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

# Allocation

> Attribute spend to applications — even when resources aren't tagged

Accurate showback and chargeback depend on knowing **which application each dollar belongs to**. In practice, plenty of resources go untagged. Koala's **application allocation** closes that gap.

## Coverage first

Start by checking how much of your spend is already attributed. In the [cost explorer](/finops/explorer), pivot by `tag:Application` to see labeled coverage and spend per application — and how much remains untagged.

## Recommending application labels

For resources with **no application tag**, Koala recommends a label. The recommendation engine is a **transparent, similarity-based classifier** — not a black box:

<Steps>
  <Step title="Learn from what's already tagged">
    Koala reads existing app tags (`Application`, `app`, `product`, `project`, and similar keys) to build labeled profiles for each application.
  </Step>

  <Step title="Find untagged candidates">
    Resources without an app tag are grouped by resource to cut line-item noise.
  </Step>

  <Step title="Score and recommend">
    Each candidate is matched to the closest application profile and returned with a **confidence score**. You control a minimum-confidence threshold, and Koala reports **diagnostics** (sample rows, how many fell below the threshold) so an empty result is always explained.
  </Step>
</Steps>

<Note>
  Because the classifier is transparent and confidence-scored, you can see *why* a label was suggested and tune how aggressive the matching is — rather than trusting an opaque model.
</Note>

## Tag untagged resources

You don't have to fix tags back in each cloud. From the allocation page you can **apply an `Application` tag to a resource in Koala**; it's picked up on the next [ingestion](/finops/connections/ingestion) run. You can also **export your Koala tags** (to CSV or as a clipboard runbook) to push them back into your cloud tagging, so the two stay in sync.

## Split a shared resource

Some costs belong to more than one owner. **Split allocation** lets you divide a single resource's cost across segments — owner, department, or cost center — with weighted percentages, so shared spend lands accurately on each business unit.

## From coverage to chargeback

With more spend labeled, application pivots become complete enough to drive **showback and chargeback** — giving each business unit an accurate, self-serve view of what it spends.

## Next steps

<CardGroup cols={2}>
  <Card title="Cost explorer" icon="table" href="/finops/explorer">
    Validate coverage and spend by application.
  </Card>

  <Card title="Insights" icon="lightbulb" href="/finops/insights">
    See what's driving each application's cost.
  </Card>
</CardGroup>