Very solid results from Day 3
What do you like best?
The demand forecast tool was quite easy to implement and we saw uplift in one of the most important KPIs (percentage of orders we can dispatch directly from our warehouse, without backordering items at our suppliers) within few days. As we have hundreds of thousands of live items at our e-shop, it is especially useful for mid- and long-tail items, with estimates of 2-5 pieces sold in next 14-30 days, which usually goes under the radar for the product managers. And we are amazed by precision of the estimations, in most cases far exceeding those of regular human product managers.
What do you dislike?
We started using the product in beta phase, so there were some glitches in forecasts (over-order or under-order). We see the system improving over time, but it still needs some human overlook of the final orders before sending them to suppliers.
What problems are you solving with the product? What benefits have you realized?
We're a bookseller with hundreds of thousands of SKUs and selling many long-tail books. This tool helps us to optimize our stock by forecasting the demand for X days in advance (we use several modes) and thus ordering necessary items in advance, so we can ship them right after the customer orders them. It has helped us to improve percentage of books that we ship directly from our stock significantly, resulting in lower waiting times for customers and higher conversion.