In most manufacturing companies the data already exists. The problem is where it lives. Some of it sits in the ERP, some in the machines on the shop floor, some in the sales team’s spreadsheets, and some in accounting’s e-invoice archive. When the owner asks “which product made money this month, which machine stopped the most,” three people spend two days pulling a report together. By the time the answer arrives, the month is already closed.
Business intelligence (BI) closes exactly this gap: it collects data from scattered sources, cleans it, and turns it into dashboards a decision-maker can read at a glance. This guide covers what BI actually does inside an SME, which metrics to start with, and what to watch out for during setup — in plain terms.
What business intelligence actually solves
Business intelligence is the shared name for tools that turn raw data into a business decision. It does three things at once: it brings together data from different systems, makes them speak the same language (matching a customer that appears under two names in two systems), then renders the result as charts and dashboards.
The key difference is here: a classic report tells you the past, a dashboard shows you what is happening now. A monthly sales report says “here’s what we sold last month.” A BI dashboard shows today’s revenue, status against target, the best-selling product and out-of-stock items on one screen, refreshed automatically. The person deciding doesn’t wait for a report; they look at the screen.
This is the natural continuation of collecting production data from machines. If you’ve gathered shop-floor data through sensors and machine-to-machine communication (we covered it in collecting production data from machines), business intelligence is the layer that makes that data meaningful on a dashboard.
Report versus dashboard
For an SME, a few concrete things separate the two:
- Freshness: A report waits for someone to sit down and prepare it; a dashboard feeds itself from the connected system.
- Depth: In a dashboard you can click a number and drill down — from “revenue dropped” to which customer and which product it dropped in.
- Single source: Because everyone looks at the same dashboard, the “your spreadsheet says one thing, mine says another” argument ends.
- Access: A dashboard opens on a phone too; the owner sees the same number when away from the plant.
Among the tools widely used in Türkiye, Microsoft Power BI stands out; the jump from spreadsheet habits is easy and it sits in the same ecosystem as Microsoft 365, so manufacturing SMEs often prefer it. Tool choice is a secondary matter — first you need clarity on which question you’re answering.
Which data comes together
In a manufacturing SME, the typical sources feeding a dashboard are:
- ERP: Orders, stock, purchasing, accounts, cost.
- Shop floor (MES/machines): Machine run and stop times, units produced, scrap, cycle time.
- Sales: Quotes, order conversion, revenue by customer.
- Finance/accounting: Collections, due dates, e-invoice data.
- Energy and meters: Electricity, natural gas, water consumption.
You don’t have to connect all of these on day one. A good start begins with the data behind the most painful problem — usually machine downtime or cash flow.
Choosing the right metrics
The most common place a BI project goes to waste is the urge to “measure everything.” A dashboard with forty metrics becomes wall decoration nobody looks at. Picking a small number of decision-linked metrics gives a better outcome.
A few examples that work on the production side:
- OEE (equipment effectiveness): How much of a machine’s theoretical capacity it actually produces. It rolls downtime, speed loss and scrap into one number.
- Downtime and its cause: Which machine stopped, how long, why.
- Scrap/reject rate: By product or line.
- On-time delivery: Share of orders delivered on schedule.
On the commercial side, gross margin (by product/customer), collection terms and stock turnover are the first to look at. For each metric ask one question: “If this number changes, what will I do?” If there’s no answer, that metric shouldn’t take up space on the dashboard.
Setup: where to start
A healthy BI setup starts not with a big software purchase but with a small, clear question. The order we follow in our digital transition approach is:
- Sharpen the question. Pick one concrete question, such as “which machine loses me the most money?”
- Find and connect the data. Identify the sources that answer it (ERP + line data). Data quality surfaces at this step; missing or inconsistent records get cleaned.
- Build the first dashboard. Make a simple dashboard that answers the one question, and test it with a real user.
- Scale out. As the dashboard proves useful, add new questions and sources.
This is the first link in the measure → software → compliance chain. At İkiz Eksen we build the layer that collects data from the field, moves it into software and reports on it — from one team, turnkey. The Qera track record behind us — more than 550 customers, over 15 sectors, more than 100 ERP deployments and a team of around 35 specialists — lets us build this infrastructure in a scalable way on Microsoft Azure. We work across Türkiye.
If you’re still deciding on the ERP side, our guide on how to choose the right ERP helps you lay a solid base for BI.
Common mistakes
- Starting with dirty data. A number on a dashboard is only as correct as the data beneath it. A wrong stock record shows up wrong on the dashboard too. Data discipline first.
- Choosing the tool before the job. Don’t get into the “Power BI or something else” debate before the business question.
- Showing everything to everyone. The production manager’s dashboard and the finance team’s dashboard shouldn’t be the same; simplify by role.
- Building it and walking away. A dashboard nobody looks at dies. Building the weekly meeting around the dashboard keeps it alive.
Same data, second purpose: the link to the green side
Once you’ve collected production and energy data properly, that data serves not only operational decisions but sustainability reporting too. Electricity, natural gas and production-volume data are also the raw material for carbon footprint and energy-intensity metrics. A well-built BI foundation spares you from redoing the work when an emissions report driven by CBAM or CSRD is later requested.
We detailed this link in AI-supported sustainability reporting. Designing operational data and green reporting together means building a single data backbone instead of two separate projects — a lens worth applying to our solutions.
Frequently Asked Questions
Do I need an ERP first for business intelligence?
Not required, but it helps a lot. If you have an ERP, it’s the richest source for the dashboard. Even without one, a meaningful start is possible with shop-floor data or sales tables. The more scattered the data, the longer the setup takes.
What does it cost for a small business?
It varies. Tool licenses run as a monthly subscription based on user count and are usually in an accessible range; the real effort is on the data-connection and dashboard-design side. A narrow start answering a single question is far more economical than a broad rollout. For a clear range it’s best to discuss your scope; talk to us for current terms.
How long until it becomes useful?
A focused first dashboard usually comes up quickly; data quality drives the real timeline. If records are clean, it’s fast; if scattered, the cleanup stage stretches. That’s why starting narrow and scaling out is healthier than entering a months-long megaproject.
Where does AI fit in?
Business intelligence shows the present and the past; an AI layer adds things like demand forecasting, anomaly detection or asking questions in natural language. But order matters: without clean data and a working dashboard, the AI layer floats. Data backbone first, prediction second.
Does our data stay secure?
Enterprise BI tools come with role-based access and data-security controls. In İkiz Eksen deployments the infrastructure is configured on Microsoft Azure and access rights are defined by role. We handle data-protection compliance and retention as part of the setup.