# Data analytics for SMEs: from documents to decisions

> Why data analytics isn’t a luxury for a small business: margins, customers, stock and cash flow, straight from the documents you already issue.

_2026-09-24 · Kronn (https://kronn.gr)_

When you hear "data analytics", you probably picture large companies, teams of analysts and expensive business intelligence tools. For a small or medium-sized business, data analytics is something much simpler and much more immediate: answering, regularly and with numbers instead of gut feeling, four questions that already keep you up at night.

The good news is that you already have the data. Every invoice you issue, every supplier invoice, every payment received and every delivery note is a record that says something about your business. The challenge isn't collecting more data. It's making the data you have speak.

![From documents to decisions: your documents, one system with the same definitions, four questions, decision and action](/insights/img/analysi-dedomenon-gia-mikromesaies-epixeiriseis-1-en.png)

## The data already exists — it's just scattered

In a typical small business, sales live in the invoicing software, purchases in folders or in email, payments in online banking, stock in a spreadsheet one person updates, and the "full picture" with the accountant, a month or two late.

Each source is right on its own. Together they don't agree, because each one counts in its own way. Is revenue net or gross? Are credit notes deducted? Does a cancelled sale count? Are "Papadopoulos SA" and "PAPADOPOULOS S.A." the same customer?

So the first step isn't a charting tool. It's **one source of truth with the same definitions**: sales, purchases, payments and stock in the same system, so every number is always calculated the same way. If you mostly work in spreadsheets today, our article on [the spreadsheet ceiling](/en/insights/esoterika-ergaleia-to-tavani-tou-excel) may interest you too.

## 1. Where is my margin?

Revenue is the most visible number and the least useful on its own. Two products with the same sales can contribute completely differently to profit.

A simple example with made-up numbers: product A sells €10,000 a month at a purchase cost of €8,500; product B also sells €10,000, at a cost of €6,000. In a sales report they look identical. In reality B brings in almost three times the margin. If your sales rep gives the same discount on both, the discount on A eats a large part of its profit.

To see this you need two things: **a purchase price per item**, updated from your suppliers' invoices, and **the actual selling price**, after discounts. With those, a "profit by item" report answers questions like:

- Which items sell a lot but earn little?
- Did a supplier raise their cost without us passing it on?
- Which discounts aren't worth it?

## 2. Which customers matter?

Most businesses know who their "big" customers are. Fewer know **how they're changing**. A customer who buys a little less every month doesn't show up in a monthly glance; they show up when you look at their trend month by month.

Three things are worth checking regularly:

- **Concentration**: what share of your revenue your five biggest customers make up. The higher it is, the bigger the risk if one leaves.
- **Trend**: which customers buy less than last year, and which items they stopped taking.
- **Payment behaviour**: who pays on time and who is always late. A big customer who pays at 120 days costs more than it looks.

The best customer isn't necessarily the biggest. It's the one who buys steadily, at a good margin, and pays.

## 3. What is sitting in stock?

Stock is money on a shelf. Every item that doesn't move ties up capital that could be in the bank or in something that sells.

Useful questions:

- Which items sold nothing in the last three months?
- What is stock worth at cost, and how is that changing?
- What's coming from suppliers, and what have we already promised customers?
- Where are we at risk of running out, so we can reorder in time?

Answering these takes something many small businesses skip: goods entering stock **from purchase documents**, at their cost, not from memory at the end of the month. That's where [AI scanning of expense documents](/en/insights/ai-sarosi-parastatikon-exodon) helps a lot.

## 4. When does the cash arrive?

Sales aren't cash. A business can have its best invoicing month ever and struggle to pay salaries, because the money will arrive in two months.

For cash flow you need two lists side by side:

- **Receivables**: what customers owe you, and how late it is (receivables aging).
- **Payables**: what you owe suppliers, and when it falls due.

When incoming payments are matched to invoices automatically through a bank connection, the first list is right every morning, not once a month. And then the call to the customer who's running late happens on day five, not in month three.

![The four questions your data answers: margin, customers, stock, cash flow](/insights/img/analysi-dedomenon-gia-mikromesaies-epixeiriseis-2-en.png)

## From report to decision

A report that doesn't lead to action is just another file. The difference between businesses that benefit from their data and those that merely collect it is almost always **rhythm**:

- **A fixed slot**: fifteen minutes every Monday with the same four numbers. The comparison with last week says more than any single number.
- **One owner per number**: someone watches collections, someone else watches stock.
- **Reports that arrive on their own**: if you have to open the software and build the report each time, one week you won't. If it lands in your inbox, you'll read it.

Examples of the decisions that follow: you change the price of an item that sells at a loss, you call a customer whose purchases are shrinking, you stop reordering something that just sits there, you ask a chronically late payer for an advance. For how to build a screen people actually look at, see our article on [dashboards that change decisions](/en/insights/dashboards-apo-dedomena-se-apofaseis).

## Three mistakes that undo the analysis

**1. Inconsistent data entry.** The same customer under three names, the same item under two codes, "miscellaneous" sales with no item. Every such record makes reports less reliable. The fix is customers created from their VAT number and items with stable codes from day one.

**2. No purchase prices, no margins.** If purchase documents aren't entered with item lines, there's no cost per item — and "where is my margin" stays unanswered.

**3. Too many indicators.** A dashboard with thirty charts impresses on day one and is forgotten on day two. Start with four numbers and add one only when it would change a decision.

## How Kronn Smart does it

In [Kronn Smart](/en/smart) sales, purchases, payments and stock live in the same system, so reports always count with the same definitions — with credit notes deducted. For example:

- **Sales** by customer, by item, by category, and the monthly trend.
- **Profit by item**, from purchase prices updated from your suppliers' invoices.
- **Receivables and payables aging**, with the bank connection matching incoming payments.
- **Customer and supplier analysis** on each one's page: sales or purchases and balance, month by month.
- **Export** to Excel, CSV or PDF, and **reports emailed automatically**, to you or your accountant.

Purchases can come in from myDATA or be read from the PDFs and photos of your documents with [AI scanning](/en/smart/ai-scanning), so cost prices exist without typing. See the [plans and prices](/en/smart/pricing).

If your data is spread across several systems and you need something more specific — an internal tool or a dashboard that joins sources — we [build that to order](/en/services/internal-tools). See [our work](/en/work) or [drop us a line](/en/contact).

---
Page: https://kronn.gr/en/insights/analysi-dedomenon-gia-mikromesaies-epixeiriseis