Frequently Asked Questions

How often is data refreshed?

SponserdIQ runs an automated scrape cycle that crawls sports property websites to discover sponsorship relationships. A supplementary web search fills in gaps that HTML scraping may miss. New data typically appears shortly after a scrape completes.

Between scrapes, existing records retain their last_seen_at timestamp. If a sponsorship is not found during a scrape, it is not immediately removed -- it stays in the database with a stale last_seen_at, which the churned_sponsors and renewal_tracker skills use for churn detection.

What does the enrichment sparkle mean?

The sparkle indicator on the dashboard marks records that have been enriched by AI. For brands, this means fields like company size, revenue range, and description were filled in by the enrichment pipeline. For sponsorships, it means category label and activation type were AI-generated.

Enriched fields are based on AI analysis, not direct scraping. They provide useful directional data but should not be treated as confirmed facts.

Can I export data?

The MCP integration enables flexible data export through your AI assistant. You can ask Claude to query sponsorships with specific filters and format the results as CSV, tables, or any other format. For example:

"Search for all NFL title sponsorships and format the results as a CSV"

The search_sponsorships tool supports pagination up to 100 results per page, so large datasets can be retrieved across multiple pages.

Is there a limit on the in-app chat assistant?

Yes. Each account includes a monthly allowance for the in-app chat assistant — it keeps responses fast and keeps the service fair for everyone. Your allowance refreshes automatically at the start of each month, you can see your current usage in the top bar of the chat, and you can always ask the assistant "how much of my chat allowance is left?"

How the allowance is shared depends on your plan:

Personal plans have their own individual monthly allowance, just for you.
Team plans share a single pooled allowance across the whole team. The pool grows with the number of seats on your plan, and any member can draw from it — so lighter users naturally leave more room for heavier ones, and you manage one combined budget instead of many small ones.

Reached your limit before it resets?

You don't have to wait for the monthly reset. You can keep working by connecting SponserdIQ to your own Claude app (Claude Desktop, claude.ai, or any MCP-compatible client). When you explore SponserdIQ's data through your own Claude connection, the conversation runs on your own Claude account instead of our in-app assistant — so the monthly chat allowance no longer applies. You get the same sponsorship data and the same tools, with no in-app cap. See Getting Started to connect in a couple of minutes.

How do merged brands work?

When SponserdIQ discovers that two brand records refer to the same company (e.g., "Coca-Cola" and "Coke"), they can be merged by an admin. One brand becomes the canonical record, and the other is marked with a merged_into_id pointing to it. The merged brand's aliases are preserved so future scrapes continue to match correctly.

Merged brands are excluded from all MCP queries and skills -- only canonical (non-merged, non-deleted) brands appear in results.

What sports and leagues are covered?

SponserdIQ intentionally focuses on under-covered segments of sports sponsorship — areas like women's soccer, minor league teams, and other properties that are underserved by traditional sponsorship intelligence tools. These are the spaces where insight is hardest to find and where our data adds the most value.

We also track sponsorships across major professional and collegiate properties for breadth of coverage, since many brands active in smaller leagues also sponsor larger ones.

Because our scraping approach is flexible, we can extend coverage to virtually any public sponsorship source. Use the list_properties tool or get_filter_options to see what is currently tracked. If you're interested in a specific league, segment, or property we don't yet cover, email us at hello@sponserd.ai and we'll look into adding it.

How does churn detection work?

SponserdIQ detects churn by comparing a sponsorship's last_seen_at timestamp against the scrape cycle. The system offers two levels of churn detection:

Renewal Tracker categorizes sponsorships in real-time. Thresholds are tied to the ~90-day property scrape cadence (plus a 3-day grace buffer):

Active -- seen within the most recent scheduled scrape (within ~93 days)
At risk -- missed one scheduled scrape (~94-183 days since last seen)
Stale -- missed two or more scheduled scrapes (more than ~183 days since last seen)
Inconclusive -- the sponsorship would be at-risk or stale, BUT the underlying property hasn't been successfully scraped recently either. In that case we can't confidently tell whether the deal is churning or our scraper just can't see the page. Rather than report a false-positive churn signal, we mark it inconclusive.

Churned Sponsors uses a configurable threshold (default: 90 days) to flag sponsorships that have likely been dropped entirely. This pulls everything older than one full scrape cycle and then checks the property's own scrape health — if the property has been re-scraped since the sponsor went missing, the deal is reported as churned; otherwise it stays inconclusive.

Removed Sponsorships is the strongest, AI-verified layer. SponserdIQ's autonomous Data Audit Agent re-validates every sponsorship on a rolling schedule using Claude + web search. When the agent concludes (with high confidence + citations) that a deal has lapsed or is pointing the wrong way, it soft-removes the sponsorship and stamps a removal_reason. These rows drop out of every public/dashboard query and are inspectable via the removed_sponsorships skill — useful for "what specifically did the agent take off Brand X's portfolio?" The historical signal is also captured per-week per-brand in the removed_count column on the snapshot table, queryable via brand_portfolio_history (single brand) or sponsorship_movement (population-wide leaderboards by net change: new - removed).

Three tiers, each definitive at its own level:

Renewal Tracker — fast, probabilistic. Fires within one missed scrape cycle.
Churned Sponsors — slower threshold-based proxy. Fires after ~90 days of no sighting once the property has been re-scraped.
Removed Sponsorships — slowest, definitive. Agent verified the deal is over with web evidence.

Renewal status is also surfaced directly on the dashboard — each sponsorship row shows an Active / At Risk / Stale / Inconclusive badge, and a summary row at the top of the page lets you filter the table to any of the four buckets with a single click. The same statuses are filterable via the search_sponsorships MCP tool using the renewal_status parameter.

All three skills are accessible via run_skill with renewal_tracker, churned_sponsors, or removed_sponsorships.

Where do the source URLs on each sponsorship come from?

Each sponsorship returned by MCP (and displayed in the dashboard brand tooltip) includes a sources list of up to 5 URLs pointing to the specific pages that evidenced that deal — a press release announcing the partnership, the property's own partner-page listing, a trade-press article covering the sponsorship, etc. Sources come from two paths:

HTML scrapes contribute the partner page URL where the sponsor's logo/listing was extracted (always present).
Web searches contribute the specific URLs Claude cited while identifying that sponsor (present when Claude attributes sources; we aggregate across scrapes to fill gaps over time).

Citations accumulate across every scrape of a brand-property pair — so even if one run doesn't attribute a source, older runs remain visible. An empty sources list means we haven't captured a machine-readable citation yet; the next scrape cycle fills the gap.

For a property-wide view (admin-curated partner pages plus every web citation from recent discovery), use the get_property_sources MCP tool. For a pipeline-quality overview (which outlets supply the most evidence, coverage rates per extraction method), run the source_intelligence skill.

What does the ⚠ next to a property name mean?

When a property's most recent scrape hasn't produced usable sponsor content within the last cycle, the dashboard shows a small warning icon next to the property name. Hovering reveals a tooltip with:

Last successful scrape -- when we last pulled at least one identifiable sponsor from the property's partners page
Last AI enrichment -- the most recent time any sponsor under this property was enriched with company intelligence

A degraded scrape signal usually means the upstream site has changed structure, added bot protection, or is temporarily returning empty content. While this state persists, at-risk and stale sponsorships on that property show as Inconclusive rather than reporting false churn.