Quattr Leads AEO, SEO, and Content Rankings on G2 Spring 2026.
Quattr Leads AEO, SEO, and Content Rankings on G2 Spring 2026.
Connect Claude, ChatGPT or Cursor to Quattr and ask questions the way you'd ask a colleague. Your assistant runs governed Quattr analyses on your organization's own data: real numbers, never a model's guess, with answers linked back to the app.
MCP (Model Context Protocol) is an open standard that lets AI assistants securely use external tools. The Quattr MCP server gives your assistant a set of ready-made, trusted Quattr analyses it can run for you. When you ask "How did my clicks trend last month vs the month before?", the assistant picks the right Quattr analysis, runs it against your organization's data (secured by OAuth sign-in with your Quattr account), and answers with real numbers, often as a rich, interactive result card with a link back to the Quattr app.
Use the same work email you use for www.quattr.com, and make sure you've signed in to the Quattr app at least once.
In your AI client, add the Quattr MCP server URL https://mcp.quattr.com/mcp and sign in when prompted. Step-by-step per client →
Try: "What were my total organic clicks and impressions for last month?" or browse the prompt library.
Quattr gives your assistant 58 analysis tools across nine areas. Your question is routed to the right one automatically, so you never need to know tool names.
Clicks, impressions, CTR and rankings: totals, breakdowns, trends, movers, CTR curves and striking-distance opportunities.
Your share of the market vs tracked competitors, by segment, intent, keyword, page or category, with gains and losses.
How often ChatGPT, Perplexity, Gemini and Google AI Mode cite your pages: trends, competitors, prompts and coverage gaps.
Google Ads spend, CPC, cost per conversion, impression share, and paid/organic keyword overlap.
Sessions, users, engagement, conversions by goal, funnels, and statistical significance of changes (GA4 / Adobe).
LCP, INP, CLS and Lighthouse grades: weakest pages, device splits, and traffic-weighted views.
AI crawler activity (GPTBot, ClaudeBot…), Googlebot verification, crawl-budget waste, orphan pages, indexation gaps.
Is a change real or noise? Do two metrics move together? Rigorous significance tests and cross-domain correlation.
Find and run your existing Quattr dashboards and saved reports from chat.
In clients that support interactive results (like ChatGPT and Claude), answers render as native Quattr cards. Every card is honest about where its numbers came from:
| Card element | What it tells you |
|---|---|
| Headline takeaway | A one-line summary of the key finding, color-coded neutral / positive / caution. |
| Scope strip | Chips echoing exactly what produced the numbers: date range, comparison period, organic vs paid, brand filter, device, segment, and any filters. If the numbers look surprising, check the scope first. |
| observational | The data is observational: it shows what happened, not what caused it. Correlations are not proof of causation. |
| Freshness note | "as of <date>" plus any known lag (e.g. Search Console data lags 1–2 days). |
| Caveats | Any honesty notes, e.g. a filter that couldn't be carried into a deep-link, or a modeled estimate. |
| Open in Quattr | Deep-links into the Quattr app with the same view, dates, segment and filters already applied, so you can keep exploring. |
| Open in Adhoc Analysis | Opens the exact data behind the card in Quattr's Adhoc Analysis view: the precise underlying table, which you can tweak, save, and add to any of your Quattr dashboards. |
Sign-in binds your session to your organization. Every result is stamped with a company_name label, and the assistant is instructed to stop if it ever doesn't match the company being discussed.
Sources update on different cadences: Search Console lags 1–2 days, web analytics loads nightly, market share and AI visibility ~daily, Core Web Vitals roughly biweekly. Ask "How fresh is my data?" anytime.
Up to 60 requests/minute per user and a daily per-account analysis budget (resets midnight UTC). Normal conversational use never hits these.
If a data source isn't part of your subscription, or a question can't be answered faithfully (e.g. an unsupported metric trend), you get a clear "not available", never a made-up number.
→ Connect your AI client (Claude, ChatGPT, Cursor, Claude Code, VS Code…)
→ Copy-paste example prompts to see what's possible
→ Browse all 58 analysis tools
→ Troubleshooting, security & FAQ
Connect the Quattr MCP server to your AI assistant in about two minutes. Setup is the same everywhere: add the server URL, sign in with your Quattr account, start asking questions.
| Requirement | Details |
|---|---|
| A Quattr account | Sign in with the same work email you use for www.quattr.com. You must have signed in to the Quattr app at least once before connecting. A brand-new email that has never logged in will be rejected at sign-in. |
| MCP access for your organization | MCP access is enabled per organization by Quattr. If sign-in is refused for your whole team, contact your Quattr customer success manager. |
| A supported AI client | Quattr approves each AI client explicitly for security. Currently approved: Claude (claude.ai web & desktop), Claude Code, ChatGPT, Cursor, and VS Code. Using something else? See Other MCP clients. |
Click your client for step-by-step setup.
Quattr Analyticshttps://mcp.quattr.com/mcpNote: custom connectors require a Claude plan that supports them; on Team/Enterprise plans an admin may need to enable the connector for the workspace.
Setup: run the command below in your terminal, then authenticate inside Claude Code.
claude mcp add --transport http --scope user quattr https://mcp.quattr.com/mcp
claude
/mcp
quattr and follow the browser sign-in flow.Note: Claude Code shows results as clean text rather than interactive cards; the numbers are identical.
MCP connectors in ChatGPT are added via developer mode (availability depends on your ChatGPT plan and workspace settings).
Quattr Analyticshttps://mcp.quattr.com/mcpTip: for best results, create a dedicated ChatGPT project for Quattr analytics and add instructions like:
When using the Quattr MCP: always state the date range you analyzed; use the tools to discover segments, trackers and filter values rather than asking me; prefer the analysis the find_relevant_analysis tool routes to; check the company_name on results matches my company; and include the Open in Quattr link when a result card provides one.
Quick install: click Add to Cursor on this row (requires Cursor installed), then authenticate. Or set up manually:
mcp.json:{
"mcpServers": {
"quattr": {
"url": "https://mcp.quattr.com/mcp",
"transport": "streamable-http"
}
}
}quattr; sign in with your Quattr credentials.If tool calls start failing later, run "Clear All MCP Tokens" from the Command Palette and re-authenticate.
Quick install: click Add to VS Code on this row, approve the browser's "Open Visual Studio Code" prompt, then authenticate. Or set up manually:
mcp.json) and choose HTTP.{
"servers": {
"quattr": {
"type": "http",
"url": "https://mcp.quattr.com/mcp"
}
}
}Any MCP-compatible client that supports remote Streamable HTTP servers with OAuth can technically speak to Quattr, but for security, Quattr only admits clients that have been explicitly approved (allowlisted) for the service. If your client shows an authorization error like "this client is not permitted", it isn't on the approved list yet.
To request approval for a new client, contact your Quattr customer success manager with the client's name and its OAuth client metadata URL (CIMD), if it has one.
You never paste keys or passwords into the AI client. The client redirects you to Quattr's secure sign-in; after you approve, it receives short-lived, signed access tokens that refresh automatically.
Tokens carry the quattr:read scope only. The MCP can analyze your data but can never change your Quattr configuration or write anything.
The MCP inherits your existing Quattr account's access: your verified sign-in email maps to your Quattr user, which determines the organization and data every answer comes from. Connecting grants no additional access beyond what your account already has.
Connections stay valid for extended periods and survive Quattr deployments. An idle session expires after 7 days, after which the client reconnects transparently or asks you to re-authenticate.
Each conversation is bound to exactly one organization: the one your Quattr account is currently associated with.
company_name label naming the customer the data belongs to. The assistant is instructed to stop and tell you if it doesn't match the company you're asking about.The MCP resolves the business objects your team has configured in Quattr, by name, so you can just say them in chat. Manage these in the Quattr app; the MCP picks up changes automatically (within ~24 hours, or immediately in a new session).
| Object | What it does in chat | Example |
|---|---|---|
| Segments | Scope market-share and ranking analyses to a tracked keyword set. The primary "Overall Market" (default) segment covers your whole domain. | "market share in my Hosting segment" |
| AI visibility trackers | Scope AI-answer citation metrics (ChatGPT, Perplexity, Gemini, AI Mode) to a tracked prompt set. | "AI visibility on my Brand tracker" |
| Goals | Conversion breakdowns use your own GA goal names. | "conversions by goal: how many Purchases?" |
| Competitor sets | Competitive comparisons run against the competitors tracked per segment. | "compare us to Hostinger and GoDaddy" |
| Intents (keyword groups) & categories | Filter or break down almost any analysis by your configured keyword groups and content categories. | "clicks for my Informational intent" |
Tool availability follows your Quattr subscription and connected data feeds. If a source isn't connected, related questions get an honest "not available" instead of a wrong number. Ask "What data sources are connected for my site, and how fresh is each one?" anytime.
| Data source | Unlocks | Typical freshness |
|---|---|---|
| Google Search Console | Search performance: clicks, impressions, CTR, position, movers, CTR curve, striking distance | Lags 1–2 days |
| GSC via BigQuery (optional) | Far more granular per-query data; used only when you explicitly ask for the BigQuery source and your account has the feed | Lags 1–2 days |
| Market share | Share vs competitors, competitive rankings, keyword gaps, featured snippets, Google AI Overviews inclusion, AEO prompt attribution | ~Daily |
| AI visibility | ChatGPT / Perplexity / Gemini / AI Mode citation metrics, trends, competitor share of voice | ~Daily |
| Web analytics (GA4 / Adobe) | Sessions, users, engagement, conversions, funnels, significance tests | Nightly |
| Google Ads | Spend, CPC, cost per conversion, impression share, paid/organic overlap | ~Daily |
| Lighthouse / page experience | Core Web Vitals grades, weakest pages, traffic × LCP joins | ~Biweekly (varies by account) |
| Site crawl feed | Orphan pages, internal link flow, indexation gaps, crawl-budget waste | Per crawl schedule |
| Server web logs | AI crawler activity, Googlebot verification | Hours – nightly |
Limits keep Quattr analytics fast for everyone. Normal conversational use won't hit them.
| Limit | Value | When exceeded |
|---|---|---|
| Requests per user | 60 / minute | "Too many requests; retry in N seconds" |
| Daily analysis budget per account | 500 analysis queries or 60 minutes of query time / day | "Budget exhausted; resets at midnight UTC" |
| Concurrent heavy queries per user | 2 | Brief "overloaded" backpressure; the assistant retries |
| Longest single query | Up to ~10 minutes | Very heavy questions may take a while; the connection stays alive |
MCP access is enabled per organization by Quattr, and each AI client is individually approved before it can complete sign-in. Contact Quattr to enable access for your team or approve an additional client.
OAuth 2.1 with PKCE; short-lived signed access tokens with rotating refresh tokens; strict per-organization isolation enforced at the data layer on every query; read-only scope.
Tool calls are audit-logged server-side (durably shipped to secure storage), so usage is traceable per user and organization.
The AI assistant you connect (e.g. Claude or ChatGPT) processes the query results to write its answers. Review your organization's policies on AI-assisted data analysis before rolling out broadly.
The Quattr MCP exposes 58 tools. You never call them by name. Just ask a question and the assistant picks the right tool automatically. This reference helps you understand what's possible and why an answer was scoped the way it was.
How the assistant finds the right analysis, entities and filter values.
| Tool | What it does | Notes |
|---|---|---|
find_relevant_analysis | Routes any analytics question to the authoritative Quattr analysis. | Called first for every question. |
resolve_entity | Resolves the names you use (segments, AI trackers, goals, competitor sets, AI platforms) to the exact object. | Say "my Hosting segment" and it resolves automatically. |
list_filter_values | Lists the valid values (and operators) for a filter: country, intent, category, device, brand, rank bucket, page, query… | Prevents guessed filter values. |
describe_data_sources | Which data sources your account has, each with its latest data date and load cadence. | Great first question after connecting. |
get_data_freshness | Latest available data date per domain (search, analytics, market share, AI visibility, ads, page experience). | Use when numbers look stale. |
Your clicks, impressions, CTR and average position. Defaults to organic.
| Tool | What it answers | Notes |
|---|---|---|
search_performance | Totals or top keywords/pages for one period: clicks, impressions, CTR, position, unique keyword/page counts; optional trend or period comparison. | "Total clicks last month", "top 10 keywords". |
gsc_breakdown | One period broken down by page, country, device, category, intent, query, rank bucket, traffic source, language or geo. | Snapshot only (no time series). |
gsc_dimensional_trend | A day/week/month/quarter series by one of those drill fields, e.g. weekly clicks by page. | Multi-metric series (any mix of clicks, impressions and CTR in one call). |
seo_ranking_movement_report | Biggest gainers and losers in clicks, impressions or position between two periods. | "What changed vs last month?" |
ctr_curve_model | Your site's own CTR-by-position curve, plus queries over/under-performing their expected CTR. | Segmented branded vs non-branded. |
striking_distance | Keywords just off page 1 (positions 8–20) ranked by modeled click upside if nudged up. | Modeled estimate with an AI-Overview haircut, not a promise. |
compare_periods | Whole-site period-over-period deltas for a summarized result. | For site-level comparisons. |
run_looker_query | Advanced escape hatch: query a dataset by explicit fields/filters/sorts. | Used only when no standard analysis fits. |
You vs your tracked competitors. Share values are percentages on a 0–100 scale; analyses default to your primary "Overall Market" segment unless you name one.
| Tool | What it answers | Notes |
|---|---|---|
market_share_overview | Your share vs competitors, the segment leader, and the gap, sliceable by intent, category, page or brandedness. | Share only; pair with clicks via growth_vs_market_share. |
growth_vs_market_share | Market share side-by-side with one of your own metrics (clicks, impressions, CTR, position). | "Are we growing faster than our share?" |
market_share_breakdown | Share ranked across intents, keywords, pages, categories, brandedness, rank buckets or countries, incl. biggest gains/losses. | The multi-row market-share view. |
market_share_query | Share for you plus several named competitors in one table. | "Us vs Hostinger vs GoDaddy." |
competitive_keyword_ranking | Per-keyword average rank position for you and competitors. | Ranking (1 = top), not share. |
competitive_overview | Snapshot of one competitive metric vs rivals: featured snippets, Google AI Overviews inclusion, brand mentions, ranked keywords/pages. | Google on-SERP AI Overviews, not ChatGPT citations. |
competitive_trends | The same competitive metrics as a time series (plus market share and average rank). | Segment-level. |
total_search_share | Organic share + paid impression share combined into one total-SERP share. | A labeled estimate; SERP geometry is approximated. |
seo_keyword_gap_analysis | Keywords where competitors beat you: losing / winning / not-ranked buckets with gap sizes. | Per segment. |
How AI answer engines cite you. Distinct from Google's on-SERP AI Overviews (see competitive tools above). Most views are scoped to an AI visibility tracker.
| Tool | What it answers | Notes |
|---|---|---|
ai_visibility_overview | Your AI-answer scorecard: citation rate, share of voice, mentions, sentiment per platform. | Whole-account or per tracker. |
ai_visibility_trend | Citations / share of voice / sentiment over time. | "Going up or down?" |
ai_visibility_competitors | Which competitors lead you on AI visibility, ranked. | Per tracker. |
ai_visibility_citation_audit | Current citation counts and rate for a period. | "Are our citations dropping?" |
ai_visibility_drilldown | Full drilldown by tab: competitors, prompts, cited URLs, intent, geography, branded. | Mirrors the Quattr app view. |
ai_citation_attribution | Which prompts cite your pages the most. | Where you win AI answers. |
ai_page_citations | Which of your pages (or a competitor's) get cited, how often, and at what position. | Blog vs comparison vs docs categories. |
ai_prompt_coverage_gap | Prompts where you're missing or under-cited: your coverage gaps. | Useful for content planning. |
| Tool | What it answers | Notes |
|---|---|---|
paid_search_overview | Spend, CPC, cost per conversion, conversions: total, trended, or ranked by keyword/campaign/ad group/page. | Whole-domain. |
paid_impression_share | Search / top / absolute-top impression share and share lost to rank. | Auction coverage, per segment. |
paid_organic_overlap | Keywords you pay for while already ranking organically (top-3 flagged). | Observational, not proof of cannibalization. |
| Tool | What it answers | Notes |
|---|---|---|
web_analytics_overview | Sessions, users, conversions, conversion rate, bounce, engagement, filtered or broken down by country, device, page, source, category, intent. | The behavioral, on-site lens. |
conversions_overview | Conversion counts by your own goal names (Purchase, Sign-up…), plus transactions, revenue, conversion rate. | Uses your configured GA goals. |
funnel_dropoff | Where users drop between funnel steps (e.g. sessions → conversions → transactions) and the biggest leak. | Macro-funnel over count metrics. |
pop_significance | Whether a conversion-rate or traffic change is statistically real or noise. | Proper z-test / t-test with a verdict. |
| Tool | What it answers | Notes |
|---|---|---|
site_health_cwv_check | LCP / INP / CLS / performance grades, distributions, weakest pages, device splits, trends, Lighthouse scores. | CWV data typically refreshes ~biweekly. |
page_traffic_cwv | Your highest-traffic pages joined with their LCP: which busy pages are slow. | Observational join, top-N pages. |
From your server logs and Quattr's site crawl. Requires the corresponding feed.
| Tool | What it answers | Notes |
|---|---|---|
ai_crawler_analysis | How AI crawlers (GPTBot, ChatGPT-User, PerplexityBot, ClaudeBot, Google-Extended…) hit your site. | Server-log activity, not citations. |
bot_verification | Which "Googlebots" are real vs spoofed, via IP verification. | Web-logs feed. |
crawl_budget_waste | Deep, parameterized or non-200 URLs eating crawl budget. | Site-crawl feed. |
indexation_gap_diagnosis | Pages blocked from indexing and why (noindex, canonical, robots, status). | Site-crawl feed. |
orphan_pages | Pages no internal link points to. | Site-crawl feed. |
internal_link_flow | Incoming/outgoing internal-link counts per page; under-linked important pages. | Site-crawl feed. |
| Tool | What it answers | Notes |
|---|---|---|
significance_check | Is a period-over-period change real or noise? Delta, p-value and a plain verdict. | Works across search, share, analytics, paid, CWV. |
correlate_domains | Do two metrics from different areas move together? Pearson/Spearman with significance. | Correlation is not causation, and it says so. |
| Tool | What it answers | Notes |
|---|---|---|
find_relevant_content | The dashboards/saved reports most relevant to a question. | Prefers a standard analysis when one answers the question. |
list_dashboards | Your dashboards, most-used first. | |
run_dashboard | Chat-native dashboard: recommend, preview (runs tiles), or describe. | Includes a deep-link to the full dashboard. |
run_dashboard_tile | Runs a single tile, optionally with overridden filters/dates. | |
run_look | Runs a saved report by ID. |
| Tool | What it answers | Notes |
|---|---|---|
render_analysis | Renders a synthesis of prior results as a native Quattr card (report, table, stat cards, notice). | How multi-query summaries become cards. |
build_chart_spec | Builds a chart specification from result rows. | |
summarize_results | Safe descriptive stats: totals only for metrics that are actually additive. | Never sums market share, CTR, or ranks. |
explain_analysis | Explains what a named analysis would compute and its cost, without running it. | "How does striking_distance work?" |
link_to_report | A deep-link into the Quattr app for the matching report page. | Powers "Open in Quattr". |
get_query_history | Your recent analysis calls with timing and status. | Handy for "what did we run?" |
Counts reflect the tools available to connected AI clients. A few additional diagnostic tools exist for Quattr-internal use only.
Copy-paste any of these into a chat with the Quattr connector enabled. No special syntax needed; they're just questions. Swap the date ranges, segments and competitor names for your own.
Most issues resolve with one of three moves: reconnect the connector, start a new conversation, or sign in to the Quattr app once. Details below.
| What you see | What it means | Fix |
|---|---|---|
| "No Quattr/Looker account for <email>" at sign-in | Your email has never signed in to the Quattr app, so no account exists to bind to. | Log in to www.quattr.com once with the same email, then retry the connector sign-in. |
| Sign-in loops or "unauthorized" after previously working | Your token expired or was revoked. | Disconnect and reconnect the connector; in Claude Code run /mcp and re-authenticate. In Cursor, "Clear All MCP Tokens" from the Command Palette also helps. |
| "This client is not permitted" during authorization | Your AI client isn't on Quattr's approved-client allowlist. | Use an approved client (Claude, Claude Code, ChatGPT, Cursor, VS Code) or ask Quattr to approve your client. Details → |
| "Session unknown or terminated" | Your session expired (7 days idle) or the server was updated. | Nothing to do; the client re-initializes automatically. If it persists, start a new conversation. |
| "Tenant changed since session initialization" | Your Quattr account switched to a different customer mid-conversation. | Start a new conversation (or reconnect). Conversations are pinned to the customer they started with. |
| "Too many requests; retry in N seconds" | You hit the 60 requests/minute per-user limit. | Wait a few seconds. Break huge multi-part requests into smaller questions. |
| "Budget exhausted; resets at midnight UTC" | Your account used its daily analysis budget. | Resume after the UTC reset, or contact Quattr if your team regularly needs more. |
| "Overloaded" / brief 503s | Momentary capacity protection (concurrent queries, or a deploy in progress). | Retry in a few seconds; the assistant usually does this itself. |
| Tools missing right after a Quattr release | The client cached an older tool list. | Refresh/reconnect the connector and start a new conversation. |
| Results show as plain text, no visual card | Your client doesn't render interactive MCP cards (e.g. Claude Code), or the host skipped the visual. | Expected. The numbers are identical. Use "Open in Quattr" style links when offered. |
Check the card's scope strip first: date range, segment, organic vs paid, brand filter, device. The most common causes: a different date window (ranges include both endpoints; weeks run Sunday–Saturday), a different segment (the default "Overall Market" covers your whole domain; a tracked segment covers only its keywords), or comparing an organic-only number to a blended one. Cards carry an "Open in Quattr" link that opens the app with the same scope applied; use it to compare apples to apples.
Market-share values are already percentages on a 0–100 scale: 0.03 means 0.03%, not 3%. Deltas are in percentage points. Shares, rates and ranks are never summed into totals; that would be statistically meaningless, and the tools refuse to do it.
Sources load on different cadences: Search Console lags 1–2 days, web analytics loads nightly, market share / AI visibility / Google Ads roughly daily, and Core Web Vitals typically refresh about every two weeks. Every card's footer shows an "as of" date. Ask "How fresh is my data?" for the live per-source picture.
By design, the Quattr MCP refuses rather than approximates. For example, shares, rates and ranks are never summed into totals; a request like that gets a clear refusal instead of a misleading number. The refusal message names what is supported; usually a small rephrase gets you there.
The data shows what happened, not what caused it. A correlation between, say, slow pages and fewer clicks is a lead worth investigating, not proof. Significance tools ("is this real or noise?") tell you whether a change is statistically meaningful, which is still not the same as knowing its cause.
No. Tool availability follows your Quattr subscription and connected feeds (e.g. Google Ads, site crawl, BigQuery). The honest "not available" prevents wrong numbers. If you believe you should have that source, contact your Quattr customer success manager.
OAuth 2.1 authorization-code flow with PKCE. No API keys or passwords ever enter the AI client. Access tokens are short-lived and cryptographically signed; refresh tokens rotate on every use and can be revoked.
The MCP inherits the Quattr permissions of the user who signs in, so you can only read data your account already has access to. Connecting grants no additional access.
Your identity maps to your Quattr account, and every query runs inside your organization's data scope, enforced at the data layer. One conversation is bound to exactly one customer, and every result is labeled with its company_name.
The connection is scoped to quattr:read. Nothing the assistant does through the MCP can modify your Quattr configuration or data.
Only AI clients explicitly approved by Quattr can complete sign-in; unknown clients are rejected before any data access.
Tool calls are audit-logged server-side and shipped to durable storage, so usage is traceable per user and organization.
The AI assistant you connect (Claude, ChatGPT, etc.) processes query results to write its answers, under that provider's terms. Review your organization's policies on AI-powered data analysis and any compliance requirements (e.g. GDPR, CCPA) before broad rollout.
Quattr supplies the data; the narrative around it is generated by your AI assistant, which can make mistakes. Verify important numbers via "Open in Quattr" before acting on them.
No. Ask questions in plain language; the assistant discovers and calls the right tools. The tool reference exists so you know what's possible, not because you need it.
Claude (claude.ai web and desktop), Claude Code, ChatGPT (developer mode), Cursor, and VS Code. Other MCP clients can be approved on request; see Connect & Settings.
Yes. Anyone with a Quattr account for your organization can connect with their own email. Access is per-user; there are no shared keys.
Switch customers in the Quattr app as usual, then start a new conversation (or reconnect the connector). Existing conversations stay pinned to the customer they started with, and every result's company_name tells you which one you're looking at.
Use "Open in Quattr" to compare the card with the app view, note the question you asked and the card's scope strip, and share both with your Quattr customer success contact. That scope information is exactly what the team needs to reproduce the issue.
If your account has the BigQuery feed, say so explicitly, e.g. "using the BigQuery source, top queries by clicks…". It returns far more granular per-query data (privacy-suppressed queries appear under "[anonymized by google]"). Without the entitlement you'll get an honest "not enabled".
