Los artículos de ayuda existen en alemán e inglés. Aquí los lees en inglés.
Informes, costes y exportación
How do I export data?
Under “Exports & AI connections” you download CSV files for conversions, customer journeys and attribution by channel, with the same values the reports show, so you can recalculate them yourself. All data and settings of your account at once come from the full export there, a ZIP with JSON and CSV per table, on every plan.
Can I recalculate the numbers myself?
| Export | Contents |
|---|---|
| journeys.csv | per journey its identifier, start, last activity, number of touchpoints and conversions, first and last channel, and value |
| conversions.csv | conversions with revenue, refund, currency, attribution status and acquisition value |
| attribution.csv | attributed conversions, revenue and acquisition value with cost, per channel, source, campaign, ad, creative, placement or publisher site |
Conversions and journeys deliberately appear individually in the exports, not as totals: anyone who doubts a figure in a report should be able to recalculate it with their own tools. An export is split into part files of 200,000 rows each; together they are the complete set for the selected period, on every plan. If a set has several parts, the page offers one button per part. “attribution.csv” contains every row of the breakdown, even when the table under “Attribution” shows only the 200 largest; the page then says how many there are in total.
Every row carries its basis: model, post-view yes/no, period, environment, conversion type and the time of the export. Model and window can be changed later, and the data keeps growing. A figure without these details cannot be recalculated a week on.
You choose the period and the model at the top of the “Exports & AI connections” page, as on every report; the page states period, time zone and environment before you download. In “conversions.csv” and “journeys.csv” the “modell” column stays empty: neither file is bound to a model. “zugeordnet” there means that at least one of the models, with or without post-view, gives the conversion a share; with “nein” the reason is next to it. The shares per model come from “attribution.csv” with the selected model and your post-view setting.
The acquisition value, meaning the value you gave a conversion in the portal, is in a column of its own in “conversions.csv” and “attribution.csv”, after the basis columns. In “attribution.csv” it is followed by “website”, the channel metrics (besuche, conversion_rate_zugeordnet, neukunden_zugeordnet, neukundenquote) and the subscription renewals (folgezahlungen_zugeordnet, umsatz_folgezahlungen_zugeordnet). In “conversions.csv” it is followed by “folgezahlung” (ja or nein) and “abo”, the subscription ID. New columns are always added at the end: anyone who reads a file by column position finds every other column where it was before.
If the text in a cell starts with =, +, -, @ or a tab, it gets a leading apostrophe. Excel, Google Sheets and LibreOffice would otherwise run the cell as a formula, and an order number or a campaign name can come from outside, for example from a link with UTM values. The spreadsheet then shows the text as it is. Numbers are not affected: a negative value stays a number. The same applies to the CSV attached to the weekly and monthly report.
Take everything with you: the full export
Under “Exports & AI connections”, card “Take everything with you”, “Start full export” starts an export of all data and settings of your account: one JSON and one CSV file per table, such as events within the retention period, journeys, conversions, costs, web analytics, websites, channels, goals, rules and instructions, plus the account and its members with name and address. We write it in the background, in several ZIP parts for large accounts, and send you an email as soon as it is ready. The page shows its status and then the parts and the index “manifest.json” for download. This is how you take your data to another provider; what each file and field means is in “Data structure of the export”.
- On every plan, including Start and Pro. The read API remains a feature from Growth upwards.
- It can be started and downloaded by whoever manages the account, that is the Admin role. It contains more than measurement data, such as the log and the members’ addresses.
- The parts are available for seven days, after closing at most until the account is deleted. After that they are deleted; a new export is possible at any time.
- A provider you switch to can fetch the export itself: with an export key that you create under “Take everything with you”. It opens the full export and nothing else, on every plan.
- It is not tied to a period: events are included as far back as retention goes. Raw rows of delivered ads that we delete from the database after 180 days come in shortened form from the analytics store into a file of their own, raw_events_archive, with the ten columns it keeps.
- It also stays available during the 30 days after closing the account and during a lock because of an outstanding payment.
- Credentials for other platforms are not included, nor are checksums of keys and invitations. A new provider connects with its own access.
- The CSV files use a semicolon as separator and UTF-8, with column names in the first row. Structured values appear as JSON in the cell.
Note
For a small account there is also the direct download below, as a single ZIP without waiting. It is created within the request and is therefore available only up to 300,000 rows; above that, use the export in the background, which is built for any size. If a restart interrupts the export in the background, it starts again from the beginning, so that all parts show the same state. After three restarts it counts as failed and you receive an email. You can then start it again or reply to the email, and we will provide the data.
Note
For programmatic access, use the read endpoints under /v1/data with an API key of scope “read” instead of the export. The read API comes with Growth, Scale and the agency plan. Without model and includePostView the endpoints use your default model and your post-view setting, like the reports, and state both in the response. If a script compares figures across weeks, it should name the model explicitly: otherwise the figures change whenever someone changes the default model. For dashboards there is /v1/data/daily: one row per day, channel and campaign, whose sums match those under “Channels & attribution”. For Google Sheets, Power BI and Excel there are ready-made templates for it in their own articles.
Revisado por última vez el 19 oct 2026
En el portal del cliente
También se pregunta
- Download CSV
- Export data
- Get raw data
- Export for Excel
- How do I get my figures into a dashboard?
- Why is there an apostrophe in front of an order number?
- Why does a cell in the export start with a single quote?
- Export all data
- How do I take my data to another provider?
- Switching provider: how do I get everything out?
- Full export
- Can I export all settings?
- Is there an export on the free plan?
- Back up data before closing the account
- Full export in the background
- The full export download breaks off
- Can the new provider fetch my data itself?
- Are impressions older than 180 days in the full export?
- What happens if the full export is interrupted?
Artículos relacionados
- Which reports are available?
What can I analyse in the portal?
The reports are in the sidebar: Overview, Customer journeys, Channels & attribution with its breakdown and model comparison, Orders & goals, plus Check budget, Platform reconciliation, Data status and Events, among others. In the period row at the top you choose the period and the model and, once your account measures more than one website, the website.
- How do I ask questions about my own data?
What can the portal chat do, and what can it not?
Under “Ask your data” a language model answers questions about your measurement data and about how to use the product. It reads only through defined tools, states only figures that came from a tool call, and openly reports which tools it used.
- How do I connect my own AI tool?
Is there an MCP server for Metrunio?
Through Metrunio’s MCP server. From Pro, Claude and ChatGPT sign in via OAuth and read totals such as key metrics, channels, campaigns, commerce and web analytics; from Growth also individual conversions and journeys. From Growth, other tools use an API key of scope “read”, the same one a script would use for /v1/data.
- How do I get my figures into Google Sheets?
Can I pull my figures into a Google Sheets spreadsheet that updates itself every day?
With a short Apps Script that fetches your daily rows per channel and campaign through the read API every morning and refills a sheet. You keep the read key in the script properties, not in the code and not in a formula. The read API comes with Growth, Scale and the agency plan.
- How do I get my figures into Power BI or Excel?
Can I load my figures into Power BI or Excel and refresh them there?
With a Power Query query that loads your daily rows per channel and campaign through the read API; Power BI and Excel both understand the same query. You create the read key as a parameter, and the query sends it in the request header, never in the address. The read API comes with Growth, Scale and the agency plan.
- How does Metrunio count subscription renewals?
How do I mark subscription renewals, and which channel are they attributed to?
You report a renewal as a purchase with the subscription ID (subscription_id) and the flag subscription_renewal; Metrunio marks renewals from Stripe itself. It counts in measured revenue, appears separately in the reports and receives the shares of the first order of the same subscription, because it has no ad contact of its own. Without a known first order or without consent, it stays unattributed with a reason.
- How do I pay, and where are my invoices?
How do I purchase a plan, switch it, and where do I find invoices?
In the portal under “Plan & usage”: there you book and switch plans, choose monthly or yearly payment and reach invoices, payment methods and, when paying monthly, cancellation at the payment provider. Paid monthly there is no minimum term, and yearly payment runs for twelve months unless you close your account; cancelling means moving to the free Start plan, and closing your account is a separate step. Through the Shopify app you can pay in Shopify instead.
- Data structure of the export
Which files and fields does the full export contain, and in which format?
The full export consists of ZIP files with one JSON and one CSV file per table, plus the index manifest.json. This article lists every file, every field with format and meaning, and the tables that are deliberately left out. It is built from the same source as the export; a new provider can use it to read the export, also by itself with an export key.