Churn model for a B2B payments platform
Built a gradient-boosted churn model that identified 80% of likely churners 60 days out. Integrated into the CS team's workflow.
Data analytics & research
We help finance, healthcare, and technology teams cut through the noise. From dashboards to predictive models, we deliver analytics that executives actually use.
Working with teams at
What we do
We work as an embedded analytics partner — not a vendor. Every engagement ships with measurement, training, and a clear path forward.
Self-serve dashboards in Looker, Tableau, and Power BI. We model the data so your team can answer their own questions.
Churn scoring, demand forecasting, lead prioritization. Models that integrate with the systems your operators already use.
A/B test design, sequential testing, and Bayesian inference. Statistical rigour that holds up to board-level scrutiny.
ETL pipelines, dbt models, warehouse architecture. Reliable data infrastructure that scales without a dedicated ops team.
Quantitative and qualitative consumer studies. Sample design, instrument validation, and reporting your CMO will sign off on.
GDPR/CCPA compliance audits, data lineage, role-based access. Compliance as a byproduct of good architecture.
By the numbers
Founded in 2019. We work with 50+ clients across finance, healthcare, and technology.
How we work
No long discovery decks. We start with a one-week diagnostic and only proceed if the value is real.
One week. We map your data sources, define 3-5 KPIs, and write a one-page brief.
Two weeks. We build the analytics architecture, get sign-off, and lock the scope.
Four to twelve weeks. Sprint-based delivery with weekly demos and clear ownership.
Handover, training, and an optional retainer. You own everything we build.
Selected work
A few projects we can talk about. Many more we can't — client confidentiality matters.
Built a gradient-boosted churn model that identified 80% of likely churners 60 days out. Integrated into the CS team's workflow.
Time-series forecasting across 14 departments. Cut overtime by 18% in the first quarter post-deployment.
Two-tower embedding model trained on 18 months of session data. Live A/B test beat the control by 22% on revenue per session.
Who we are
No offshore handoffs. Every project is led by a partner with 8+ years of experience.
Founder & Partner
Head of Analytics
Lead Data Engineer
Research Director
Common questions
Most engagements begin with a one-week paid diagnostic. We audit your data, define 3-5 KPIs, and produce a one-page brief. If the value is clear, we scope a build phase. If it isn't, you keep the brief.
Our sweet spot is Series A to mid-market. We have one client with 12 employees and one with 4,000 — the common factor is executive buy-in for data-driven decisions.
Yes. We're tool-agnostic. Most of our clients use a mix of Snowflake or BigQuery, dbt, and either Looker, Tableau, or Power BI. We integrate with what you have.
Diagnostic is a fixed fee. Build phase is priced either as a fixed-scope project or a monthly retainer for ongoing work. We do not charge by the hour.
Our analysts are based in London, Berlin, and Toronto. We work async-first with overlap hours for live sessions.
Get in touch
We respond within one business day. No salespeople, just an analyst on the team.
If you'd rather talk to a person first, give us a call or send an email. We pick up.