> For the complete documentation index, see [llms.txt](https://help.rankability.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.rankability.com/track/understanding-spi.md).

# Understanding SPI

SPI is Rankability's 0–100 **Search Performance Index**. It combines the tracked search surfaces that apply to a project so you can monitor overall visibility and then drill into the evidence behind a change.

SPI is not traffic, conversions, market share, sentiment, or a prediction of future performance. Use it to compare the same tracking scope over time.

## Standard category weights

| Category           | Weight |
| ------------------ | -----: |
| Traditional search |    30% |
| Video              |    10% |
| AI mentions        |    35% |
| AI citations       |    25% |

## Local-query category weights

When a project is configured as a local query and includes Local Pack evidence, SPI uses a local-first weighting:

| Category           | Weight |
| ------------------ | -----: |
| Local              |    60% |
| Traditional search |    10% |
| Video              |     5% |
| AI mentions        |    14% |
| AI citations       |    11% |

## How the calculation works

1. Rankability calculates a 0–100 score for each active category.
2. Results inside a category are weighted by platform and, where relevant, ranking or citation position.
3. Category scores are combined using the standard or local weights above.
4. Categories with no applicable selected data are removed and the remaining category weights are normalized to total 100%.

This means a missing category is not automatically a measured zero. By contrast, a completed check where the brand is absent can contribute zero inside an active category.

Traditional and video position credit declines as the result moves down the rankings. Positions below 30 receive no position score. AI mention credit is based on whether the visible answer names the brand. AI citation credit uses citation presence and position. Local scoring uses the center result and available grid positions.

Platforms inside each category have different weights. SPI is therefore not a simple average of every selected platform.

## Selected surfaces change the scale

SPI uses the surfaces that the project actually measures. Adding or removing a platform, changing locations, or changing a per-platform AI tracking mode can alter the scoring scope even when existing results do not move.

Where AI tracking-mode controls are available:

* **Both** includes the platform in AI mentions and AI citations.
* **Citations only** removes its mention contribution.
* **Mentions only** removes its citation contribution and can hide the AI citations tab when no citation-tracking AI platform remains.

When comparing projects in a client-level or portfolio view, check for mixed scoring scales. Two projects with different selected surfaces or tracking modes are not perfectly equivalent even when their SPI values match.

## Video and Local Pack details

YouTube Search, Google Video Pack, and TikTok Search can contribute to the video category when selected and available. A confirmed absence of a Google Video Pack removes that pack from the video calculation instead of treating the missing surface as an owned ranking failure.

Supported video mention evidence can receive partial visibility credit even when the brand does not own the ranking video. That is different from position credit for an owned result.

For local projects, the Local category uses the configured Local Pack center and grid evidence. Grid coverage and rank are therefore more important to the local SPI than the standard organic category.

## Score bands

| Score  | Label          |
| ------ | -------------- |
| 90–100 | Very Strong    |
| 70–89  | Strong         |
| 50–69  | Moderate       |
| 30–49  | Weak           |
| 15–29  | Very Weak      |
| 0–14   | Extremely Weak |

The label summarizes the calculated visibility score; it is not a judgment about the business, campaign quality, or reputation.

## Diagnose an SPI change

1. Confirm the project, keyword, location, and comparison dates.
2. Check whether the selected platforms or AI tracking modes changed.
3. Open the category breakdown to identify which component moved.
4. Open the corresponding traditional, video, AI answer, citation, or local evidence.
5. Distinguish a completed absence from **No scan**, **No history**, **Not tracked**, stale data, or a failed platform.
6. Compare more than one completed run before treating normal volatility as a trend.

GSC clicks, impressions, and GA4 engagement can help explain business impact, but they do not directly become SPI. Compare those connected-data metrics separately instead of assuming an SPI increase caused a traffic or conversion change.

## Common interpretation mistakes

* Comparing projects with different platform or location coverage.
* Treating an unavailable category as zero.
* Treating an AI citation as a brand mention.
* Treating positive sentiment as part of SPI; sentiment is a separate evidence dimension.
* Comparing a local-query SPI with a standard-query SPI without considering the different category weights.
* Reporting a one-run change as a durable trend.

## Next steps

* [Read the Track overview](/track/reporter-executive-summary.md)
* [Read traditional and video results](/track/traditional-search-tab.md)
* [Read AI answers](/track/ai-answers-tab.md)
* [Read AI citations](/track/ai-citations-tab.md)
* [Connect GSC and GA4 evidence](/track/data-integrations-overview.md)