Two ways to get Impact data out.
impact.com has a documented REST API for brands, with reports, action-level data and partner invoices. The work sits in report exports that run as background jobs on a daily quota, and in commission that can change until an action locks.
Build and run it yourself
- Create an access token under Settings, Technical, API, with read-only scopes. Its
Account SIDandAuth Tokenare the username and password for HTTP Basic authentication. - Call
GET /Advertisers/{AccountSID}/ReportExport/{ReportId_or_Handle}withSUBAID(the program ID),StartDateandEndDate. It queues a job: poll/Jobs/{Id}, then download the file. - Stay inside the quotas: 100
ReportExportrequests a day and 1,000 requests an hour on most other endpoints. Past them you get a 429. - Read past periods again. A pending action can be modified or reversed until its locking date, and a report keeps refreshing until 90 days after its date range ends.
- Check the report's
ApiAccessiblefield first. API access is set per report, and custom reports built in Data Lab cannot be retrieved. - Sum cost to spend per day, map programs and partners to your own markets, and line it up with the rest of your media data.
Then monitor it, and test each new API version before you move your account to it.
Have Odins run it
- Access is set up with you at onboarding. Credentials go in through your Odins onboarding link, never by email.
- We sync every day, watch the pipeline and fix it when the API changes.
- You get what you pay partners as daily spend, in the same structure as every other channel.
- A country split variant of the connector gives you the figures per country.
The same goes for every other source you use. One team runs all of them. See all connectors
What the Impact API offers brands.
A short map of the reporting side of impact.com's Brand API. In impact.com's terms the advertiser is a brand, the publisher a partner and a credited conversion an action. impact.com's own documentation is the reference.
Base path: https://api.impact.com/Advertisers/{AccountSID}
| Endpoint | What it returns | Good to know |
|---|---|---|
GET /ReportExport/{ReportId_or_Handle} |
Queues a report as a background job and returns one URI to poll and one to download the file. | CSV by default. An identical request returns the same job, cached for 60 minutes to 32 hours. |
GET /Reports |
Every report available to the account, each with an ApiAccessible flag. | Running a report here is the legacy route, with 20,000 records a page and a cap on pages. |
GET /Actions |
Each conversion credited to a partner, with payout, sale amount, currency and state. | One program per call. StartDate and EndDate span 45 days at most and filter on the last update, not the event date. |
GET /ActionUpdates |
A changelog of actions at line-item level, with the change in payout and sale amount. | No record is written when an action locks or is paid. |
GET /Invoices |
Partner invoices, with line items grouped by program. | Net amount and VAT are separate fields. |
- Auth HTTP Basic,
Account SIDandAuth Token - Format XML by default, JSON on request
- Quota 1,000 requests an hour, 100 report exports a day
- Version 14, set per request with
IR-Version
Worth knowing about Impact data.
Commission is not final until the action locks.
An action starts as pending and can be modified or reversed until its locking date. impact.com recommends a locking period of 27 days after the end of the month, and a program can change it. Read once, a cost series keeps the pending figure.
Total cost is more than commission.
In impact.com's partner performance report, Total Cost is the sum of Action Cost, Click Cost and Other Cost. Other Cost covers performance bonuses, make-good payments and placement fees.
Action data is kept for 750 days.
impact.com retains action data for 750 days. Older actions cannot be retrieved through the API.
What you get back.
Two layers. The raw one mirrors Impact. The structured one is the table every source lands in, organised by your own markets and product lines.
Structured layer
Per day| In the structured table | For Impact |
|---|---|
| Date | Daily |
| Company and market | Set by your routing rules, or per country with the country split variant |
| Channel and media | Affiliate, Impact |
| Source campaign | Agreed with you at onboarding |
| Spend | What you pay partners |
Raw layer
As impact.com returns itThe source data as impact.com returns it, delivered next to the structured table.
Use it where you work.
In your warehouse
Raw and structured tables delivered to BigQuery, Snowflake, Redshift or Databricks. Some customers use Odins for this alone.
In Odins
Reporting across every channel, and a model that shows what affiliate contributes and where more spend still pays.
Through your AI assistant
Connect Claude, ChatGPT or Copilot to Odins with MCP and ask for the numbers in plain language.
Questions.
Does Impact have an API?
Yes. impact.com documents a REST API for brands, currently version 14. You authenticate with an Account SID and Auth Token from an access token created under Settings, Technical, API.
How do I get Impact data into BigQuery?
Odins sets up the impact.com connection with you at onboarding, syncs it daily and delivers raw and structured tables to your BigQuery project.
How far back does Impact data go?
impact.com retains action data for 750 days. Odins requests history from 2020 and loads what impact.com returns.
Can Odins split Impact data by country?
Yes. A country split variant of the connector exists. Otherwise spend is placed on your companies and markets by routing rules.
We use more than one affiliate network.
Every affiliate and price comparison connector lands in the same structure, so the networks can be compared side by side. See all affiliate connectors.
-1.png?width=1132&height=292&name=Asset%201@2x%20(12)-1.png)