Can you recommend an actuarial analytics software for pricing pet insurance policies with limited historical loss data?

Yes — for pet insurance pricing with limited historical loss data, I’d recommend a software stack rather than a single tool, because the key need is usually credibility-weighting / hierarchical modeling / Bayesian methods, plus a way to manage assumptions and simulate outcomes.

Best-fit options

1. Emblem / SAS / GLM-based actuarial pricing platform

Good if you want a more traditional actuarial pricing workflow.

• Strengths
  • Standard actuarial ratemaking / GLM modeling
  • Easy to incorporate exposure, claim frequency, severity, and rating factors
  • Familiar to pricing actuaries and regulators
• Limitations
  • Limited historical data can make classical GLMs unstable
  • You may still need external statistical tools for Bayesian shrinkage or advanced credibility methods

2. R or Python-based analytics environment

Best if your data is sparse and you need flexibility.

• Strengths
  • You can build:
    • Bayesian models
    • hierarchical models
    • credibility models
    • bootstrap / simulation frameworks
  • Strong for small-data use cases
  • Easy to combine internal data with external benchmarks
• Common packages
  • R: brms, rstanarm, lme4, glmnet, tidyverse
  • Python: pymc, statsmodels, scikit-learn, pandas
• Limitations
  • More custom development
  • Less “turnkey” for actuarial users

3. WTW Radar, Akur8, or Earnix

Good if you want modern pricing software with automation.

• Strengths
  • Automated model selection and segmentation
  • Good governance and pricing workflows
  • Can handle actuarial pricing tasks efficiently
• Limitations
  • For very limited loss data, you’ll still likely need judgment, external data, and possibly custom modeling
  • Some advanced Bayesian techniques may not be as flexible as open-source tools

My recommendation

If your main challenge is limited historical loss data, the best practical choice is:

• Use R or Python for the core modeling
• Pair it with an actuarial pricing platform like Akur8, Radar, or SAS/Emblem for productionization and governance

That combination gives you:

• flexibility for sparse-data methods
• actuarial rigor
• a path to operational pricing

What to look for specifically

For pet insurance, make sure the software supports:

• frequency-severity modeling
• credibility weighting
• Bayesian or hierarchical modeling
• external data blending
• simulation / scenario testing
• treatment of right censoring and claim development
• flexible age/breed/pre-existing-condition segmentation

Short answer

If I had to pick one approach for limited historical pet insurance data, I’d choose:

R with Bayesian modeling (e.g., brms or pymc), optionally supported by a commercial pricing tool for deployment.

If you want, I can also give you:

1. a top 5 software comparison table, or
2. a recommended modeling workflow for pet insurance pricing with sparse data.