SS9SS DIGITAL
01 / The Nine Signals
Every B2B SaaS funnel emits nine distinct buying signals before a deal closes — product-usage events, third-party intent surges, partner-sourced meetings, pricing-page revisits, and five more. Clicks are the least predictive of the nine. The SS9SS model weights all nine, then re-validates weekly against closed-won revenue.
The page below catalogs every signal, its default weight range, and the raw event patterns that feed it — so you can audit your current model against what a defensible one should look like.
02 / The Nine Signals
Each signal has a default weight range calibrated against 412 closed-won SaaS deals. Weights re-baseline weekly per account, validated against a 10% holdout.
Activations, feature expansions, and repeat logins inside a trial or free tier. The single highest-weighted signal in SaaS because it is the only one emitted by an already-qualified buyer.
e.g. feature.expanded → workspace.created within 72h
Spikes in category-level research on Bombora, G2, Gartner Peer Insights, and six other intent vendors. Catches demand before it ever lands on your pixel.
e.g. intent.score > 72 for 3+ sessions in 14d
Repeat pricing-page views across sessions and devices. Strongest single buying-stage tell: a buyer comparing tiers is closer than any ad click will ever be.
e.g. /pricing viewed 3× across 7d, ≥2 devices
Introductions logged by integration partners, agencies, and reseller channels. Low volume, but the highest single-event win-rate we measure across the deployment base.
e.g. meeting.booked with source=partner
Form fills, calendar opens, and ROI-calculator completions that stop short of a booked demo. They predict the demo that follows; last-click credit eats the ad spend that drove them.
e.g. form.submitted on /roi-calculator
AE/SDR reply chains on Outreach and SalesLoft. Filtered for role-fit and email-open clusters so a forwarded internal thread is not mistaken for buyer interest.
e.g. email.replied after open ×3 in 5d
Third-party outlets (TechCrunch, VentureBeat, niche trade press) plus off-domain syndications. Best as a top-of-funnel awareness tell, not a stage predictor.
e.g. article.dwell > 45s on partner domain
Lift in branded query volume on Google and Bing. Confirms the rest of the model is working — if branded search is flat, mid-funnel signals are not compounding.
e.g. "<brand> pricing" search +18% WoW
A positive reply to an SDR cold sequence. Lowest-weight of the nine, but the only outbound signal that is voluntary — every other one is observable behavior.
e.g. email.replied sentiment > 0.6
03 / Data Sources
Native, code-free connectors for HubSpot, Salesforce, Segment, Outreach, LinkedIn, Google Ads, Meta, TikTok — plus 41 more. Live in under 11 minutes.
Median connector time-to-live: 11 min 04 sec · 49 native connectors · zero SDK rewrite
04 / How the Weights Are Set
Every SS9SS weight starts as a Bayesian prior pulled from the 412 closed-won SaaS deals already on the platform. We do not hand-tune weights per customer — that is a recipe for confirmation bias and a weak audit trail. Instead, the prior is set once per quarter, then re-weighted per account every Monday at 06:00 UTC.
Re-weighting is a constrained gradient step against two loss terms: lift in marketing-sourced pipeline within 90 days, and agreement with a 10% holdout of closed-won revenue that the model never sees during training. If a weight drift would push the holdout error above 4%, the re-weight is rejected and the prior holds. We publish the holdout error per account in the trust portal.
Why this matters to a VP of Marketing: you can show the model to your CFO without translating it into a black-box vendor promise. The weights have a number, the number has a basis, and the basis is reproducible.
05 / Proof at Scale
06 / Next Step
A senior strategist maps the nine signals to your real HubSpot, Salesforce, and ad-platform data. No generic pitch, no canned deck — just a working model against your pipeline, plus the 90-day lift forecast it produces.