Target Data Analyst Interview Questions

“At a previous job I studied when online orders arrived at our locations. Staff for packing were scheduled evenly through the day, but orders came in two waves, late morning and early evening. Before presenting, I showed the pattern to two site leaders and asked what I was missing. They told me evening staff also handled returns, which my data did not show. I added that and proposed moving some morning hours to the evening. We tried it at a few sites first. The result was your result. Asking first made the leaders partners in the change.”

“Our dashboard showed that almost every order was packed on time. Staff told me they were always behind. I went to watch a shift. Orders were marked as packed when the first item was scanned, not the last, so the clock stopped too early. The number was true to its definition and useless in practice. I changed the measure to use the final scan and showed the old and new versions side by side, so nobody felt tricked by the drop. With the honest number, we saw your result. I now check how each event is recorded.”

“First I would reshape the events into one row per order, with the arrival time and the handover time as columns, using MAX with a CASE for each step and GROUP BY order. The wait is handover minus arrival. Some orders have no arrival event because the guest never checked in, and some have a handover before arrival, so I would set those aside and count them. Then I group by store and by the hour of arrival. I would report the median and a high percentile, not the average, because a few very long waits would hide the usual experience.”

“Shipping from a store should get the parcel to the guest sooner because the store is nearby. So I would measure time from order to delivery, cost per parcel including store labour, and how often the store could not find an item it had accepted. I would also watch the store itself: did shipping empty shelves that walk-in guests needed? For a fair view, I would compare similar orders, the same kind of item going a similar distance, because warehouses and stores handle different mixes. One number will not settle it, so I would show the measures together.”

“Launch week sales mostly measure advertising and shelf position. To see whether guests like the products, I would look at repeat purchase: of the guests who bought once, what share bought from the line again within a sensible period for that kind of product. I would track the return rate and the reasons given, and the tone of reviews. I would watch what happens to weekly sales after the launch displays come down. I would also check whether buyers of one item go on to try others in the line. Comparing with earlier launches shows what normal looks like.”

“I would begin with a simple baseline: the average for each day of the week and hour over recent weeks, because order patterns repeat weekly. Then I would add what we know ahead of time, such as a promotion, a holiday or the start of the school term. I would test the method on past weeks it has not seen and measure how far off it was. For the store leader, I would give a likely number and a high case for each day, since staffing for the high case on the busiest days is cheaper than keeping guests waiting.”

“I like data where every row is a small event in someone's day: a guest parked, a bag handed over. At Target the store is a shop and also the place orders are packed and collected, so the same team and shelves serve several purposes. That creates real trade-offs, and good analysis can ease them. In my past work on your project, the most satisfying moment was a schedule changing because of what I found. I want to learn how store leaders use data during a shift, and build reports that fit that moment.”

“I would first ask whether anything changed in how late is recorded, such as a new cutoff hour or a delayed feed from the carrier. If the measure is sound, I would split each parcel's journey into two parts: order to packed, which belongs to the store, and packed to collected, which belongs to the carrier. Then I would break the rise down by store, region, carrier and day. A rise in one region on two days points to something like weather or a missed collection. Once I have a likely cause, I would call a few stores to confirm before I report it.”