Origin Story

Sandip and Chris have spent their careers working for various e-commerce businesses of all sizes and across multiple industries. In our roles, we’ve worked on all aspects of engineering, analytics, and data science. We’ve noticed that it’s hard to do analytics right. Most business analytics tools give their users a deluge of information, charts, and dashboards and then leave it up to the user to glean actionable insights from this information. It takes time, energy and precious headcount for the business owner to take all of this data and know what specific actions they should take. If large companies with teams of highly paid analysts struggle with this, what chance do small businesses have that don’t have the resources to dedicate to a dedicated analytics team? Frustrated by the status quo and having serendipitously come across the Square hackathon, we set out to build dataStacker

What dataStacker does

We set out to build dataStacker with the aim of making it :

powerful for small businesses to get straightforward insights and clear actions from their customer data.

Small businesses are facing increasing headwinds, dealing with economic instability, severe challenges due to the pandemic, and an increasingly complex competitive landscape. Given all of this, they don’t have the resources to hire on expensive outside analytics expertise, or spend hours upon hours developing the level of knowledge needed to glean the appropriate actions from an analytics tool. Our goal was to hide the complexity of sophisticated machine learning techniques on the backend, by only surfacing tailored clear-cut actionable recommendations that highlight important trends to power customer retention, loyalty and enabling targeted upselling.

We built smart retain by dataStacker to help small businesses understand their customer base. The focus is on understanding the value that each organic customer segment has to the business, and helping them figure out what business actions to take. By leveraging machine learning technology we find complex patterns in our merchant’s sales and customer behavior data to create an action plan for each customer segment. We surface insights such as which customers are most likely to be flight risks, which are the most loyal customers, and identifying customers for upselling opportunities.

How we built it

We power dataStacker by leveraging the square POS API’s, using the orders, invoices, customers products and loyalty endpoints. In a secure fashion built with cloud security best practices, each small business merchant has multiple custom machine learning models built in the background. These machine learning models are used to discover patterns in the sales and customer behavior data, and then power the actionable business recommendations that dataStacker presents to our users.

Challenges and accomplishments

Despite our combined decades of technology experience, this was a challenging project to pull together, especially in such a short period of time. Learning the data points available, nuances of the Square API, and designing a data ingestion process had to be done quickly. In parallel, the development of machine learning models that would find the types of insights we wanted to leverage was challenging, especially without live customer data to work with. This necessitated the development of multiple realistic, synthetic datasets that would challenge our models in different ways. In addition, we also had to develop the technology that could translate complex numerical insights and trends into human readable format, conveying what our models find as interesting into easily understood text. In short, this was a highly complex project leveraging our full technical skillset. In the end, it is rewarding to see a working product that has the ability to change the way millions of potential small businesses operate. We look forward to building on this early iteration and building an even more powerful product.

What's next for smart retain by dataStacker

Our goal is to provide the same set of powerful tools that Macy's or Gap would use to understand their customers in the hands of small to medium Square e-commerce business owners. We believe that smart retain would make a great and valuable fit into the Square app marketplace ecosystem and look forward to making it a part of the marketplace.

Share this project:

Updates