Guest post from Joannes Vermorel, Founder of Lokad.com, a software company that specializes in Big Data for Commerce, with a special interest in demand forecasting and inventory optimization.

Over the years, as my company specializes in quantitative optimization for commerce, I had the chance to audit the IT systems of many retailers. There, the WMS - warehouse management system - is typically a central component with the IT landscape; especially for online merchants who do not manage actual stores.

In retail, on average, the WMS is a rather hellish enterprise bloat-ware that was chosen years ago thanks to some broken RFQ process. It’s big, it’s slow, usability is abysmal, it took months to setup, every update still takes months, the license fees are crazy high (6 digits in annual license are fairly common) and the staff only preserve their sanity by restricting the number of hours they spend per week interacting with the WMS.

When I discovered SkuVault a bit more than 1 year ago, it was a small revelation: a slick WMS did exist. Until that point, “slick” and “WMS” were antonyms in my book. SkuVault was a WMS without the enterprise fuss; an app crafted like what you expect to find in, say, the AppStore.

For Lokad, the availability of an app such as SkuVault matters a lot. We do not deliver inventory management – like SkuVault does – only inventory optimization through an app named Salescast and the No1 reason why our prospects struggle to get started with Lokad is simply because their WMS makes it insanely complicated to do something as simple as exporting their historical data.

In contrast, thanks to the support of the SkuVault team, it’s now fairly straightforward to export from SkuVault the data needed by Salescast to perform the on-demand forecasts and generate the inventory optimization report that specifies when to reorder and how much to reorder for all items in store.

SkuVault is not your average WMS, and, I believe that Salescast is not your average sales forecasting software either. At Lokad, we realized a few years ago, that the classic forecasting methodology simply does not work for retail. Period. In fact, all classic forecasting models have been designed for long, stable, high-volume time-series, which is a complete misfit for commerce, especially only commerce, where products are short-lived, volume are low, and demand is erratic.

The solution consists of adopting a very different type of forecasts, named the quantile forecasts. Quantile forecasts give you directly the minimal quantity of inventory you need to make sure that less than X% of your clients (say less than 5%) will suffer a stockout.

Between Lokad and SkuVault, eCommerce retailers need look no further for their forecasting and inventory management needs. I believe that Lokad delivers exceptional demand forecasting technology, but you don’t need to take my word for granted: we offer an Express Plan that is free and that does not expire so that you can freely challenge our technology with your data.