Precision Causal Modeling — BCT Analytic Solutions

The core idea

Causal, not just correlational. Precise, not average.

Most analytics tell you what happened, on average, in the past. PCM identifies the interventions and conditions that actually cause better outcomes for specific, meaningful subgroups, using the administrative data organizations already collect.

🧩

Meaningful subgroups

PCM surfaces distinct subgroups within a population, not one blended average, so services can be matched to who actually benefits.

🔗

Causal drivers

It isolates the service features and conditions associated with better outcomes, separating signal from noise.

🧑‍🔬

Human-guided ML

Domain experts and people with lived experience guide the modeling, keeping results interpretable, responsible, and grounded in practice.

♻️

Uses existing data

No costly new data collection. PCM works with the administrative, service, and assessment data agencies already hold.

🎯

Actionable evidence

Findings are translated into tools and guidance leaders and practitioners can use in real decisions, not just reports.

⚡

Faster and cheaper

On one federal engagement, PCM-style methods produced findings in ~15% of the time required for manual analysis.

Proven in the field

Not a lab experiment, a working capability.

PCM powers BCT's proprietary platforms, including the Community Impact Compass and Outcomes Generator, and has been applied across child welfare, behavioral health, education, housing and homelessness, philanthropy, and economic opportunity.

Identified eight matched subgroups of children and the treatment features linked to post-discharge outcomes in a residential program
Investigated risk and protective factors for elder and adult maltreatment for a federal agency
Extracted insight from vast research corpora in a fraction of the time of manual coding

Read a case study

What works, for whom, and under what conditions.