For many businesses, choosing a location to set up shop can be a rewarding hit, or a costly miss, so one Singapore software vendor is tapping the use of data and predictive analytics to help mitigate the risk.
Omnistream’s technology analyses proprietary and open-access data to help retailers and brands figure out key operational issues, such as the customer profile of individual retail outlets based on location and the most optimal inventory each outlet should stock up on. It also offers data insights on outlet location, store layout, and marketing.
Founded in 2013, the company currently supports a clientele that includes convenience store chain 7-Eleven in the Philippines, and wellness retail chain Osim in Singapore.
Omnistream bases its analytics on three main types of datasets, its founder and CEO Wendy Chen said in an interview with DTC Daily. Namely, it looks at first-party data that is proprietary to the retailer, including transaction and POS (point-of-sale) data, as well as open data sources such as traffic, weather patterns – since weather can affect demand for certain kinds of products – locations of amenities, such as schools and transport hubs, and distance to the competition.
The third dataset is generated from a pre-trained algorithm that understands the hyperparameters of similar customers in similar markets, leveraging this data to boost the accuracy of its predictions.
Chen explained that the machine-learning algorithm runs on hyperparameters, where the bigger the customer base and the more times the data is generated, the better trained the algorithm and the better businesses are at identifying whitespace.
Omnistream takes the pre-trained algorithm, using this as building blocks, and builds its proprietary machine-learning software on top of it to generate data insights.
The software vendor then comes up with a list of recommendations, such as potential locations in which retailers can set up shop and the expected sales these locations can generate.
Touting the company’s “outcomes as a service” pitch, Chen said Omnistream would only get paid if its clients saw a sales uplift from its recommendations, receiving a percentage of every dollar in incremental profits made.
“We want to help retailers make profit; and the way we do that is take a lot of the manual processing and information overload out of operational decisions, such as what to put in each store and where. They don’t care [about the technology or machine learning behind it]. They just want to know how to sell more and manage their operations better”, she said.
Omnistream claims the data insights it provides has helped retailers clock a growth of up to 24%.
Its software also builds “personas” based on customer data, such as shopper purchases and psychographic methodologies, to offer customer attributes that can help brands and retailers better engage consumers and influence their behaviours.
Chen said she was looking to further expand Omnistream’s operations in Vietnam, the Philippines, Indonesia, Thailand, and Myanmar, targeting markets in Southeast Asia that were “fast-moving and low-cost”.
The startup last May raised USD$1.7m (£1.3m) in seed funding from investors that included the venture capital arm of Singapore shopping mall operator Capitaland’s VC C31 and Wavemaker.