1. Probability Scoring Fundamentals
Probability scoring quantifies the likelihood that a specific allocator will deploy capital into your fund within a defined time horizon — typically 90 days. It replaces subjective pipeline labels like "warm" or "interested" with a numeric score between 0 and 100, grounded in observable data rather than intuition.
The purpose is operational, not academic. A probability score answers the question every distribution team faces each morning: "Which allocators should we prioritize today, and why?" Without scoring, teams default to recency bias (whoever responded last gets attention) or relationship bias (whoever the senior partner knows best stays at the top of the list). Both patterns leave capital on the table.
Effective probability scoring enables three outcomes. First, pipeline prioritization — your team focuses outreach on the allocators most likely to convert, not the ones who happen to be top of mind. Second, capital forecasting — you can weight your pipeline by probability to produce realistic projections instead of aspirational totals. Third, resource allocation — you deploy your best salespeople against the highest-probability opportunities instead of spreading effort evenly across the entire pipeline.
AllocatorBase's scoring methodology is built on five dimensions, each weighted by its predictive power based on historical conversion data across alternative asset fundraises. The sections below walk through each dimension, the data inputs behind them, how scores are calibrated, and how they integrate into your CRM workflow.
2. The Five Scoring Dimensions and Their Weights
Each allocator is scored across five dimensions. The weights reflect the relative predictive power of each dimension based on conversion analysis across multiple fund cycles.
Mandate Alignment (30%) is the strongest single predictor. An allocator with a perfect mandate match — same strategy, geography, vehicle type, and check size range — converts at roughly 3x the rate of an allocator with a partial match. This dimension is scored 0–100 based on the overlap between the allocator's stated mandate and your fund's characteristics.
Engagement Intensity (25%) captures how actively the allocator is interacting with your firm. Multiple substantive meetings in the past 30 days score highest. Email exchanges without meetings score lower. No recent engagement scores near zero. Engagement is weighted by recency — a meeting last week matters more than a meeting three months ago.
AUM & Deployment Pattern (20%) evaluates whether the allocator has the capital capacity and the behavioral pattern to deploy. Growing AUM combined with recent alternative allocations scores highest. Declining AUM or no recent deployment activity scores lowest. This dimension uses SEC Form ADV data as its primary input.
Decision Timeline (15%) measures proximity to an active allocation decision. An allocator with a known decision window within 90 days scores highest. An allocator with no visible timeline scores lowest. Timeline data comes from direct conversation, RFP activity, and market intelligence.
Organizational Fit (10%) accounts for the allocator's internal decision-making structure. Firms with streamlined IC processes, prior experience allocating to alternative strategies, and a governance structure that can move within your fundraise window score higher. Large institutional allocators with multi-year evaluation cycles may score lower on this dimension even if other dimensions are strong.
3. Data Inputs: Where the Scores Come From
Probability scores are only as good as the data feeding them. AllocatorBase draws from three primary data sources, each mapped to specific scoring dimensions.
SEC Form ADV filings. These are the foundation. Form ADV provides verified data on AUM, number of clients, types of clients (high net worth individuals, pooled investment vehicles, pension plans, etc.), advisory activities, and compensation structures. AllocatorBase ingests quarterly ADV updates for 125,000+ registered firms. This data feeds the AUM & Deployment Pattern dimension and provides baseline inputs for Mandate Alignment.
Engagement signals from your CRM. When AllocatorBase integrates with HubSpot or Salesforce, it reads engagement data directly: email opens, meeting logs, document downloads, data room access, and deal stage progression. This data feeds the Engagement Intensity dimension and contributes to Decision Timeline scoring. The bidirectional sync means scores update as new engagement data enters the CRM — no manual refresh required.
Market intelligence and behavioral signals. This includes RFP activity tracked through industry channels, conference attendance patterns, public statements about allocation strategy, and peer deployment behavior within the same allocator segment. These signals are harder to systematize but contribute meaningfully to Decision Timeline and Organizational Fit scoring.
4. Score Calibration and Thresholds
Raw dimension scores are combined using the weights above to produce a composite probability score between 0 and 100. That composite is then calibrated against historical conversion data to ensure the scores are meaningful — that an allocator scored at 70 actually converts at roughly the rate you would expect from a 70-score opportunity.
AllocatorBase uses three threshold bands to translate scores into operational categories:
These thresholds are not arbitrary. They are calibrated against conversion data from alternative asset fundraises across multiple fund strategies and AUM ranges. The thresholds can be adjusted for your specific fund — a niche strategy with a smaller addressable market may need different cutoffs than a broad-mandate fund.
Calibration is an ongoing process. AllocatorBase recalibrates score thresholds quarterly as new conversion data becomes available. If your fund closes allocators that were scored as "medium" at a rate that looks more like "high," the model adjusts. The goal is predictive accuracy, not static labels.
5. How Scores Flow Into Your CRM
Probability scores are only useful if they show up where your team actually works. AllocatorBase writes scores directly into your CRM as custom properties — no separate dashboard to check, no export-import cycle.
HubSpot implementation. Scores appear as custom contact and company properties: ab_probability_score (composite 0–100), ab_probability_band (High / Medium / Low), and individual dimension scores. These properties are available in list filters, workflow triggers, and dashboard reports. A common automation: when an allocator's score crosses from Medium to High, enroll them in a priority outreach sequence and notify the assigned rep.
Salesforce implementation. The same properties are written as custom fields on the Account and Contact objects. Salesforce users can build reports, dashboards, and Process Builder automations using score data. The integration supports both Salesforce Classic and Lightning.
Views and automation triggers. Most teams create three saved views in their CRM: a "High Priority" view filtered to scores 65+, a "Rising" view filtered to allocators whose score increased by 15+ points in the past 30 days, and a "Stalled" view for allocators whose score has declined. These views become the daily operating system for the distribution team.
Score updates happen automatically as new data enters the system. When an allocator opens an email, attends a meeting, or files a new ADV, the relevant dimension score updates and the composite recalculates. Your team sees the current score every time they open a contact record — no manual refresh, no stale data.
6. Worked Example: Scoring a Single Allocator
Consider a hypothetical allocator: Meridian Wealth Partners, an RIA managing $2.4B with a stated mandate for alternative strategies including real estate, private credit, and hedge funds. Your fund is a $250M private credit vehicle targeting 8–12% net returns.
Meridian scores 75.0 — solidly in the High band. The mandate alignment is strong: private credit is within their stated allocation targets, and the fund size is appropriate for their AUM. Engagement is active — two meetings in the past 45 days and a DDQ request. AUM is growing, and they deployed into two alternative funds in the past 12 months. The decision timeline is moderate — they have indicated interest but no formal IC date. Organizational fit is strong — the firm has a streamlined three-person investment committee with prior alternative allocation experience.
The operational implication: Meridian should be assigned to a senior rep, receive personalized follow-up within 48 hours of any engagement signal, and be prioritized for the next available meeting slot. If the Decision Timeline dimension improves (e.g., they set an IC date), the composite score would rise further.
7. Deployment and Monitoring
Deploying probability scoring is not a one-time event. The system requires ongoing monitoring to ensure scores remain predictive as market conditions, allocator behavior, and your fund's characteristics evolve.
Monthly score reviews. Review the distribution of scores across your pipeline monthly. If 80% of your pipeline is scored "Low," either your targeting is off or your thresholds need adjustment. If 80% is scored "High," the model may not be discriminating enough.
Quarterly calibration. Compare predicted conversion rates against actual outcomes. If allocators scored 65+ are converting at 8% instead of the expected 15–25%, investigate which dimensions are overweighted. Adjust weights and recalibrate.
Score change alerts. Configure CRM alerts for significant score movements. When an allocator's score increases by 20+ points in a single update cycle, it typically signals a material change — a new mandate, a leadership change, or a competitive fund closing. These moments are high-leverage for outreach. When a score drops by 20+ points, investigate: budget cuts, strategy shifts, or a competitor winning the allocation.
8. Advanced Techniques
Segment-specific models. Build separate scoring models for each allocator type. RIAs, family offices, endowments, and pensions have fundamentally different decision-making processes, timelines, and conversion patterns. A single global model averages across these differences and loses predictive power. Segment-specific models typically outperform global models by 10–15% on conversion prediction accuracy.
Time-decay weighting. Recent engagement signals should carry more weight than older ones. An email opened yesterday is more predictive than a meeting held six months ago. AllocatorBase applies exponential decay to engagement signals, with a half-life of 30 days. This means a signal from 30 days ago carries half the weight of a signal from today.
Ensemble scoring. For teams with sufficient historical conversion data (200+ closed allocators), combining multiple scoring approaches — logistic regression, gradient boosting, and rule-based heuristics — into a weighted ensemble often produces more stable and accurate predictions than any single method. AllocatorBase's scoring engine supports ensemble configurations for clients with the data depth to support them.