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

# Evidence and data states

> Interpret observed, calculated, inferred, partial, unavailable, and not-measured results correctly.

Alesta separates what a source directly reported from what the system calculated or inferred. It also separates missing evidence from a real measured zero.

## Evidence states

| State      | Meaning                                                                 | How to use it                                          |
| ---------- | ----------------------------------------------------------------------- | ------------------------------------------------------ |
| Observed   | Read directly from an available page, provider response, or measurement | Treat it as source-bound evidence, not universal truth |
| Calculated | Derived from observed values with a defined transformation              | Keep the inputs and scope in mind                      |
| Inferred   | Interpreted from observations                                           | Review and correct before relying on it                |
| Proposed   | Suggested as a candidate, such as a competitor                          | Accept, reject, replace, or investigate                |
| Confirmed  | Accepted or supplied by a person                                        | Treat it as deliberate workspace context               |

## Availability states

| State                 | Meaning                                                                    |
| --------------------- | -------------------------------------------------------------------------- |
| Preparing or queued   | The relevant stage has not started yet                                     |
| Generating or running | Work is active for that surface                                            |
| Ready                 | A usable result is available for the stated scope                          |
| Partial               | Some requested evidence is readable and some is missing                    |
| Unavailable           | The source or operation did not produce a usable result                    |
| Not measured          | The current run did not ask or could not ask that question                 |
| Disconnected          | A required provider is not connected                                       |
| Needs selection       | Authorization exists, but you must select a property, account, or resource |
| Locked                | The current plan does not include the reading or action                    |
| Coming soon           | The product surface is visible, but the capability is not available yet    |

## Unavailable is not zero

This is the most important reading rule in the product.

If Google does not have enough eligible Chrome UX samples, Alesta cannot report real-visitor Core Web Vitals for that domain. That does not mean the site had zero visitors or perfect vitals.

If a backlink provider returns no usable record, the panel can be unavailable. That does not prove the site has zero backlinks.

If an AI-visibility run did not execute, Alesta must not display zero mentions. It should say not measured or unavailable.

<Warning>
  Never copy an unavailable value into a report as `0`. Preserve the state and
  the source limitation.
</Warning>

## Lab data and field data

PageSpeed scores come from Lighthouse lab runs. Core Web Vitals in Alesta are real-visitor field data when eligible data exists. The two can disagree because they answer different questions.

* Use lab data to reproduce and diagnose a controlled page load.
* Use field data to understand eligible real-user experience over the provider's reporting window.
* Do not substitute lab values when the field-data panel is unavailable.

## Plan state and data state are separate

An upgrade can unlock a run or panel, but it cannot guarantee that a provider has data for a particular domain. Read the sequence this way:

1. **Access:** Does the plan include the capability?
2. **Readiness:** Is the required connection, selection, or service available?
3. **Data:** Did the source return usable evidence?
4. **Quality:** Is the result complete, partial, inferred, or otherwise caveated?

## References

* [Availability, plans, and access](/concepts/availability-plans-and-access)
* [Read partial and unavailable results](/guides/read-partial-and-unavailable-results)
* [Connections and missing data](/troubleshooting/connections-and-missing-data)
* [Glossary](/reference/glossary)