Every department keeps its own Excel
We write definitions down once. Sales counts orders, accounts counts invoices, logistics counts shipments. All of them are right, yet their figures never meet.
Area 14 · Data, AI and consulting
Data and analytics
Think of a farm machinery wholesaler selling through dealers in Slovenia and Croatia. When a new month begins, the sales manager exports turnover from the business software, adds webshop sales, subtracts returns by hand and emails the spreadsheet to the board. Accounts then bring a different total for that period, since they go by invoice date and sales go by order date. The first half hour of the meeting is spent establishing who is right. Repeated monthly, this ritual swallows several days of work and splits the company into two camps, each holding its own numbers. We aim for a single version: tidy master data, all sources gathered together, reports that update overnight and indicators defined in writing. All of it is done from a distance, through secured access and video meetings.
3
services in this area
1 revenue
figure shared by sales, accounts and the board
€50
per hour plus VAT, or a fixed price after an initial review
EU
data centres for your data, in an account owned by your company
How the three services fit together
You can order any one of them separately, yet the sequence has a logic. Reports built over untidy records simply present old mistakes more nicely, so cleaning comes first. Reading documents can run at the same time, because it cuts manual entry whether or not you have a warehouse yet.
Master data management
One entry per customer, supplier and product, carrying the correct legal name, address and VAT identification number. Duplicates are found, merged once you approve and kept out afterwards by rules inside your business software. E-invoices returned because of wrong buyer details become a rarity.
Data warehouse
Pantheon, Minimax, Metakocka or Business Central, the webshop, Amazon and eBay, ad spend and timesheets flow into a shared database every night. Reports in Power BI or Metabase read from it and agree with the books to the cent.
Intelligent document processing
Foreign PDF invoices, paper delivery notes, CMRs, customs papers and statements: anything someone is still retyping today. The software reads each document, links item codes to yours and lines it up for a colleague to approve.
Does any of this sound familiar?
We are almost never asked to clean data. We are asked for a dashboard, and once we open the sources we find items from the list below. The more of them you recognise, the less your reporting software is to blame.
The same customer three times
We merge and set rules. “Novak d.o.o.”, “NOVAK DOO” and an entry with no address are three separate customers to every report, each with a credit limit of its own.
The report arrives too late to matter
We automate the data loads. The monthly review lands on the tenth, because it is compiled by the same person who is closing the books and preparing the VAT return at that moment.
E-invoices keep bouncing back
We fix the root cause. The recipient rejects an e-invoice because the name or VAT identification number in your records is wrong. Correcting the invoice helps once; correcting the record helps for good.
Incoming invoices are still retyped
We introduce automatic reading. eSLOG e-invoices arrive neatly structured, but PDFs from abroad and paper delivery notes from smaller suppliers are still copied out line by line.
Nobody knows which channel pays
We work out contribution per channel. The webshop, Amazon and dealers are compared on turnover, not on what remains after commissions, delivery, advertising and returns.
A dashboard with no owner stalls within three months. Reports rarely die because the numbers are wrong; far more often nobody is responsible for opening them and acting on what they show. So for every dashboard we agree with you who uses it, in which meeting and for which decision. An indicator nobody needs for a decision does not make it onto the screen. Before the first chart is drawn, we check how clean the data is and report the error count without varnish.
How a project unfolds
The first topic is narrow, typically revenue or stock, with five to ten indicators. The next one starts once people open that first report every day. We work over a secured remote connection and talk on video calls.
01
A shared vocabulary
With accounts and sales we pin down what revenue, margin, a return and an “active customer” mean. Skip this agreement and the first report falls apart at its first meeting.
02
Source map
Business software, webshop, marketplaces, ad accounts and unofficial spreadsheets. For each source we record the owner, the access route and any personal data fields that belong in your record of processing activities.
03
Old and new together
For a few weeks the board receives both the manual and the automated report. The manual one is dropped only when the source of each difference is known.
04
Each new topic costs less
New reports reuse the model and the nightly loads already in place, so they take far less work than the first one did.
Frequently asked questions
Not necessarily. Most small and mid-sized firms need someone who knows the business and can read a dashboard, not a programmer. We build and maintain the warehouse and reports, and your colleague learns to add simple views and filters. If you hire an analyst later, they inherit a tidy model and documentation rather than a pile of spreadsheets.
For a company on Microsoft 365 and Entra ID, Power BI tends to be the obvious fit: permissions follow existing accounts and Excel is a click away. Teams that know SQL and want to avoid per-user licences often pick Metabase hosted on a server of their own. Google Workspace users can also consider Looker Studio. All of them read from the same warehouse, and we compare total costs before you choose.
It should not. Data is pulled overnight or in small batches, preferably through the software's interfaces or from a copy of the database, never with heavy queries in the middle of the morning. That is precisely why a separate warehouse exists: reports query it, and the colleague issuing invoices at the time notices nothing.
It sits in an account belonging to your company: at Hetzner, OVHcloud or IONOS, in an Azure or AWS region inside the EU, or on a server you run yourselves. Apply acts as a processor under GDPR, which we formalise in a data processing agreement. Permissions follow roles, so directors have full visibility while a regional salesperson sees just one region. Every access is recorded.
That depends mostly on the state of your master data. If it is roughly in order, the first report comes quickly and most of the time goes on agreeing definitions. If there are thousands of duplicates, the real project is the clean-up. So we give you a timeline once we have seen your data. Pricing is either fixed or €50 per hour plus VAT against an agreed estimate.
No. Excel is a great tool for analysis; the trouble is only the monthly manual job of feeding it. Connected to the warehouse or via Power BI, Excel brings fresh figures, identical for everyone, each time it opens. Everyone keeps their own pivot tables; only the source underneath becomes common.
Related areas
Area 13
Artificial intelligence
Where machine learning helps once your data is reliable
Area 10
E-commerce and integrations
Orders landing in your business software with nobody retyping them
Area 15
IT consulting and audit
Picking indicators that really help when deciding
Solution
GDPR compliance
Personal data in reports kept under control as well
Let's sort out your numbers together
Tell us which reports are still put together by hand and where the figures fail to match. We will look at your sources and say where it makes sense to start.
When we work
Weekdays 9:00-17:00 CET, answers the same or next working day
Weekdays 9:00-17:00 CET, answers the same or next working day
Calls
Video meetings, Teams and Meet both welcome
Video meetings, Teams and Meet both welcome
Enquiry received
An engineer will write back no later than the following working day. Faults that leave your staff unable to work are always looked at first.
- Questions first. If something is missing before we can estimate, we email you or set up a brief video chat.
- A written quote. It spells out the work, a euro figure excluding VAT and a sensible start date. No fine print, no surprise line items down the road.
- You choose. It lands in your mailbox. Read it at your own pace, query anything that seems vague, and only then make up your mind.