Driving product innovation through data ‹ Prime Holding JSC – Premium software development.

Digitalization and the relentless focus on operational excellence are putting enormous pressure on product companies to innovate faster and faster. A study by McKinsey shows that in 2020 digital offerings flashed through seven years of progress within the span of just a few months. This has led to an exponential growth in data, which now companies are looking to leverage as a competitive advantage in a new economy.

Traditional approaches to data analysis can’t quite keep up with new trends and standards. But traditional approaches to business can find new life in this new, revolutionized digital reality. Specifically, businesses have always invested in trying to figure out the trajectory of the market and forecast consumer behaviors. Now, technology can bring them closer to their goal.

Developments in predictive analytics are shaping the future of marketing. In fact, the predictive analytics market is estimated to almost triple in size by 2027. Smart companies have started adopting predictive modeling tools in order to be able to forecast consumer trends and behaviors as a way to better understand their customers. This understanding is the cornerstone of true product innovation.

One such company offers an OKR (Objectives & Key Results) platform to its customers and is dedicated entirely to understanding the customer objectives and ensuring these objectives are met. This is why utilizing past and current data for the purpose of predictive analytics became a vital element of their strategy – and they chose to partner with us to make it a reality.

The Challenge

Our client’s goal was to introduce data-driven enhancements to their OKR software. As their product promotes efficiency – the seamless integration of objectives and implementation – they were under pressure to deliver it in a tangible way within a highly competitive fast-moving multimillion-dollar market. The key to gaining a competitive advantage was to make the most of customer insights. 

Developing and integrating data-driven products requires a specific set of skills and industry know-how. Being a product company specializing in software engineering, our client lacked the resources to embark on a product transformation journey of such magnitude. 

The Prime Holding Solution

Our transformation journey started by performing agile analytical work in various segments of customer and product data. This is always the first step in any project and serves the purpose to grasp key concepts and gain knowledge about the product and how to ensure our solution aligns with business goals.

For the analytical workloads, we developed a PostgreSQL database that we connected to an automated data pipeline within Azure Cloud, thus leveraging Azure Data Lake and Data Factory to source data from MongoDB. As development efforts progressed, we enabled Azure Synapse to handle data orchestration and supersede Azure Data Factory. This enabled us to collect valuable data about customer habits so we can bridge the gap between what they type into the platform and what result they are looking for.

Once enough insights were gained, we proposed to develop a Long short-term memory (LSTM) deep learning model based on TensorFlow and integrate it as a core natural language processing (NLP) feature in the product. 

Our new feature provides real-time semantic suggestions inferred from customer-supplied text fields. Once the user writes out the text, this information is sent over to the model API and predictions about the form fields are returned to the front end. As product architecture is based on microservices, embedding the model service within this context was done using gRPC.

By leveraging technologies such as Docker, MLflow, Jenkins, and SonarQube, we achieved high level of automation of the data science and machine learning operationalization workflow for seamless model training, testing, evaluation, and deployment. As part of the ongoing monitoring setup, we developed Grafana dashboards with KPIs and metrics to monitor real-time accuracy and performance of the service in production.

The Results

Our model service exists in a live Kubernetes cluster serving thousands of customer requests daily. Through vigorous code optimization and testing, we achieved a mean response time of 150 milliseconds per request and model accuracy of 95%. We continue to bring value to the customer by researching, building, testing, and delivering new, data-driven product features.

Each client is unique, as are the ideas and challenges they bring. But there are elements that are identical everywhere – the push to innovate and the desire to better understand your customers. The Prime Data Science team has the skills and experience to help you achieve exactly that. Depending on the specifics of your project, there are several ways we can help you drive product innovation and scale your data science and analytics.

Get in touch with us to talk about your idea and figure out the best step forward.