How Jones Road Beauty became an AI-native company on a data warehouse they own
connected
per month
At A Glance
We have one data analyst and in the past it used to take two weeks to get a report. We have our entire data warehouse. That was one of our Q1 projects, that we can access the warehouse directly from Claude, so you can get anything in a heartbeat and you don’t have to be technical at all.
The challenge: one analyst and a company impatient to use AI
Jones Road Beauty moves fast, with a constant stream of product launches and pricing tests across its DTC site, TikTok Shop and owned retail. But its analytics ran through one person. Ben Rosenwald, Director of Data & AI Initiatives, is the only technical data hire in-house. “I’m the only data person internally,” he says. Every report, and every new cut of the data, ultimately landed on his desk.
The previous setup made the bottleneck worse. On Daasity, the data was managed for them on a warehouse they didn’t fully control. Routine changes meant filing a request and waiting. As Ben recalls, “even something as easy as adding a colleague, I’d have to go through them.” Pulling a more nuanced metric or wiring up a custom source was the same story.
Meanwhile, the company had bet hard on AI. As Ben put it, Claude had rolled out so that “everyone in the org has a Claude login,” and leadership was impatient to point it at their numbers. “Cody, our CEO, and others are chomping at the bit, like, when can I start asking Claude about data?”
The ambition was there. The data simply wasn’t in a form Claude could reach safely.
Why Polar: data they own, governed for AI
Jones Road came to Polar while actively planning their exit from Daasity. The pitch matched the goal exactly: keep the ownership and control of a real warehouse, but make the data instantly usable by AI. Polar stood up a dedicated, fully isolated Snowflake warehouse the team owns, fed by 40+ out-of-the-box connectors plus the long tail of custom sources Jones Road relies on. On top of the raw data sits Polar’s Synthesizer semantic layer, a governed commerce ontology that defines metrics like blended ROAS, true CAC, and contribution margin once, so they stay consistent everywhere.
That governance is what makes the final step trustworthy. Through the Polar MCP, the same governed data connects directly into Claude, so every answer runs against agreed definitions instead of rogue, on-the-fly SQL. The team came to describe Polar as “an open API to feed Claude with anything that’s connected.”
Ben framed the whole project simply. The core goal was “making our data accessible in Claude.”
Onboarding in days, not months
Warehouse migrations are supposed to be painful. This one wasn’t. Jones Road connected its core marketing sources and was exploring data almost immediately. “Appreciate the speed guys… LFG,” the team posted on day one. Polar publicly describes the cutover as a nine-figure brand that transitioned in 72 hours and is now running production agents on the semantic layer.
In action: from same-cart questions to in-store CVR agents
The first business question set the tone. When Jones Road relaunched its hero product, the Mini Miracle Balm, as an everyday item, the team needed to know how customers were actually buying it. Were they purchasing one unit or several? Were they only adding a second to clear the free-shipping threshold the team had just lowered from $85 to $55? Did customers acquired on the mini retain differently than those on the full size? Polar’s team delivered a same-cart and launch-performance analysis, and Jones Road started asking the follow-ups itself, in plain English.
Today most of the team, not just the data function, turns to Polar to answer its own questions, whether in Claude, in Ask Polar, or through a Slack databot. The kinds of things Jones Road now uses Polar to figure out span the whole business:
The connector list keeps growing to feed it. Alongside Shopify, Meta, Google, TikTok, Pinterest, GA4, Klaviyo and the Polar Pixel, the team has wired in sources like AppLovin, Postscript, RetailNext, Northbeam, KnoCommerce, Junip, heatmap.com and NetSuite. Rollout is moving from the data team outward, with team-by-team training that began in marketing, turning Polar into the single governed source every Jones Road agent reads from.
The results: a faster company, not just faster reports
The reporting bottleneck is gone. Work that used to take “a week, two weeks” now comes back in seconds, and the people asking don’t need to be technical. From the CEO to the growth and demand-planning teams, everyone queries the same governed source of truth, so the numbers stay consistent even as more people pull them. Adoption reflects it. Most of the team is active in Polar every month, the analyst is no longer the bottleneck, and the company runs well over a hundred AI-driven analyses through the platform in a typical month.
For Cody Plofker, accessing the warehouse through Claude was “step one.” With reporting effectively instant, the team is now pushing into the execution layer on top of it, chasing faster launches and the standard he describes as a 24-hour full-funnel launch. The throughline is the one Polar is built for. Give AI clean, governed data, and a lean team can move like a much bigger one.
The fastest growing Shopify brands use Polar Analytics
