PriceShare

Help: Origins Back to Help

The PriceShare project began in May 2014, when Eric Savage performed a basket price check across 6 stores in the Upper Highway area near Durban, South Africa. His motivation was to determine how much more expensive Woolworths were as a chain, given their status as a premium brand. The basket contained 24 common products, with a mixture of in-store products and known national brands.

The results were drafted into an Excel document showing the price differences, and the image was shared to an 8000-strong Facebook group of local residents. There was a considerable reaction to the post, with several hundred comments reflecting a range of views about the implications of the data.

Birth of an Idea

While coming to the end of the price capture, Eric Savage started to formulate a web concept on how the practice might be extended. While websites were already available to perform price comparisons, these relied on data that was available electronically. In other words, if a website listed products with prices, a "bot" could scan these pages automatically and store the data.

South African supermarkets, however, had no such public-facing data available on the web, so the process of capturing the data would need to be manual. Obviously this would be tedious as a one-at-a-time job. The key, Eric realised, was receipts. These could be scanned via image text recognition in order to speed up the process.

First Implementation

As a software developer, Eric had the skillset to implement a price-capture solution fairly quickly. The first version of the site was completed in the space of a few weeks.

Development of the site was done on the principle of "MVP" - minimum viable product. In other words, the least number of features were to be introduced in order to make the site usable. Other features would be introduced as demand arose.

The prime risk of MVP is that first impressions can make an indelible impression. Potential customers could visit the site for the first time, be unimpressed, and never come back. As such, a certain minimum number of features would be needed to at least make a lasting impression.