Historical Metrics Dashboard

The Historical Metrics Dashboard shows how your AI visibility changes over time. It displays trends in brand consistency, share of voice, mentions, citations, and gap opportunities, helping you measure the impact of your optimization efforts.

What You'll Learn

In this guide, you'll learn:

How the AI Visibility Funnel separates awareness, consideration, and decision metrics
The difference between Share of Influence and Share of Voice
How to interpret each chart on the dashboard
Why AI metrics naturally fluctuate
How to identify real improvements versus normal variation
How to use time range, platform, and tag filters to slice the funnel

Prerequisites

This dashboard requires data from AI Visibility Reports. Run reports regularly to populate historical data and track trends over time.

Understanding Data Variability

Important: AI output is non-deterministic by nature. Ask an AI assistant the same question twice and you may get different responses. This means your brand might appear in one answer but not the next, even for identical queries.

Because of this uncertainty, charts display data as ranges rather than exact values. The shaded bands around metrics represent typical variability (approximately ±8%). This is normal and expected.

What Counts as Real Change

Trust trends, not single points: One report showing 45% SOV and the next showing 50% doesn't necessarily mean improvement. But three consecutive reports trending upward likely indicates real progress.
Look for sustained movement: A metric that stays outside the confidence band for three or more consecutive reports probably represents genuine change.
Consider context: If metrics improve right after publishing major content, that's likely causal. Random improvements with no changes are probably just variation.

It's not uncommon to see small decreases in reports run close together. The goal is to capture statistically significant improvements over time, not to react to every fluctuation.

AI Visibility Funnel

The top of the dashboard is a three-stage funnel summarizing your AI visibility at every point in the buyer journey:

AWARENESS                       │   CONSIDERATION         DECISION
Share of Influence (SoI)        │   Share of Voice (SoV)  Share of Voice (SoV)
"Is our category framing        │   "Is our brand surfaced when AI
showing up before any brand     │    recommends brands at the bottom
is named?"                      │    of the funnel?"

A dashed divider separates the awareness segment from the consideration/decision pair to make one thing visually explicit: the awareness segment is measured differently from the other two. SoI and SoV are not the same metric at different stages — they're two metrics measuring two different things.

Share of Influence is the average rate at which your positive key messages surface in AI answers to awareness-stage prompts — the prompts a buyer asks when they've identified a problem but haven't formed brand preferences yet.

For each positive-sentiment Key Message, Spyglasses computes the percentage of awareness-stage AI responses that contain at least one sentence semantically matching the message. The funnel hero displays the average across all positive key messages, restricted to the selected platform.

What feeds SoI:

Prompts with query type category_entry_point, jobs_to_be_done, or buyers_journey_awareness. Buyer's Journey and Category Entry Points frameworks (see Generating Prompts with AI Frameworks) generate them; the Coverage Matrix tab on the Prompts dashboard helps you spot gaps.
Key messages marked as Positive sentiment. Negative and Neutral messages don't contribute to SoI.

A low SoI means buyers asking AI about your category aren't hearing your framing yet — the moment-of-truth happens before they know to ask for your brand by name. A high SoI means your messaging strategy is shaping the AI's mental model of the category.

Share of Voice is your brand mention rate in AI answers to consideration- and decision-stage prompts — the prompts where AI surfaces specific brand recommendations.

Crucially, the funnel hero's SoV — and the summary card and chart below — is scoped to consideration + decision only. Awareness-stage executions are excluded because they rarely mention any brand by definition (and including them would deflate SoV with results that aren't measuring what SoV is supposed to measure). The headline SoV percentages you see on this dashboard are (brand mentions across C+D) / (opportunities across C+D).

Consideration and Decision are split into two cards inside the SoV group because clients often want to distinguish them: consideration-stage prompts ("best CRM for marketing teams") tend to enumerate several brands, while decision-stage prompts ("CRM under $50/user with HubSpot integration") tend to converge on fewer.

Clicking a Stage

Each of the three cards is clickable. Clicking expands a per-stage trend chart below, scoped to the current platform and tag filters. The trend uses the same per-day collapsing semantics as the other charts: multiple measurements on the same day collapse to one point with the latest measurement winning.

Filtering the Funnel

Three filters re-scope the funnel:

Filter	Effect
Platform	Re-scopes all three segments. Picking ChatGPT shows SoI and SoV for ChatGPT-only executions. "All Platforms" shows the rolled-up view.
Tag	When the property has at least one tagged prompt, a Tag select appears. Picking a tag re-scopes the funnel to just the prompts carrying that tag. See Custom Tags.
Time range	Funnel headline numbers reflect the latest measurement in the selected window; the per-stage trend chart shows every point.

A note on legacy data: reports written before the funnel feature shipped may have null breakdown JSON. Those data points still chart against the overall totals, but won't move when you change platform or tag. New reports and the next nightly run populate the breakdown JSON, after which the filters work as expected for that point onward.

How Measurements Are Recorded

Every measurement is a point in time. Each AI Visibility Report writes its own data point. Nightly prompt runs that test a subset of your discovery queries write a separate data point per day. As a result, you'll commonly see multiple measurements on the same date — especially during setup or testing — and that is expected:

Running two reports on the same day produces two separate points on the chart.
A report run and the nightly aggregation on the same day produces two separate points as well.
The dashboard's summary card always shows the latest measurement in the selected window. The chart shows every point.

If a date appears to have a "different number" from a report you just ran, scrub the chart to that date — you'll usually see two points, with the summary card reflecting whichever ran last.

Aliases, Competitors, and Recomputation

Brand mentions and Share of Voice are computed by matching your property's company name + aliases against the names that appear in AI responses, and your competitors + their aliases to determine the SOV denominator. When you edit any of these, the historical metrics recompute in the background — typically within 10–30 seconds — so the dashboard becomes eventually consistent with your latest brand definition.

In practice: edit an alias, give the recompute a brief moment, then refresh. Old data points will reflect what your current brand definition would have produced for the underlying AI responses captured at the time.

How Data Is Scoped for Locations

If your brand has locations, historical metrics are scoped to the property you're viewing:

On the parent brand: charts show only the brand-wide rollup — the parent's own prompt runs without any location context. Per-location data is excluded from the parent's series so the trend lines read cleanly. To see how each location compares, use the Share of Voice by Location widget on the parent's main dashboard, or open Historical Metrics on a specific location.
On a location: charts show only that location's data points — its own nightly runs and any AI Visibility Reports run against it. The parent's brand-wide series is excluded.

This scoping applies to every chart on the dashboard (consistency, share of voice, mentions, citations, gap trends) and to the metric summary cards.

A few details worth knowing:

Brand identity always follows the parent. Even when viewing a location's dashboard, the brand-mention matcher uses the parent brand's company name and aliases. The location's own label (e.g. "Downtown") is never used for matching. Editing the parent's aliases recomputes historical metrics for every location automatically.
Competitors are per-location. Each location maintains its own competitor list, so the SOV denominator on a location's dashboard reflects its local competitive set — which may differ from the parent's.
Locations and parent accrue independently. Adding a new location does not retroactively affect the parent's brand-wide series; existing data points keep their original meaning. New nightly runs after the location is created accrue into the location's bucket alongside the parent's brand-wide bucket.

Charts Explained

Brand Consistency Over Time

This chart shows your consistency score (0-100) across all AI platforms. Brand consistency measures how reliably your brand appears when AI platforms answer questions in your space.

What to look for:

Upward trend: Your optimization efforts are working
Flat trend: You're maintaining position but not improving
Downward trend: Either you're slipping or competitors are improving faster
Large jumps: Often indicate new content or authority signals taking effect

The shaded band shows typical variation. Small changes within this band should be interpreted directionally, not as definitive improvements.

This dual-axis chart shows your Share of Voice percentage (left axis) and total mention count (right axis).

Share of Voice is your primary competitive metric. It shows the percentage of AI responses where your brand appears compared to all opportunities — restricted to consideration- and decision-stage prompts (see Share of Voice in the funnel hero for why awareness-stage executions are excluded). The series and the summary card use the same C+D-only definition, so they line up with the funnel hero.

Mentions provide context for your SOV percentage:

Rising SOV with flat mentions means you're holding steady while competitors decline
Flat SOV with rising mentions means everyone's improving together
Rising both indicates you're genuinely growing visibility

Citations Over Time

Citations indicate how often AI platforms reference your domain as a source. This is a strong signal of authority and trustworthiness.

Citation patterns to understand:

Steady growth indicates building authority
Sudden spikes often follow publication of data-rich content
Citations often preceed mentions (you get cited first, and mentions are drawn from them)

Grounding Search Gap Trends

This chart tracks two metrics:

Gap count (area): Total searches where competitors rank but you don't
Average competitor rank (line): How well competitors are positioning

Interpreting gap trends:

Declining gaps: You're successfully closing visibility opportunities
Rising gaps: Competitors are finding new ranking opportunities
Improving competitor ranks: The competitive bar is rising
Declining competitor ranks: Optimization opportunities competitors haven't found

Using Time Range Filters

The time range selector lets you focus on specific periods:

7 Days

Use for:

Testing impact of specific content published this week
Monitoring immediate response to technical changes
Checking if a competitor suddenly improved

30 Days

Use for:

Regular monthly performance reviews
Evaluating recent optimization efforts
Comparing this month to last month

90 Days

Use for:

Preparing quarterly business reviews
Assessing long-term strategy effectiveness
Identifying seasonal patterns

Custom Range

Use for:

Comparing before/after a site redesign
Measuring impact of a content campaign with specific dates
Analyzing performance during a product launch period

Top Gap Opportunities

Below the charts, you'll find a list of your highest-priority gap opportunities. These are grounding searches where competitors rank but you don't.

For each gap, you'll see:

The specific search query competitors rank for
How many discovery queries use this grounding search
Which competitor ranks best and at what position

Using This List

Start with gaps used in multiple queries (highest leverage)
Check if you have content targeting this search
Review what competitors rank for this search
Create or optimize content to rank in the top 30
Re-run your AI Visibility Report to verify improvement

Best Practices

Track After Major Changes

Whenever you publish significant content or make technical changes, note the date. Check Historical Metrics 7-14 days later to see the impact.

Set Quarterly Goals

Based on your current baseline, set realistic quarterly targets:

5-10% improvement in Share of Voice
10-20% increase in citations
20-30% reduction in gap opportunities

AI visibility improves gradually. Expecting month-over-month doubling sets you up for disappointment.

Document Your Efforts

Keep a simple log of optimization work with dates. When metrics improve, you'll know what likely drove the change.

AI Visibility Reports - Run reports to generate historical data
AI Visibility Rankings - See detailed ranking data for grounding searches
Prompts - Manage prompts, custom tags, and the awareness-stage prompts that feed Share of Influence
Key Messages - Define the positive narratives that drive Share of Influence
Locations - Track AI visibility separately per market

Historical Metrics Dashboard

On this page