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

# What a customer costs and what they return

> The five inputs behind an account's economics, the two reads built from them — one backward, one forward — and the verdict Beacon reaches at a renewal point.

Retention decisions get made on instinct more often than anyone admits — usually on revenue alone, which says nothing about whether the account is worth the effort. Beacon reads five things instead.

<Note>
  A designed model. The inputs below are published today as separate values; the two composite reads and the verdict are not.
</Note>

## Five inputs

| Input | What it is |
| - | - |
| What it cost to win them | Acquisition cost attributed to the account |
| What it costs to keep them | Ongoing cost of serving them — published as [`cost_to_serve`](/definitions/cost-to-serve) |
| What they have returned | Margin realised so far, not revenue |
| What they are likely to return | Remaining margin over the relationship, plus room to expand |
| How likely that is | The inverse of [`churn_risk`](/definitions/churn-risk) |

Margin rather than revenue, throughout. Two accounts on the same contract value can differ by a wide margin once support load and delivery cost are in.

## Two reads

**What has happened.** The first three inputs, combined — what the relationship has actually been worth to date.

**What is ahead.** The last two — what it is likely to be worth from here, weighted by how likely it is to continue.

They answer different questions and they can disagree. An account that has already returned its acquisition cost several times over can still be a poor place to put the next unit of effort, and an account that has not yet paid back can be the best one.

## At a renewal point

Beacon reaches one of three positions: **worth further investment**, **review**, or **not worth further investment**.

Getting to that position is arithmetic — the same inputs give the same verdict every time, and it is reproducible from your data. **Choosing what to do about it is not Beacon's.** The verdict identifies; a person decides, and there is a floor below which Beacon will not put an account in the bottom category on its own at all.

## Where the numbers come from

Your finance system. Beacon reads cost and margin figures and never replaces them — the canonical version of any of these stays where your accounting lives. If a figure here disagrees with your ledger, the ledger is right and Beacon has a stale or partial read, not a better one.

## Related

<Columns cols={3}>
  <Card title="Cost to serve" href="/definitions/cost-to-serve">
    The one input published in full today.
  </Card>

  <Card title="Customer value" href="/definitions/customer-value">
    A related published measure of what an account is worth.
  </Card>

  <Card title="Proposals, not actions" href="/proposals-not-actions">
    Why a verdict is not a decision.
  </Card>
</Columns>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.