What you love

Relentless Problem Decomposition. Every question that lands on your desk will be ambiguous. “Why is Burrow behind plan?” is not a question you can answer with one query. It’s a question that decomposes into traffic vs. conversion vs. AOV, then further into channel mix, campaign-level performance, inventory availability, pricing changes, competitive shifts, and macro factors. Your job is to decompose faster and more completely than anyone else in the room, and then reassemble the pieces into a clear story with a recommended action.

Cross-Functional Truth-Seeking. You will be the person in the room who asks the uncomfortable question. Marketing presents a ROAS number and you ask whether it accounts for the halo effect on brand search and whether we’re double-counting. Finance presents a forecast — you ask what conversion rate assumption it’s built on and whether that assumption still holds given what you’re seeing in real-time data. Merchandising wants to expand a collection and you ask what the incremental contribution margin is after accounting for cannibalization of existing SKUs. You are not adversarial. You are rigorous. There’s a difference, and you know how to walk that line.

Deep, Original Analysis. The work that matters most in this role is the analysis nobody asked for. You notice that the relationship between Meta spend and revenue has changed shape over the last 90 days and you dig in. You see that one brand’s repeat purchase rate is quietly climbing while another’s is flat and you want to understand why. You realize that our pricing strategy is optimized for gross margin but might be costing us market share, and you build the model to prove it. This is the work that changes how the company operates — and it only happens if you’re obsessively curious about the business, not just responsive to assignments.

AI as a force multiplier. You’ll use Claude Code to analyze rich data sets, build automations, generate first-draft analyses, and move through data at a pace that would have required a team of three five years ago. But you understand that AI output is a starting point, not an answer. You read what Claude produces and immediately see what it assumed, what it skipped, and what it got wrong. You iterate, push back, and refine. The person who runs one prompt and presents the output is not who we’re looking for. You’re the person who runs five follow-ups, questions a number that looks too clean, and cross-references a second data source before forming a view.

What You’ll Actually Do

• Own the analytical foundation the CEO and leadership team use to make decisions. This includes daily revenue and marketing performance analysis, but extends far beyond it — pricing impact studies, competitive analysis, customer cohort behavior, channel attribution deep dives, and portfolio-level pattern recognition.
• Sit in cross-functional meetings not as a note-taker but as the person who pressure-tests what’s being said. When someone presents a conclusion, you’re the one who asks what the data actually shows — and whether the conclusion survives contact with a different cut of the numbers.
• Build and improve automated systems that surface signals across all six brands. Revenue vs. forecast, ROAS by channel, inventory alerts, pricing change impacts. You’ll use AI tools to build these faster, and your own judgment to decide what deserves attention and what’s noise.
• Conduct deep investigations that start with a hunch and end with a recommendation. “I noticed X, I dug into Y, here’s what I found, here’s what I think we should do.” This is the highest-value work in the role and the thing that will earn you a seat at the table.
• Analyze increasingly rich and complex data sources. Work with Rockerbox attribution data, Google Ads and Meta platform data, and internal operational datasets. You’ll often need to join datasets that have never been joined before to answer a question that’s never been asked.
• Produce clear, opinionated written analysis — not slide decks with charts, but concise narratives that say what’s happening, why it matters, and what to do. Our CEO reads these directly, so precision and clarity matter more than polish.

Who You Are

• You are constitutionally incapable of seeing a number that doesn’t make sense and moving on. You have to know why. This is the single most important trait for this role.
• You decompose problems instinctively. When someone says “revenue is down,” your brain immediately starts breaking it into components — traffic, conversion, AOV, channel mix, new vs. returning — before anyone finishes the sentence.
• You are genuinely, demonstrably obsessed with AI tools. You’ve used Claude Code, Cursor, or similar tools to build something real.
• You are comfortable being the person who says “I don’t think that’s right” to a VP or a CMO — and you can do it in a way that’s rigorous, not combative. You bring the data. You show your work. You let the analysis do the talking.
• You have strong SQL skills and have worked directly in a data warehouse — not just queried a BI tool. You understand data structures, joins, window functions, and how to structure a query that answers a business question, not just returns rows.
• You think in systems, not snapshots. You don’t just look at yesterday’s ROAS — you think about how ROAS connects to attribution methodology, which connects to budget allocation, which connects to customer acquisition cost, which connects to long-term brand health. You see the chain.
• You have 2–4 years of experience in analytics, strategy, consulting, investment banking, or a similarly rigorous environment. The exact title and industry matter less than the quality of your thinking and your hunger to understand how things actually work.
• You write clearly. You can take a complex, multi-variable finding and explain it in three sentences that a non-technical executive would act on.

Why Havenly Brands

Our CEO writes SQL, builds automated briefings with Claude, queries our data warehouse directly, and expects the same intensity from everyone around her. This is not a company where analytics is a support function that produces reports no one reads. This is a company where the right analysis, delivered at the right moment, changes what we do tomorrow.

You’ll have direct access to leadership, real problems with real stakes, and an environment where the best idea wins regardless of who’s in the room. You’ll work across six brands with different customers, price points, and dynamics — which means the analytical puzzles never get repetitive.

Havenly is a multi-brand home furnishings company operating six brands: Havenly, Burrow, Interior Define, The Citizenry, St. Frank, and The Inside. We are looking for someone whose default setting is “why?”

This is not a reporting role. We don’t need dashboards maintained or decks formatted. We need someone who wakes up thinking about why one brand’s conversion rate diverged from another’s last Tuesday, who notices that a pricing change three months ago is quietly reshaping our product mix, and who can’t let go of a number that doesn’t make sense until they’ve chased it to the source.

You will work directly with the CEO and cross-functionally with marketing, merchandising, operations, and finance — not to support their requests, but to challenge their assumptions. When the marketing team says Meta is working, you ask what “working” means and whether the attribution model agrees with the P&L. When merchandising says a collection is underperforming, you ask whether it’s a demand problem, a pricing problem, an inventory problem, or a traffic problem — because those are four completely different actions.

You’ll use AI tools — Claude, Claude Code, Cursor — as force multipliers to move through data faster than any traditional analyst could. But the tools are not the point. The point is what you do with what they surface: the second question, the missing variable, the connection between two datasets that nobody else thought to join.