ClusterSignal

Blog

How to use ClusterSignal to find high-conviction setups

A practical guide to the table dashboard, cluster detail pages, Short Radar, and the AI analysis feature.

June 2026 · 5 min

Start with the table

The dashboard is a dense, scannable table — not cards. Columns show Ticker, Company, Cluster Type (BOUGHT / SOLD / SHORT RADAR), Score (0-100 with grade badge), Net Volume, Insider count, Date, and a direct SEC EDGAR link. Sort by grade, capital, or most recent activity using the filter bar.

Click any row to go to the cluster detail page. It loads fast and shows the full picture without cluttering the dashboard.

Use the score to prioritize

Grade A (>=80) clusters have the strongest combination of factors: typically 3-4+ insiders, meaningful capital ($500K+), C-suite participation, and tight timing. Grade B (>=50) is still significant. Grade C is worth tracking but interpret with more caution.

The score breakdown on the detail page shows exactly how each factor contributed — which insiders drove the seniority score, how tight the timing was, and whether historical win rate data is available.

Check the direction and context

BOUGHT clusters and SOLD clusters are different research inputs. Buyers commit real capital with their own money — that is a clear positive signal. Sellers may be diversifying, meeting tax obligations, or executing a pre-scheduled 10b5-1 plan.

If you see a Short Radar badge (orange), the sell cluster passed additional filters for size and coordination. That warrants deeper investigation: check the detail page for the insider track records, earnings proximity, and whether a dilution risk flag appears.

Use AI analysis for faster context

On the cluster detail page, signed-in users can generate an AI analysis powered by Claude. It produces a 2-4 paragraph summary covering what the pattern means, what the insiders roles suggest, and what to watch next. It always ends with a disclaimer: this is not financial advice.

Use the AI summary as a starting point for your own research — not a conclusion. Review the underlying filings, upcoming earnings, sector backdrop, and valuation before forming a view.

Validate before you act

ClusterSignal is a research tool, not investment advice. A high-scoring cluster is an interesting pattern, not a trade signal. Cross-reference with SEC EDGAR directly (the link is on every row), check earnings dates, review recent news headlines, and consider the company sector context.

Past patterns do not guarantee future results. Use ClusterSignal as one input in a broader research process.