Case Study: Dynamic Pricing AI
A.I dynamic pricing combined with pedestrian identification to increase profits by 18% for US brick and mortar retailer.
Background
Bringing Dynamic Pricing to Brick and Mortar stores has always been difficult.
Dynamic Pricing in Brick and Mortar Stores has always been has always been a difficult problem in comparison to their E-Commerce cousins. Firstly, unless the Store has digital price tags, it's a long and arduous task to update the prices for every item in the building. Whereas in an online store, prices can be changed storewide in a matter of seconds.
Secondly, it has been a difficult challenge to model the demand and conversion rates for each item in the brick and mortar context. Conversion rates are a great indicator that artificial intelligence price optimisation can use to understand pricing dynamics for each product, and are of critical importance in online dynamic pricing models. Unfortunately it's been impossible to accurately understand conversion rates in the offline stores, until now.
Working with a large Retailer in the US with over 30 Brick and Mortar Stores, Remi AI drew upon their previous work in image recognition and pedestrian modelling to deploy a customer tracking artificial intelligence in the client's CCTV.
This image recognition platform tracked how many customers looked at individual products. This was used to calculate conversion rates for individual products and was then fed into the dynamic pricing artificial intelligence to better understand how pricing changes impacted conversion rates.
Approach
By using pedestrian tracking in-store to accurately understand conversion rates for products, Remi Ai were able to the utilise this information to run parallel tests with different pricing strategies.
Our AI Price Optimisation involves numerous simultaneous tests of pricing strategy to quickly optimisation to the highest performing strategy.
Profit Margins increased by more than 18% with image recognition.
Results
Sales results show that our image recognition strategy outperforms the rule-based strategy of an experienced seller by a profit increase of more than 18%.
The client were able to understand for the first time which products their customers were interested in but not buying.
The client was able to use the image recognition to test different floor layouts and to study their customers engagement with the space.