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

# Quickstart

> From a fresh account to your first prediction

This walkthrough takes you through the full loop once. Each step links to a detailed guide if you need more.

<Note>
  You'll need an Allyy account (follow the invitation email) and access to a data source — your CRM, a warehouse like BigQuery, or even a CSV export.
</Note>

<Steps>
  <Step title="Create a dataset">
    A dataset is the container for everything that follows. Create one for the body of work you have in mind (e.g. "Direct mail appeals").
  </Step>

  <Step title="Connect a data source">
    Inside the dataset, add a data source and enter its credentials. Preview the data to confirm it loaded correctly.

    → [Connect a data source](/processes/data)
  </Step>

  <Step title="Map your data">
    Map your columns onto the Allyy data model — Contacts, Responses, Offers, Subscriptions. This is what lets Allyy understand gifts, campaigns and supporters generically.

    → [Data mapping](/processes/data_mapping) · [The Allyy data model](/data/data-model)
  </Step>

  <Step title="Train a model">
    Create a model, choose the question you want to answer (e.g. **DM propensity** — who is likely to respond to a mailing), and start training. Allyy builds the features and trains on your history.

    → [Create & train a model](/processes/model) · [What Allyy can predict](/models/what-allyy-can-predict)
  </Step>

  <Step title="Generate predictions">
    Point the trained model at the population you want to score. Allyy produces a score for each supporter.

    → [Generating predictions](/predictions/generating-predictions)
  </Step>

  <Step title="Make a decision and export it">
    Turn the scores into an optimised, ranked contact list, then export it to your campaign tools.

    → [From predictions to decisions](/predictions/decisions) · [Exporting decisions](/predictions/exporting)
  </Step>

  <Step title="Measure the result">
    After the campaign runs, use the model-evaluation dashboards to compare predicted vs actual, and the analytics dashboards to see the effect on your supporter base.

    → [Model evaluation](/dashboards-evaluation/index) · [Analytics dashboards](/analytics/overview)
  </Step>
</Steps>

## What next?

<CardGroup cols={2}>
  <Card title="Understand your file first" icon="chart-line" href="/analytics/overview">
    No model needed — the analytics dashboards describe your supporter base as it is today.
  </Card>

  <Card title="Automate the loop" icon="arrows-spin" href="/automation/workflows">
    Once it works manually, schedule it with a workflow so it runs on its own.
  </Card>
</CardGroup>