> ## Documentation Index
> Fetch the complete documentation index at: https://woku.app/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# External study import

> Load your own research or a third party's as a data source and unify it with your clients through a data flow

The **external study import** lets you bring into woku your own research or a
third party's, for example an NPS benchmark or the results of a survey run
with another tool, and put it to work alongside your data: unify the
respondents with your clients and send them new measurements.

The path relies on two modules of the **Data** group in the side menu:
[Data sources](/docs/en/datos/fuentes-de-datos) to load the study and
[Data flow](/docs/en/datos/flujo-de-datos) to activate it.

## Use case

You have a file with the results of a study: each row is a respondent with
their email, their score and their comment. You want those respondents to
exist as clients in woku, without duplicating the ones you already have, and
you want to invite them to a new measurement to give continuity to the
study.

## Import the study as a data source

The study is loaded by creating a **data source** with the four-step
wizard:

1. **Connect**: upload the study file in **CSV, Excel/XLSX, JSON or XML**
   format (up to 50 MB). If the study provider exposes an API, you can also
   connect it directly by providing the URL, the method (GET or POST) and an
   optional API key.
2. **Review data**: woku detects the study fields and shows the type and an
   example of each one. Include the fields you care about, such as email,
   score and comment, and exclude the rest.
3. **Define schema**: with the AI assistant, or manually, you define the
   display name, the type, the format (for example `email` or `date`), the
   description and whether each field is required.
4. **Publish**: the source becomes available to build data flows on top of
   it.

The full detail of the wizard is in
[Data sources](/docs/en/datos/fuentes-de-datos).

## Unify with your clients and activate the study

With the source published, create a **data flow** that uses it as the
origin. That is where unification and activation happen:

* The **create or find client** action looks up each respondent by their
  **email**. If a client with that email already exists in your company, the
  record is linked to it. If it does not exist, a new client is created. This
  way the study's universe is unified with your base without duplicates.
* The same flow can **send measurements** to the study's universe: **NPS**,
  **CSAT** or **CES** surveys, invitations to a **woku** or **forms**.

Before running you can **simulate** the flow to review what it would do with
the source records, and then launch the **manual execution** with real
sends.

The detail of building, simulation and execution is in
[Data flow](/docs/en/datos/flujo-de-datos).

## Current limits

* Flows operate on the **sample of up to 50 records** that each data source
  stores. For larger studies, process the study in parts.
* Unifying respondents with clients is done **by email**. Rows without an
  email cannot be linked to a client.
* Loading is by file or by an API query. There is no real time event
  ingestion.