Demand Planner Interview Questions
& Model Answers, 2026

The answer should match method to pattern: simple exponential smoothing for stable demand without trend, Holt's method where a trend exists, Holt-Winters where trend and seasonality both do, and Croston's method or a similar intermittent-demand approach for sporadic items where standard smoothing produces a misleadingly small non-zero forecast every period. A planner who applies one model across the portfolio is guaranteeing poor results on the tails.

Good answers separate true demand from what was shipped: adding back lost sales during stockouts, removing or flagging one-off events such as a single large tender or a recall, treating promotional periods as an identified uplift rather than baseline demand, and correcting for a lost or gained customer that changes the level permanently. Fitting a model to uncleansed shipment history is the most common cause of a structurally wrong forecast.

Expect analogue-based reasoning: pick comparable products by category, price point and channel, adjust for distribution build and launch support, express the forecast as a range rather than a point, and set an early review cadence to correct within the first weeks of real data. Strong candidates also cover the supply consequence — how much launch inventory the range justifies, and what the downside costs if the launch underperforms.

The method should be modelled rather than guessed: build a baseline excluding promoted periods, quantify the uplift from previous events of the same mechanic and depth, adjust for cannibalisation of related products and for the forward-buying dip that follows, and store the promotion calendar as an input so the model learns. Planners who fold uplift into the baseline permanently inflate the forecast after the promotion ends.

Because a forecast with random error around the truth is absorbed by safety stock, while a persistently biased forecast either builds excess every period or generates chronic stockouts, and no amount of safety stock fixes it. Strong answers track bias by product segment and by contributor, since bias usually has a source — a sales team incentivised to under-commit, or a marketing plan consistently over-optimistic — and naming that source is the fix.

The good version starts from the statistical baseline, requires commercial input to be specific and evidenced — a named customer's listing, a distribution gain, a competitor exit — rather than a percentage uplift, records every override with its owner and reason, and reviews the accuracy of previous overrides so the process learns. Meetings that simply average opinions produce a number nobody owns.

Interviewers want the post-mortem: whether the miss came from the model, the history, an override, or an event nobody could have known, and what changed afterwards. A planner who traces a miss to an override they accepted under pressure, and describes how they now handle that conversation, is demonstrating exactly the judgement the role needs.

Strong answers ask for the evidence behind the number and offer to reflect anything specific and verifiable, while declining a blanket uplift. They also show the consequence being made visible — the inventory or capacity commitment the higher number triggers — so the change becomes a business decision rather than a planning courtesy.

Concrete interventions land best: resegmenting the portfolio and applying appropriate models, cleansing history properly, introducing an override discipline, or fixing a hierarchy so forecasts were made at the right level. The answer should quantify the before and after, and name what did not work, because most accuracy programmes include at least one intervention that made no difference.

Look for someone who communicates uncertainty rather than a single number: flagging which items are volatile, sharing the range and the key risks, and giving early warning of a step change rather than letting it appear in the next cycle. Demand planners who treat their output as final and immovable make supply planning's job impossible.

Anchor on published data with its caveat. The BLS OEWS May 2025 national median for logisticians is $82,320 a year ($39.58/hr), with the top 10% above $133,160, and this series bundles several analytical logistics and planning roles rather than demand planning specifically. Then place yourself on portfolio size, planning system depth, sector — consumer goods with heavy promotional activity is more demanding than stable industrial demand — and statistical capability.

Ask carefully, because accuracy targets can be gamed and are heavily influenced by demand volatility outside the planner's control. If accuracy is in the bonus, ask which metric, at what level, and whether bias is measured alongside it, since accuracy alone can be improved by forecasting conservatively. A well-designed target pairs accuracy with bias and is set by segment.

Evidence the transferable core — statistical capability, data handling, and any forecasting you already do — then be direct about the gap and propose a review point tied to demonstrated ownership of a segment. Also negotiate the tools and access, because a demand planner without proper system access and training spends their first year rebuilding the process in spreadsheets.