Otto's autonomous stock ordering: a deep-learning system that buys inventory on its own

The problem

Otto, one of Germany’s largest online retailers, sells goods from many third-party brands. Forecasting demand and ordering the right stock at the right time is the core operational problem: order too much and surplus and returns pile up, too little and items sell out.

What was built

Otto deployed a deep-learning system from Blue Yonder, a startup in which it holds a stake, whose algorithm “was originally designed for particle-physics experiments at the CERN laboratory in Geneva”. Per The Economist, the system “analyses around 3bn past transactions and 200 variables (such as past sales, searches on Otto’s site and weather information) to predict what customers will buy a week before they order” (source). The algorithm was developed by physicist Michael Feindt while at CERN (source).

The outcome

Accuracy and autonomous ordering. The Economist reported the system “predicts with 90% accuracy what will be sold within 30 days” and that Otto “allows it automatically to purchase around 200,000 items a month from third-party brands with no human intervention” (source). Otto’s director of category support, Michael Sinn, defined the accuracy metric first-hand: “We consider it accurate when we sell out of items ordered from our retail partners within 30 days. With automated replenishment decisions from Blue Yonder, we achieve this 90 per cent of the time” (source).

Stock and returns. “Overall, the surplus stock that Otto must hold has declined by a fifth. The new AI system has reduced product returns by more than 2m items a year” (source). An HBS write-up relayed the same, that the application let Otto “predict with 90% accuracy on any given day what items will be sold over the next 30 days and reduce inventory levels by 20%” (source).

No job cuts. The Economist was explicit: “Otto did not fire anyone as a result of its new algorithmic approach: it hired more, instead. In many cases AI will not affect a firm’s overall headcount, but will perform tasks at a level of productivity that people could not achieve” (source).

Earlier gains and delivery times. An earlier Blue Yonder deployment achieved a 40% improvement in demand forecasting (source), and Otto later reduced delivery times of third-party items “from up to seven days down to one or two days, without risking overstocking” (source). The relationship between the earlier 40% figure and the later 90% accuracy metric is not reconciled in the public record.

How this was verified

This case carries a green badge under TIN’s current standard: independently validated by TIN against the public record. The headline 2017 figures rest on a single top-tier independent article, The Economist, sourced from Otto and Blue Yonder, with a named Otto executive and an HBS write-up echoing them. Read the badge as TIN validated this account is accurate, including the honest part: the figures are roughly 2017-vintage and unaudited. Green certifies they are reported exactly and in context, not that they were independently re-measured. Direct client confirmation is the higher bar TIN grows into as the platform matures.

Sources

Cited in this case file. Tier 2 = independent press; Tier 3 = first-party or trade material. Each quote above was checked against an archived capture of the source.

1. The Economist, “How Germany’s Otto uses artificial intelligence,” 2017-04-12 (Tier 2). https://www.economist.com/business/2017/04/12/how-germanys-otto-uses-artificial-intelligence
2. Retail Systems, “OTTO / Blue Yonder AI delivery times” (Michael Sinn quoted), 2018 (Tier 3, first-party executive). https://www.retail-systems.com/rs/OTTO_Blue_Yonder_AI_Delivery_Times.php
3. Harvard Business School (RCTOM), “Autonomous stock replenishment at online retailer OTTO,” 2018-11-15 (Tier 3). https://d3.harvard.edu/platform-rctom/submission/autonomous-stock-replenishment-at-online-retailer-otto/
4. Computer Weekly, “German retailer Otto invests in neural software to net future sales,” 2013-11 (Tier 2). https://www.computerweekly.com/news/2240170763/German-retailer-Otto-invests-in-neural-software-to-net-future-sales