Service · Data and analytics

Data warehouse

Picture an outdoor equipment retailer in Salzburg selling through Shopware, Amazon and Kaufland.at. Stock and orders are handled in Business Central, while the bookkeeping is done by an outside tax firm in BMD. Every Monday someone spends three hours assembling a spreadsheet for the managing director, and every week it disagrees with the accountants, because returns, vouchers and shipping charges are treated differently in each system. A data warehouse puts an end to that ritual. It replaces none of the programs. Instead it pulls data from every source, converts it into a common shape and preserves the history even if a system is swapped out later. Reports sit on top of it, so purchasing, marketing and management are finally looking at the same numbers.

Send an enquiry Inclusions overview

Single source

for every key figure

EU hosting

in accounts your company owns

Nightly load

with checks and alerts

Reconciled

with the books to the cent

Everything this covers

Smaller companies can skip the enterprise blueprint. What they require is robust connectors, clear definitions and numbers your tax adviser will not dispute.

Settle the details with an engineer

Definitions before tables

What does revenue mean here: gross, net, after returns, with or without shipping? Who counts as an active customer? We write the definitions down and have them approved by management and accounts.

Connectors to your sources

Business Central, Odoo or exports from BMD and RZL; Shopware, WooCommerce or Shopify through their APIs; settlement reports from Amazon Seller Central and Kaufland.at; Google Ads and Meta; spreadsheets held in SharePoint.

An Austrian calendar

Public holidays, school breaks by federal state, promotional periods such as Black Friday or the Advent shopping Saturdays, and the 20, 13 and 10 % VAT rates kept apart. That way you compare weeks that genuinely are comparable.

Where it is hosted

PostgreSQL at Exoscale in Vienna, at Hetzner or IONOS, or a managed database on Azure, AWS or Google Cloud in an EU region, always under accounts that belong to your business.

Tested loads

Transformations are written in dbt and covered by automated tests. Whenever a job breaks or a total lands outside its usual band, someone is notified and each dashboard displays how fresh its underlying data really is.

Minimal personal data

Reporting seldom needs to know who a buyer is by name or email. Anything unnecessary is pseudonymised, and access is granted by role.

Dashboards for the people using them

Power BI or Metabase with one view per department, readable on a phone and protected by your Microsoft 365 or Google Workspace sign-in.

Our working method

We start with a single source and a single report. Further ones follow only after the first has become a fixture of your weekly management meeting.

Collecting the questions

A two-hour video workshop: which ten questions should the warehouse answer, and which metrics sit behind them?

First report from one source

Usually the ERP. It is reconciled with accounts before anything else is connected.

Growing source by source

Web shop, marketplaces and advertising accounts join one at a time, every one of them tested and balanced separately.

Handover or ongoing care

We coach one or two colleagues so they can run things, or we keep operating the pipelines ourselves if that suits you better.

Amazon does not pay out revenue; it pays out a balance. Each disbursement mixes sales, fees, refunds, advertising charges and sometimes reserves held back for future returns. Load only the payouts and the marketplace seems to earn next to nothing. Load only the orders and it looks wildly profitable. We break the settlement reports down into their components, so that margin per channel can be compared on equal terms.

Frequently asked questions

For a single system and a handful of questions, it is fine for a time. Once the web shop, marketplaces and advertising come in, every report rebuilds the same joins in its own way and the figures drift apart. The warehouse does that work once, centrally and with tests.

The designated person receives an alert, the job runs again, and each report states the date up to which its numbers are current. Nobody ends up acting on figures from the day before yesterday while believing they are fresh.

The warehouse is not a bookkeeping system; it only reads. Each month we reconcile revenue with the totals in BMD or RZL and document every difference, such as credit notes dated on the first of the month or deposits not yet matched to a delivery.

Power BI licensing is per person and varies by tier; Metabase also comes in a free open-source edition you can run yourself. Infrastructure costs scale with data volume and tend to stay small at SME size. Our quote separates fixed and variable costs, and our own work is quoted on the basis of €110 per hour plus VAT.

Certainly. All code and documentation sit in a repository owned by your business, and we coach the colleague who inherits it. Should you prefer not to babysit the nightly jobs, that can stay with us indefinitely.

Related work in the same area

Open the full area page

Service Master data management A single clean entry for each client, vendor and product, protected by rules that stop the drift Service Intelligent document processing Delivery notes, foreign invoices and freight papers read automatically and checked by people

Start your data warehouse

List the reports that are still assembled by hand and the programs the figures come from. We build the first report and reconcile it against the numbers you use today.

Email
office@apply.at

Availability
Monday to Friday, 8:00-17:00 Austrian time (CET/CEST), reply within one working day

Meetings
By video on Microsoft Teams or Google Meet

By ticking this box I allow Apply to process my details for the purpose of answering this request (see our privacy notice for details)