Refyner was built by people who spent years doing this work by hand – and got tired of waiting on engineering to ship a number the business needed yesterday.
AJ is a data and analytics leader with over 10 years of experience across Unilever, Deloitte and Tapi – most recently as Data & Analytics Lead, building the data functions and models that consumer and subscription businesses run on.
He built Refyner after watching the same problem play out brand after brand: analysts who knew exactly what they needed, stuck waiting on engineering to ship it. Refyner is the tool he wished his teams had had.
We're the team behind CXDATA, an analytics partner to eCommerce and subscription brands. Across dozens of live brands – from supplements to pet food to skincare – we kept solving the same problem in the same way.
A brand needed its real numbers: lifetime value by cohort, repurchase curves, MRR movement, churn, channel attribution. And every single time, an analyst would sit waiting – for a data engineer to wire up a pipeline, for an orchestrator to be configured, for someone with the right access to push it to production. The analyst knew exactly what they wanted. They just couldn't ship it themselves.
Meanwhile the models – LTV, RFM, retention, subscription economics – got rebuilt from scratch, brand after brand, as if no one had ever modelled an eCommerce business before.
Refyner is what happens when analysts build their own tooling. It does the two things that were always split across two tools and a queue – prep your data and schedule it – in one place, on the Snowflake a brand already has. And it ships with the eCommerce and subscription models already built, because we'd built them a hundred times before.
The result: an analyst connects their warehouse and ships a production dataflow themselves, with the right models populated, without waiting on an engineer. The data team is freed for the genuinely hard problems instead of being a ticket queue for routine work.
Everyday data prep shouldn't require an engineer in the loop. The person who knows what they need should be able to ship it.
The models an eCommerce or subscription brand runs on are well understood. They should come ready, not from a blank canvas.
Self-serve only works when someone has your back. We're an agency at heart – real people who know your stack, not a ticket queue.
Two brands, one team. They sit at different points on the same value chain, and they're built for different people.
A managed analytics partner for brands that want their data function run for them – strategy, models, and the judgement of what to do next. No analyst required on your side.
The product for brands with their own analysts who want to move fast and self-serve – the same models and rigour, in their hands, on their Snowflake.
Refyner is a CXDATA product. The same institutional knowledge that powers our managed service is what's built into the tool – now available to teams who want to run it themselves.
Connect your Snowflake and prep, schedule, and ship a dataflow yourself – no waiting on engineering.
Refyner