Open the seller health dashboard in Looker and the first thing you see is a long table, one row per seller, sorted by gross orders. Across the top run the columns that matter: cancellation rate for the trailing 30 days, share of orders delivered late, and a small sparkline of the average rating over the past quarter. Most rows sit in neutral grey. This week one of them, a mid-sized home goods seller, has turned amber, because its late deliveries crossed the threshold someone set months ago. Explaining why that cell changed colour, and deciding what happens next, is the core of remote marketplace operations analyst jobs. Expect a full-time week from your own home, in whatever country you live, at $3,750 USD per month.
Read the rest as a marketplace operations analyst job description told from the dashboard down: the scorecard first, then the queries beneath it, then the decisions it feeds.
A seller scorecard is a compact answer to one question: is this seller giving buyers the experience the marketplace promises?
Seller-initiated cancellations divided by orders placed in the window. A rising figure usually means stock counts are wrong or the seller is listing items they cannot source. Buyer cancellations belong in a separate field.
Orders that reached the buyer after the promised date, as a share of delivered orders. This column often points to a carrier problem or a warehouse change rather than a careless seller, which is why you never read it alone.
Not the lifetime average, which barely moves, but the direction over recent weeks. A seller with a strong history and a falling trend deserves a closer look before the average catches up.
Supporting fields sit alongside: order volume, tenure, category and region. Two late orders out of five looks alarming and means almost nothing, so the scorecard should mark small samples as small.
Every tile on that dashboard starts as a query, and in this role you write most of them. A typical late delivery query joins the orders table to shipment events, filters to delivered orders inside the window, compares the delivery timestamp to the promised date, and groups by seller. The hard part is agreeing on definitions: does an order count as late if the carrier scanned it on time but the buyer marked it missing? Which time zone sets the cutoff?
Expect to use window functions for rolling 30-day rates, CTEs to keep long logic readable, and careful handling of NULLs where shipment data arrives late. You will also write checks that run before the numbers reach anyone, such as row counts against yesterday and sellers that vanished from the feed. Catching a broken feed before a seller gets an unfair warning is the most useful thing you will do some weeks.
In Looker, you define measures and dimensions once in LookML so that "cancellation rate" means the same thing in every explore, then build the operational dashboards that the seller support and trust teams check daily. Conditional formatting drives the amber and red cells, and drill-downs let someone click from a seller row to the individual orders behind the number.
Tableau tends to carry the wider analysis: category comparisons, regional maps of late deliveries, cohort charts showing how new sellers perform in their first months. Keep the operational view sparse: people acting on a seller need five numbers and a clear flag, not forty filters.
A threshold turns a cell amber, but you decide whether that amber row becomes a case. Before raising one, check the obvious alternatives. Did a carrier outage hit the whole region? Did the seller just move into a slow-shipping category? Is the volume high enough for the rate to mean something?
When the pattern holds, write the flag so another person can verify it without asking you anything: the metric, the window, the comparison group, the order count, and a link to the underlying query or Looker view. That record matters later if the seller disputes the outcome.
Most marketplaces would rather fix a seller than remove one. Your analysis supports the policy team as they choose between a warning, a time-bound improvement plan, listing limits or, in persistent cases, suspension. The analyst's job is to make those decisions fair: the same rule applied the same way to a large seller and a small one, with the evidence attached.
For an improvement plan, you set measurable targets drawn from the scorecard, agree on the review date, and build a simple tracking view so everyone sees progress. Then you check whether plans work at all; if sellers on plans recover no faster than sellers left alone, the design needs rethinking.
You would earn $3,750 USD every month for full-time hours, all worked from home. As with most marketplace operations analyst jobs remote worldwide, the team may span several time zones, so expect a few overlapping hours for reviews. Any marketplace operations analyst salary elsewhere depends on region and employer, so read this figure as specific to this role.
Desks for marketplace operations analyst work from home jobs are simple, though they must stay dependable. You need a stable broadband connection, a computer that handles a browser with several heavy dashboards open alongside a SQL client, and ideally a second monitor so the query and the chart sit side by side. Since you will handle seller and order data, follow the employer's rules on VPN use, screen locking and where files may be saved.
Employers generally want a bachelor's degree plus one to three years in marketplace, e-commerce or business operations analytics. The core marketplace operations analyst skills are strong SQL, working fluency in Looker or Tableau, and clear writing, since much of your influence comes through short notes explaining what a number means.
Becoming a marketplace operations analyst usually means arriving from e-commerce operations, seller support, supply chain reporting or general business intelligence work. Build a small portfolio: a public dataset of orders, a set of queries computing seller-level rates, and a dashboard with sensible thresholds. A SQL course or a Tableau or Looker credential shows initiative, but a portfolio that explains its definitions usually carries more weight. Entry level candidates with no analytics background will find this role looks for at least a year of related work.
Etsy, eBay and Uber are useful examples of marketplace businesses where seller or driver performance gets measured this way; reading their published seller standards is good preparation.
A business analyst usually works across functions, gathering requirements, mapping processes and supporting projects that may have nothing to do with sellers. A marketplace operations analyst sits closer to the daily running of a two-sided platform. Your outputs feed decisions about specific sellers within days, your metrics are tied to buyer experience, and your SQL runs against order and shipment tables every morning. Because the toolkits are so similar, people move between the two in both directions.
After a couple of years, analysts in this area often take on ownership of the whole seller quality framework: choosing which metrics count, setting thresholds and testing whether a policy change helped. From there, paths lead towards senior or lead analyst roles, marketplace strategy, trust and safety analytics, or operations management. Some drift into analytics engineering because they liked LookML more than policy debates. Nothing guarantees those steps, but the habits this role builds carry into each of them: precise definitions and evidence attached to every claim.