Send Power BI sales to Kunads

For a business whose real sales live in Power BI. Kunads compares what Google and Meta claim with what actually sold — so it needs the net sales, once a day. Two ways: a daily flow (about an hour to set up), or a CSV export once a month. Nothing is installed in your Power BI and Kunads never changes anything in it.

Before you start

  1. In Kunads: Connect → Sales → Sheet · CSV · API → API → Create import key, for the right account. Copy the key (it starts kimp_); it is shown once.
  2. Know which Power BI dataset (semantic model) holds the orders, and the names of its order number, date and net value columns.

Option A — a daily Power Automate flow

Runs every morning and re-sends the last 60 days, so returns and late edits are corrected — an order sent again is updated, never doubled. Sending an order again updates it. Needs one Power Automate Premium licence for the person who owns the flow (the HTTP step is a premium connector, about $15 a month).

  1. In Power Automate: Create → Scheduled cloud flow. Name it Kunads daily sales, every 1 day at 06:00.
  2. New step → Power BI → Run a query against a dataset. Pick the workspace and the dataset. Paste this query and change the table and column names to yours:
EVALUATE
SELECTCOLUMNS (
    SUMMARIZECOLUMNS (
        'Sales'[OrderNumber],
        'Sales'[OrderDate],
        'Sales'[Status],
        FILTER ( ALL ( 'Sales'[OrderDate] ), 'Sales'[OrderDate] >= TODAY () - 60 ),
        "Net", SUM ( 'Sales'[NetRevenue] )      -- net of VAT, shipping and returns
    ),
    "id",     'Sales'[OrderNumber],
    "date",   'Sales'[OrderDate],
    "status", 'Sales'[Status],
    "value",  [Net]
)
-- one row per ORDER (its lines summed), the last 60 days every morning,
-- so a return made weeks later still corrects its order
  1. New step → HTTP. Method POST, URI https://app.kunads.ai/api/import.
  2. Headers: Authorization = Bearer kimp_YOUR_KEY, and Content-Type = application/json.
  3. Body — type {"orders": } and put First table rows from step 2 (Dynamic content) after the colon. In code view it reads like this; Kunads reads Power BI's [id]-style column names as they are:
{
  "orders": @{outputs('Run_a_query_against_a_dataset')?['body/firstTableRows']}
}
  1. Save → Test → Manually. The HTTP step should answer {"ok": true, "stored": …}. In Kunads the orders appear on the account within a minute.
  2. Change only the table and column names on the left of the query to yours. One run sends up to 5,000 orders: a shop with more than that in 60 days uses 30 days, and the first history load runs a month at a time (change the date filter, run once, change it back).

Option B — a CSV once a month (free)

  1. In a Power BI report, make a table visual with the order number, date and net value (plus status, email or campaign if you have them).
  2. … (More options) → Export data → Summarized data → .csv.
  3. In Kunads: Connect → Sales → Sheet · CSV · API → CSV file, choose the account and upload. Columns are found by name, in English or Lithuanian; uploading again updates what is already there.

What Kunads needs in each row

The full field list and other ways to send data: the import guide.