Select the data source
First we clarify whether the source is a database, API, file, accounting system, shop or another system. The key question is not the connector name, but which report it should support.
Understand in 90 seconds what Klarzahlen does – and watch the complete flow in about six minutes: source, data run, data quality, KPI logic, monthly report and monitoring. All with demo data, no real customer data.
The short overview shows the problem and the Klarzahlen cockpit: five stations that every monthly report passes through.
Stand-alone short video (70 seconds) – the full end-to-end demo follows below.
The complete demo with the demo tenant for June 2026: 12,482 rows loaded, revenue CHF 42.8k, costs CHF 30.6k, margin 28.4%, 3 data quality findings and export as dashboard, PDF or Excel. No real customer data and no secrets are shown.
Deutsche Version: zum deutschen Demo-Video.
0:00 Every month, the same routine: the monthly report is assembled by hand from Excel, the shop, ads, accounting and a database. Numbers get copied, formulas get patched – and nobody is sure everything is current and complete. Klarzahlen turns this into a repeatable process.
0:25 This is the Klarzahlen cockpit with a demo tenant for June 2026. Every monthly report passes through the same five stations: data source, data run, data quality, KPI logic, and the finished report.
0:50 Everything starts with the data source: accounting such as Bexio, a Shopify shop, CSV and Excel files, a database or an API. What matters is not the connector name but which report should come out of it. Every source stays cleanly assigned to one tenant.
1:35 Credentials are stored separately and used only for data runs – that is why you never see a password or token on screen. Klarzahlen then checks which tables or endpoints exist: orders, customers, payments.
2:15 Next, the load strategy: full load, delta load or a scheduled refresh – for example every night. A monthly report rarely needs realtime; a reliable, traceable run matters more.
2:55 Every data run lands in the run history: 12,482 rows loaded, duration and error status. If a run fails, the error is visible right away – instead of surfacing in the report.
3:35 Before the report, Klarzahlen checks data quality. The demo tenant has three findings: bookings without a cost center, a duplicate customer number, and a CSV export older than the cut-off date. These findings stay attached to the report.
4:15 KPIs are not a black box: the margin is defined as revenue minus product costs, minus ads, minus payment fees, minus returns. For June 2026 that gives 28.4 percent.
4:55 The result is the monthly report: 42.8k revenue, 30.6k costs, a 28.4 percent margin – with the three data findings right next to it. The report is available as a dashboard, PDF or Excel, with fixed recipients. Next month, the same flow runs again automatically, including monitoring.
5:30 If your monthly report is still built by hand: book a free initial consultation or a system check.
0:00 Jeden Monat dasselbe: Der Monatsreport wird aus Excel, Shop, Ads, Buchhaltung und Datenbank von Hand zusammengebaut. Zahlen werden kopiert, Formeln angepasst – und niemand ist sicher, ob alles aktuell und vollständig ist. Klarzahlen macht daraus einen wiederholbaren Ablauf.
0:25 Das ist das Klarzahlen Cockpit mit einem Demo-Mandanten für Juni 2026. Jeder Monatsreport durchläuft dieselben fünf Stationen: Datenquelle, Datenlauf, Datenqualität, KPI-Logik und der fertige Report.
0:50 Am Anfang steht die Datenquelle: Buchhaltung wie Bexio, ein Shopify-Shop, CSV- und Excel-Dateien, eine Datenbank oder eine API. Entscheidend ist nicht der Connector-Name, sondern welcher Report daraus entstehen soll. Jede Quelle bleibt sauber einem Mandanten zugeordnet.
1:35 Zugangsdaten werden getrennt gespeichert und nur für Datenläufe verwendet – im Bild ist deshalb nie ein Passwort oder Token zu sehen. Danach prüft Klarzahlen, welche Tabellen oder Endpunkte es gibt: Bestellungen, Kunden, Zahlungen.
2:15 Dann wird festgelegt, wie geladen wird: Full Load, Delta Load oder ein geplanter Refresh – zum Beispiel jede Nacht. Realtime braucht ein Monatsreport selten; wichtiger ist ein zuverlässiger, nachvollziehbarer Lauf.
2:55 Jeder Datenlauf landet in der Run-Historie: 12'482 Zeilen geladen, Laufzeit und Fehlerstatus. Schlägt ein Lauf fehl, ist der Fehler direkt sichtbar – statt erst im Report aufzufallen.
3:35 Vor dem Report prüft Klarzahlen die Datenqualität. Der Demo-Mandant hat drei Hinweise: Buchungen ohne Kostenstelle, eine doppelte Kundennummer und ein CSV-Export, der älter ist als der Stichtag. Diese Hinweise bleiben am Report sichtbar.
4:15 Kennzahlen sind keine Blackbox: Die Marge ist definiert als Umsatz minus Produktkosten, minus Ads, minus Payment-Fees, minus Retouren. Für Juni 2026 ergibt das 28.4 Prozent.
4:55 Am Ende steht der Monatsreport: 42.8 tausend Umsatz, 30.6 tausend Kosten, 28.4 Prozent Marge – mit den drei Datenhinweisen direkt daneben. Den Report gibt es als Dashboard, PDF oder Excel, mit festen Empfängern. Nächsten Monat läuft derselbe Ablauf automatisch wieder, inklusive Monitoring.
5:30 Wenn Ihr Monatsreport heute noch von Hand entsteht: Buchen Sie ein kostenloses Erstgespräch oder einen System-Check.
The demo walks through the practical flow: select a source, configure access, inspect tables or endpoints, choose the load pattern and check the first run before a report is trusted.
First we clarify whether the source is a database, API, file, accounting system, shop or another system. The key question is not the connector name, but which report it should support.
Then URL, host, user, API key or file location are configured. Secrets do not belong in reports or screenshots; they are stored separately and used only for data runs.
Klarzahlen checks which tables, files or API endpoints are relevant. Technical names are translated into understandable report terms.
Depending on volume and freshness, Full Load, Delta Load or a scheduled refresh may be enough. Realtime is chosen only when the business value justifies the extra operating effort.
Before a report is used, a run is tested: row counts, duration, missing values, duplicate keys and clear error messages matter more than a nice-looking chart.
The result is not only a dashboard. What matters is clear KPI logic, freshness, the last successful run and an export or report that finance, management or an accounting firm can use.
This is how a recurring monthly report with example data could be structured.
June 2026: revenue up, margin stable, 3 data quality findings open.
CHF 42.8k revenue from 12,482 order rows, up 7.4% versus the previous month.
CHF 30.6k costs including ads, shipping, product costs and payment fees.
CHF 12.2k profit, 28.4% margin according to the defined KPI rules.
CHF 8.7k open items, 5 invoices overdue, incoming payments visible.
Ads +14%, shipping costs +9%, return rate slightly lower than in the previous month.
3 findings: missing cost centers, duplicate customer number, stale CSV export.
Add cost centers, review ad campaigns, confirm the KPI rule for returns.
For companies with manual monthly reporting and several data sources.
For recurring client reports and standardized analysis.
For revenue, marketing, and order data from several systems.
For teams with multiple sources, open items, and management reports.
The video shows a guided example flow with demo data. In a call we additionally look at your concrete reporting problem, your data sources and which freshness actually makes sense.
Bring a short list of today’s data sources and manually built reports.
We show how data sources, runs, reports and monitoring work in a typical B2B reporting workflow.