Decision Support Products
Concentric Market.
A what-if platform for strategic decisions.
Organizations have plenty of analytics, but what ties them together? We all know the benefits of centralizing data as “one source of truth” but what about having one model as “one source of analysis”? Applying market simulation we created a place for various outputs from existing analyses and disparate data sources to feed into a live model of any consumer market. Concentric Market is used today across industries like automotive, entertainment, finance, and travel.
Concentric Scenarios.
Imagine if you could provide decision-makers around the organization a simple, clean user interface where they can ask what-if questions and get immediate answers? And not simply visualizing or querying data, but having a live model ready to be explored. Concentric Scenarios is exactly that.
It is tied to Concentric Market as the engine, and connected to the organizational data and internal computing ecosystems through APIs and connectors. In that way decision-makers have a real-time system available to run quick simulations about their domains.
Market Predict.
Forecasting voter behavior is hard. Most techniques today rely on polls and trend analysis. But what if we could add voter surveys, social media analysis, detailed campaign media spend, and candidate messaging? At E.W. Scripps we created an election simulation as an enterprise software platform for campaign strategy optimization. Armed with an understanding of the potential impact of campaign investment and positioning, E.W. Scripps could support advertisers in real time about the best next action.
Decision Conjoint Analysis.
A conjoint analysis is a survey method that uncovers respondents’ preferences. But what if we wanted to understand the heuristics and biases people use when making decisions rather then their preferences? The Decision Conjoint Analysis does that using genetic algorithms to generate screens to show respondents. Each option shown corresponds to one of a library of behavioral economic choice strategies.
The novel analysis became a part of a global media agency toolkit.
Artemis Attribution.
A tool for attribution management built for Havas Media.
Built on top of individual-level data, we created a first-of-a-kind multi-touch attribution software that integrated paid, owned, and social media. Decision-makers in CPG, airline, and insurance used it to design the right message to the right people at the right time.
Data Science Work
Created a prototype and a product development plan for “Alexa for the boardroom,” an interactive what-if scenario spoken language interface and real-time decision support system for an automotive CEO’s executive leadership team
Built an industrialized decision-support system to analyze new sparkling beverage product launches based on operations, marketing, finance, product, and syndicated data for 60 markets
Replaced a multi-touch attribution solution with an in-house, self-service platform for a leading parks and resorts that cut costs by a third and sped up time to insights 4x (weekly from monthly)
Organized a process for collaborative analytics that integrates methodologies and algorithms, data and tools, and people and process across all companies of a global media holding company
Deployed a data integration software and simulation technology on top of a leading data provider proprietary products, systems, and databases
Built a system dynamics model of the global fertilizer value chain in advising the king of a North African country about allowing foreign investment in phosphorus factories
Developed an application to forecast and optimize demand and utilization of a leading semiconductor company’s globally distributed resources
Built a library of decision-making heuristics and an algorithm for fast creation of scale-free simulated social networks to test what-if scenarios in casual dining choice of restaurant locations
Forecasted online video streaming company’s subscribership, consideration, and perception with <1% MAPE (mean absolute percent error) six months in advance to provide the CMO strategies for anticipating Amazon Prime, iTunes, and HBO moves
Developed an agent based model for the Chief Actuary at a leading health insurer to optimize plans, formularies, and marketing