A finished data product needs evangelists to curb the data literacy gap. When a data product finishes development, adoption is critical. Data science works closely with marketing to create valuable content that helps translate complex abstract data science into tangible assets for the end-user.
Q2 2019: Release of Core Metrics and real-time ETL
Q1 2019: Release of self-serve report builder and core reports
Creating a standard, consistent, and effective method of measuring marketing investment effectiveness can be very challenging. Adding a global investment environment, with dramatically varying global consumption patterns, can compound those difficulties. Throw in macro-economic impacts, and the desire to determine the best investment amounts to be allocated to marketing a product, and these challenges manifest themselves into the journey of American Express’ Travelers Cheque marketing investment analytics work over the past few years. The purpose of this paper is to share a versatile marketing investment optimization framework, focusing on clustering consumer behavior, leveraging SAS/ETS® software for time series sales pattern trends, and utilizing a Monte Carlo type of approach to determine variability and confidence.
Data projects go through a quick prototype. It usually starts with digging deep into the problem that is trying to be solved, and identifying what data science techniques can be applied to solve it. Once those techniques are identified, a prototype is ‘drawn’ up of what the outcome would look like. This helps product leaders understand the value and impact upfront. The risk of a data project failing is drastically reduced just by having a prototype process in place.
Successful implementations of data product development start by ensuring that data teams are working on the most important feature. Part of being a data scientist is how you communicate a data products roadmap to executives and business leaders. Getting buy-off is also critical. Voting and scoring the roadmap items ensure that all aspects of your business are focusing on the most import data initiatives.
Top priority features are managed in an agile backlog. Each data product feature (like core dashboards) are broken into small development ‘stories’. Stories are prepped with input from data science, development, UX, and product management. By the time a roadmap feature makes it to seeing actual development work, data scientists can put their head down and build amazing data products with minimal blockers.