Pricing
Invest in decisions, not dashboards
Choose the plan that matches your location count, channel complexity, and measurement needs. Every tier includes causal attribution and audit-grade traceability.
Starter
$1,200/mo
For single-brand operators running their first controlled pricing tests across up to 25 locations.
- Up to 25 locations
- Single-channel deployment (POS or online)
- Test design advisor with recommended holdouts
- Difference-in-differences causal readouts
- Contribution margin lift reporting
- Basic guardrails (margin floor, anomaly alerts)
- Audit log and change history
- Email support
Most Popular
Growth
$3,500/mo
For multi-location chains that need cross-channel orchestration and advanced causal measurement across up to 150 locations.
- Up to 150 locations
- Multi-channel deployment (POS, delivery, online)
- Synthetic control + DiD with Bayesian intervals
- Profit-weighted readouts with uncertainty ranges
- Advanced guardrails (anomaly pausing, channel consistency)
- One-click rollback across all channels
- Decision logs with Ship / Extend / Stop recommendations
- Evidence trail with cross-functional sharing
- Priority support and onboarding
Enterprise
Custom
For large restaurant groups and prepared-foods retailers with complex channel landscapes and custom integration requirements.
- Unlimited locations
- All Growth features included
- Custom POS and delivery aggregator integrations
- Multi-brand and multi-region support
- Advanced experiment templates and playbooks
- API access for workflow automation
- Dedicated customer success manager
- Custom SLA and security review
- Quarterly business reviews
Compare plans
| Feature | Starter | Growth | Enterprise |
|---|---|---|---|
| Locations | Up to 25 | Up to 150 | Unlimited |
| Deployment channels | Single | Multi-channel | Multi-channel + custom |
| Test design advisor | Yes | Yes | Yes |
| Causal methodology | DiD | DiD + Synthetic control | DiD + Synthetic control |
| Bayesian intervals | No | Yes | Yes |
| One-click rollback | No | Yes | Yes |
| Anomaly auto-pause | No | Yes | Yes |
| Decision recommendations | Basic | Advanced | Advanced |
| Evidence trail | Basic | Full | Full + API |
| Custom integrations | No | No | Yes |
| Support | Priority | Dedicated CSM |
Frequently asked questions
- How long does it take to set up the first experiment?
- Most teams design and deploy their first controlled test within one week. The test design advisor recommends holdouts, duration, and guardrails based on your location count and traffic patterns, so you spend less time on configuration and more on measurement.
- Do we need a data science team to use MarginSignal?
- No. The platform handles experiment design, causal methodology selection, and result interpretation automatically. Readouts surface contribution margin lift with credible intervals in business-ready terms that Revenue, Ops, and Finance can review without statistical expertise.
- What POS and delivery systems do you integrate with?
- Growth plans support the most common POS, delivery aggregator, and online ordering platforms. Enterprise plans include custom integration work for proprietary or legacy systems. Contact us for your specific channel landscape.
- What happens if an experiment causes unexpected margin erosion?
- Operational guardrails monitor every running test. If a guardrail threshold is breached — such as a margin floor or anomaly detection — the system can pause the experiment automatically and notify your team with the impact and recommended next step. One-click rollback reverts changes across all channels.
- Can we run experiments across multiple brands or regions?
- Yes, on the Enterprise plan. Multi-brand and multi-region support lets you manage separate experiment portfolios with shared learnings and distinct channel configurations.
- How is contribution margin calculated?
- MarginSignal uses the contribution margin definition your business provides (revenue minus variable costs). You configure the cost inputs during onboarding, and all readouts, guardrails, and decision recommendations reference that consistent definition.