defined starting point, a clean data source and a professional user as a champion, first real added value can be displayed in 4-8 weeks.

Fall knits – bad data quality, lack of governance, broad scope – are known and avoidable.

Frequently Asked Questions (FAQ)

Do we need a Data Engineer or Data Scientist?

For entry no. With Power BI or Metabase and a clean data source, professionals can start with number affinity.

Data Engineers become relevant when data from many sources need to be consolidated or transformed.

What does a BI project cost?

Pure tool license: 100–2,000 €/month depending on the number of users and tool. Implementation (data pipeline, first dashboard): €10,000 per complexity. Current data maintenance: internally or as a managed service.

Can BI replace our Excel evaluations?

Yes – but not overnight. Acceptance increases when BI dashboards provide more comfort than Excel. This requires that the correct key figures are available in the correct granularity.

How does BI relate to our ERP?

ERP and BI are complementary. The ERP is the transactional system (data is collected and processed). BI is the analytical system (data are evaluated and visualized).

The connection runs through interfaces or direct database connection.

Business Intelligence in mid-sized businesses: Make or Buy – and start with?

can be successfully implemented when technology, organization and measurability fit together – instead of insulated tool rollouts without process reference. Use the overview in this article as a basis for discussion on priorities, risks and the first loadable pilot. .Intensify matching topics in category overview Blog category and check operational support via software development, IT consulting. Groenewold IT accompanies analysis, implementation and operation – from the first classification to scalable releases.

Additional notes

Measurability and quality assurance

Define Erfolg on measurable criteria – for example reduced processing time, lower escalations or higher conversion – and not only managed via “Go-live”.

For business, a slim set of automated tests is worth on the most important user journeys plus targeted manual exploratory tests before releases.

Quality is also created by code reviews, architecture decision logs (ADR) and clear handovers to the operation: runbooks, escalation paths and documented border cases.

Knowledge remains in the company – regardless of individual persons or service providers.

Technical sources and further links

The following independent references complement the classification on the topics of this Article:

• Bitkom – Digital Economy Association
• BSI – Federal Office for Information Security
• European Commission – Digital Strategy
• MDN Web Docs (Mozilla)
• W3C – World Wide Web Consortium

"Privacy by Design is not a subsequent checkbox, but an architectural question – especially for personal master data."

— *Björn Groenewold, Managing Director, Groenewold IT Solutions *

Conclusion and next steps