Project Title: SmartShelf – Automatic Shelf Space Allocation Software for Small Retail Shops
Student: Mark Hopkin
Course: MSci in Computer Science

Abstract:
Shelf space allocation is an incredibly relevant issue in today's turbulent market. Large shops
use advanced and expensive software to aid and generate their shelf space arrangement in
order to maintain competitiveness and maximise profit. However, research performed in
this project indicates that there is a gap in the market for an affordable and useable
program that automatically generates planograms for small retail shops.
The aim of the project is to create a piece of software which would be useable by a wide
range of small shops, providing enough depth of customisation to allow each shop to tailor
the generated planograms to their shop's needs.
In order to gain a full understanding of the requirements of owners of small shops, a broad
field study was undertaken. The results of this field research informed the requirements of
the system. After studying existing shelf space allocation algorithms it was decided a tailormade
algorithm was the best approach to make the software as useful as possible to small
shops. The challenging aspect of the project was to take the qualitative rules that
shopkeepers use every day to arrange their shelves, and quantify these when creating a
novel shelf space allocation algorithm, as well as creating useable interfaces to enable
shopkeepers to easily create arrangements.
The solution is a piece of software which allows shopkeepers to manage a catalogue of
products, categorising these products into custom categories, as well as create promotions,
which are rules that affect the profitability of certain products based on date or
temperature. There was a large focus on making interfaces clear and useable, which was a
challenging design task.
The user can then create a complex shelving arrangement using a simple interface and
select the products they wish to place on this arrangement. The custom shelf arrangement
algorithm then uses all of this information to create a profitable and sensible shelving
arrangement. The result of all of this input is a PDF file that contains a visual representation
of the shelving arrangement in the form of a planogram, in a clear and easy to implement
style.
After extensively testing the system and critically evaluating a generated planogram against
a manual arrangement, it is determined that the software created is both efficient and
effective in creating profitable shelving arrangements that a shopkeeper would be able to
implement. Testing also showed that there are certain elements of shopkeeper's knowledge
that the system could be expanded with, such as the matching of product heights to create
'neater' looking arrangements.