top of page
forecast-comparison-cover-large.png

Company

Flieber →

Flieber is a technology company focused on helping multi-channel online retailers optimize their inventory and efficiently operate their supply chains. The system uses advanced data analytics and machine learning to connect sales forecast, inventory availability, and supply-chain decisions in real-time. The result is a drastic reduction in the stock-outs that hurt sales and overstocks that hurt margins and capital allocation. The co-pilot for inventory decisions

ltmgswa3e6zeihhp23ff.webp
Glossary

Stockout

A stockout is an event in which a product is exhausted in inventory and, therefore, unavailable to fulfill an order.

Overstock

Having too much stock in a warehouse that has not sold which increases storage costs and reduces working capital

Forecast

Forecasting in retail involves utilizing existing data to predict future events and, more specifically, consumer behavior. The goal is to avoid stockouts and overstocks. 

Forecast Model

Different statistical tools/algorithms designed to predict future trends and outcomes based on historical data. It involves analyzing past patterns and trends to make informed predictions about future events, sales, demand, or inventory levels.

Project overview

Role

Lead designer

Development

~ 3 months

Release

Currently in development

Context

Forecasting plays a pivotal role in the retail landscape. Flieber, with its advanced Forecast Models, is equipped to scrutinize sales and purchase orders, harnessing the power of machine learning to fine-tune forecasts. These models adeptly account for seasonal fluctuations, outlier occurrences, inventory constraints, and a multitude of variables to deliver optimal predictions. The Flieber Forecast model takes the lead as the default forecasting model for all new products added, automatically incorporating considerations for growth and decline based on Flieber's cutting-edge algorithms.

Nonetheless, when it comes to understanding a product's performance, those actively involved in selling it possess unparalleled insights. In light of this, Flieber assumes the role of a co-pilot, complementing this expertise by providing alternative models that can be applied to each product as the need arises.

 

To gain a more in-depth understanding of each available model and determine the best fit for a specific situation, the integration of a comparison tool becomes indispensable.

At the end of the day,
why should we build this Feature?

Goal

Compare and select between 6 different forecasting models for each product, including custom imported forecast and current customized forecast.

Scenario

ANALYSIS

POSITIVE

  • Flieber has its own algorithm and forecast model

  • Flieber can already provide other forecast models

  • Flieber has a dedicated forecast page and it's the most accessed page on the platform

  • Customers have shown interest in a scenario analysis tool to compare different models

NEGATIVE

  • Model selection doesn't offer a comparison 

  • Model change is not shown in real-time

  • no way of saying the effects of customization in real-time

  • Limited customization
     

BIGGEST CHALLENGE

Make the flow easy too use and not too convoluted. Data visualization needs to be clear.

User profile

SMB

Single brand and operations team with up to 3 people usually.

User

Single operator dedicated to forecasting.

  • Excel user

  • Medium tech trust

  • Medium tech familiarity

Enterprise (Aggregators)

Multibrand and operation team with more than 3 people usually.

User

Single operator per brand dedicated to forecasting.

  • Excel user

  • High tech trust

  • Medium tech familiarity

Development strategy

01.

Improvements to the Forecast Page

Small changes to the forecast page by repurposing preexisting components and choosing better data to display. The goal is to create a better relationship with the page and encourage mode customization.

02.

Default forecast settings

Small changes repurposing preexisting components to focus on creating a more customizable forecast experience by providing a way of selecting which model will be automatically applied when a product is first introduced to the inventory.

03.

Adding the ability to consider sesonality

Small changes repurposing preexisting components to focus on creating a more customizable forecast experience by providing a way to choose if seasonality calculations should be applied or not to the chosen model.

04.

Forecast comparison tool

Final design of the new tool that allows users to compare different forecast models, customize each model, and visualize the variations in real time. The action of comparing and applying a new model can be performed for single products or in bulk.

Style definition  Comparison tool

Card elevation and charts

Cards with elevation applied were chosen as the wrappers for each model with the goal of creating enough separation and breathing area. Attempting a less visually cluttered and more easily readable page.

The charts with full control of what's being shown and filters applied were a must to facilitate data visualization and comparison. It's also a familiar tool for the target user, used to macros on Excel.

Design output  Comparison tool
bottom of page