Purseia | Our Thinking
page,page-id-342837,page-template,page-template-full_width,page-template-full_width-php,eltd-cpt-1.0,ajax_fade,page_not_loaded,,purseia-ver-1.4, vertical_menu_with_scroll,smooth_scroll,blog_installed,wpb-js-composer js-comp-ver-4.12,vc_responsive

Our Thinking

How we see it

At Purseia, we create lasting value by strengthening the bonds between our clients and their consumers. We do this by integrating four key elements – context, data, models, and algorithms. The unification of all four elements provides a blueprint for building reliable and scalable systems.


The set of circumstances, ideas, knowledge or facts surrounding a problem or event. This is the foundational element of organization and strategy. We use context to provide meaning to data and to establish a general framework for understanding growth.


Information or output that must be processed to be meaningful. Consumers are generating unprecedented amounts of behavior data creating huge opportunities in marketing and advertising, but data must be processed to be meaningful. By combining data with the right context, models, and algorithms we advance our ability to understand the key drivers of growth.


A set of assumptions and a mathematical formulation for data generation. Models provide statistical or mathematical structure that allows data and assumptions to be converted into more interpretable insights. The model provides a window into the ecosystem enabling better, more informed business decisions.


A set of rules or calculations to be explicitly followed. Algorithms allow computers to process data and improve efficiency with minimal human effort.


We know that building thoughtful algorithms creates value through efficiency and scalability. We have seen the power of relatively simple practices with commerce platforms to code triggers into the system based on browsing behavior. One of these triggers sends a coupon code every time someone browses without buying for a week. While this doesn’t address the problem of whether or not the triggers are effective, it scales to millions of users with very minimal labor demands.


Utilizing data to estimate statistical models is the conventional approach to analytics. It’s powerful and it reveals interesting insights. Raw data listing beer sales and individual characteristics of the people who buy it isn’t very useful by itself, but a simple regression model can reveal insights that inform marketing strategies and in-store displays. However, data and models alone don’t deliver repeatable business transformation so it’s important to use them effectively.


The explosion of consumer data has increased the demand for complex analytics, but sometimes it’s more valuable to use the available to data to validate old heuristics. For example, many traditional manufacturers design strategies based on assumptions like “more women between 25 and 34 buy our product than any other group.” Simple summaries of digital behavior and demographics can help keep such assumptions up to date.


Our approach to Data Science is the fusion of context, models, algorithms, and technology to transform data into value at a scale previously unattainable.