Podcast 171: Real Time Is Just a Marketing Concept, with Yuliia Komaryn

“Real time” is a marketing concept, not a technical one. Yuliia Komaryn, Business Analytics Team Lead at DashDevs, opens with a distinction that reshapes the rest of the conversation: real time isn’t a technical requirement, it’s a marketing and stakeholder-facing concept. A fraud alert genuinely needs millisecond response. A monthly profit metric doesn’t. The real design question isn’t “can we make this instant,” it’s “what’s the right time for this specific metric to reach the person who needs to act on it.”

A real BNPL fraud case that proves the point

Yuliia shares a case DashDevs solved directly: a BNPL client whose user base evolved to include people who took credit access with no intention of ever repaying. Without dashboards properly flagging suspicious transaction patterns in time for the compliance and support teams to act, that exposure keeps growing. The dashboard isn’t the deliverable — the reaction it enables is.

Why “customer” doesn’t mean the same thing to every team

One of the most concrete examples in the episode: a marketing team defining “customer” as anyone who clicks an ad, while the finance team only counts someone who has actually paid. Yuliia argues that misaligned field definitions like this, repeated across LTV models, acquisition cost, and every other core metric, cause more damage than any dashboard’s visual design ever could.

That definition problem sits at the center of practical data science in the fintech industry: models inherit whatever the business means by each field, including the contradictions.

Why data lakes matter less than trustworthy data

Yuliia’s view on 2026: the market has shifted from “can you build the dashboard” to “can you trust what’s in it.” AI makes it trivially easy to generate a dashboard that looks polished and complete. It doesn’t make the underlying data any more accurate, traceable, or aligned with how the business actually defines its own terms.

Where AI actually helps analytics, and where vibe coding fails

Yuliia is direct about where AI adds real value for her team: removing routine, repetitive work so analysts can focus on getting the underlying concepts right. She’s equally direct about where it fails: vibe coding a financial system without understanding what’s inside the black box guarantees compliance and business-logic mistakes, because someone still has to be accountable for what the code actually does.