• Home
  • Services
    • Predictive Analytics Services
    • Data Science Training
    • Data Science Consulting
  • Training Classes
    • Intro to Data Science Using R & RStudio
    • Data Science For Beginners
    • Predictive Modeling with R & RStudio
  • Use Cases
  • Platform
  • Blog
  • Contact
High Time for Marketing Analytics
October 2 2017

High Time for Marketing Analytics

Mark Redfern Analytics

High Time for Marketing Analytics


Marketing Strategy is ripe for a gigantic change with the advances made in predictive analytics. Thus Analytics should be embedded in every organization’s marketing strategy instead of the occasional use for ad-hoc reports. Advanced Analytics can guide and support the Strategy for better customer understanding, targeting and life-cycle management. This insight can drive marketing campaigns, market share expansion, up-sell, cross-sell, churn or retention, credit and risk, and last but not least sales funnel management.

Using internal, external and social media data, the multidimensional segmentation of clients can be enriched and segmented along many profiles. Such a segmentation could be based in Value, Behavior, Propensity, Loyalty, Socio-Economic, Demographics, Life-Stage, and Attitudes.

By using analytics and data mining, the organization can build Best Practices strategies and develop marketing activities for each customer segment and select “individualized” approaches that might include:
• An offer for preventing churning, mainly for high-value, at-risk customers.
• A promotion for the right add-on product and a targeted cross/up/deep selling offer to desired customer segments with growth potential.
• Impose usage limitations and restrictions on customers with bad payment records and bad credit risk scores.
• The development of a new product/offering tailored to the specific characteristics of an identified segment and so on.

Some of the components that should be taken into account in the design of the Best Practices strategy are:
1 The propensity or likelihood of accepting offers as opposed to rejection or marketing offers
2 The market basket analysis across all the organization’s products
3 The current and future customer profitability and lifetime value.
4 Cost of customer acquisition across the various channels
5 The type of customer, the differentiating behavioral and demographic characteristics, the identified needs and attitudes revealed through data analysis and segmentation.
6 The growth potential as designated by relevant cross/up/deep selling models and propensities.
7 The attrition risk (churn propensity) as estimated by a churn model.
8 The payment behavior and credit score of the customer.
9 Cost-Benefit analysis that compares various

An example of how these pieces can be applied to work together in the Best Practices strategy in action, let’s consider the following example. A high-value banking customer has a high potential of accepting a mortgage loan offer but at the same time is also scored with a high probability to churn. What is the best approach for this customer, having been identified as more likely to accept an offer, and how should she be approached by the organization? Considering the high acquisition costs, a high-value, at-risk customer, the top priority is to prevent her leaving and lure her back with an offer that matches her particular profile. Therefore, instead of receiving a cross-selling offer, she should be included in a retention campaign and contacted with an offer tailored to the specific characteristics of her segment.

Anova Analytics is a Data Science partnership focused on advanced Data Driven Decision management, predictive and prescriptive analytics technology. We apply math, science and superior business expertise to build robust solutions for our clients. Our scientific expertise includes statistical algorithms, machine learning, logistic optimization, simulation, neural networks, and intelligent systems. Our scientists have a proven record in producing effective predictive models to extract hidden patterns to deliver efficient and robust decision support solutions. This is complemented by our Data Driven Analytics framework for rapid agile execution of predictive models and rules.

Our solutions that includes: Anomaly Detection, Best Practices Discovery, Dynamic Pricing, Sensory Data Analysis, Neuro-Linguistic Computation, At-Risk Identification, Market Segmentation, Advanced Customer Intelligence, Marketing Analytics, Web Analytics, Anomaly Detection, Risk Analysis, Simulation and Process Optimization.

Hospital Emergency Simulation
October 2 2017

Hospital Emergency Simulation

Mark Redfern Analytics

Hospital Emergency Simulation

ER Simulation
This simplistic model of an Emergency Department was designed primarily to demonstrate the usage of network markup elements in conjunction with Process Modeling Library. Specific markup shapes such as nodes and paths are used to define facility layout. Resources of different kinds are placed in the network.

In this case the resources are:
– Nurses, PAs and Technicians – of type moving [can move on their own]
– Triage rooms, Express care rooms, X-ray – of type static [are bound to their home locations and cannot be moved]
– Ultra sound devices – of type portable [can be moved by resources of type moving]
Technicians have their own sub-process to prepare for the ultra sound process and wrap-up afterwards.
Upon arrival, a patient registers and proceeds to the waiting room, from where he is escorted by the nurse to a triage room. After triage the patient goes to an express care room and then either X-ray or ultra sound is done with the help of a technician and a PA. X-ray process requires the patient to go the X-ray room, whereas ultra sound device is moved to the EC room where the patient is located.

The model presentation screen is organized as a number of pages (animation, main flowchart, etc.) with hyperlinks between them to enable easy navigation.

The above is Simulator Caption of Control Panel. Please note all the Parameters, Resource Utilization and Length of Stay
Action Take to improve Length Of Stay:
1- Add 1 Nurse
2- Add 1 Technician
3- Add 1 Ultrasound

Please note all the Parameters, Resource Utilization and Length of Stay in comparison to the 1st caption, and take note the Length Of Stay in the Bar Charts of the last 2 screens, reduction from about 50 minutes to and 30 minutes.

The above was accomplished by running the Simulation for 5 minutes for the Baseline state, and 5 minutes for the Projected state.

October 2 2017

Intelligent Performance Assistant

Mark Redfern Analytics

Intelligent Performance Assistant

Overview
Run-time performance of a reservoir simulator is significantly impacted by the selection of the linear solver preconditioner, iterative method, and their adjustable parameters. The choice of the best solver algorithm and its optimal parameters is a difficult problem that even experienced simulator users cannot adequately solve by themselves. The typical user action is to use the default solver settings or a small perturbation of them that are frequently far from optimal and consequently the performance may deteriorate.

Working with our customer, ExxonMobil Upstream Research Company, we developed an adaptive control, on-line system that optimizes simulator performance by dynamically adjusting the solver parameters during simulation. The systems starts with a large set of parameters and quickly choose the best combinations. These parameters are continuously adapted during the simulation using the solver’s runtime performance measurements (e.g. solver CPU time) to guide the search.
This machine learning-based software system, called the Intelligent Performance Assistant (IPA), works with SparSol, the sparse linear solver also developed by NeurOK Software for ExxonMobil. The pair have been integrated into ExxonMobil’s proprietary reservoir simulator and deployed worldwide.
The system can handle a large number of combinations of solver parameters, currently in the order of 108, and consistently improves run time performance of real simulation models, frequently by 30% or more, compared to the performance with the default solver settings.

IPA also includes a persistent memory of solver performance statistics. The runtime statistics from these individual runs is gathered, processed using data mining techniques and integrated back into the IPA system, thus allowing for continuous improvement.

Will Analytics Help Us Find Our Way?
October 2 2017

Will Analytics Help Us Find Our Way?

Mark Redfern Analytics

Are you in the Analytics & Big Data Promised Land?

The level of interest and investment in advanced forms of “data analytics” is unquestionably on the rise. Spending on Big Data solutions will continue in the middle double digit for the near future. Are we in the Promised Land yet?

An increasing number of organizations understand the need to become more competitive and become Analytics Driven throughout the enterprise, as well as the need to manage the realization of this new data-centric culture. Gartner Inc. found that by 2018, Big Data will be such a central requirement for information management that it will become “table stakes” in organizations’ bids to use data more creatively in making business decisions.
The promise of Data Driven Enterprise may have begun with Big Data, but turning Data to Decisions is not a Big Data component. Big Data is the highway infrastructure, Data to Decisions is resources, skills, competencies, processes and cultures. It is all new, and to reap the benefits and experience the promise, organizations have to become Analytics Driven Cultures.

Modeling Operations at Pharmaceutical Distribution Warehouses
Cardinal Health, a billion dollar pharmaceutical distribution and logistics firm, manages multiple products from brand name pharmaceuticals and generic drugs to over the counter drugs, health & beauty items and their own private label. They face a multitude of typical distribution warehouse challenges that are further complicated by the nature of pharmaceutical products, which are smaller in size, consumable, expensive, and could be life critical. Brian Heath, Director of Advanced Analytics at Cardinal Health, and an experienced user of The Simulation software, employed agent based modeling to solve various business problems, saving Cardinal Health over $3 Million annually.

The Problem:
Cardinal Health is an essential link in the healthcare supply chain, offering next day delivery to over 30,000 locations including hospitals, retail pharmacies, physicians’ offices, and direct to consumer. Other value added services including efficiency and demand management, working capital management and contract credit management add to the difficulties of poor manufacturing reliability and supply disruptions in the market due to FDA and DDA regulations. In summary, Cardinal Health must keep up with the variability in pharmaceutical distribution management.

Cardinal Health considers facility layout, flow of product, order picking, labor planning & scheduling, customer order requirements and congestion for analysis and day to day operations management. Traditional analysis tools such as empirical trial and error, are risky, expensive and difficult to make changes. Industrial engineering operations researchers would suggest mathematical models, inexpensive, but the models do not capture unexpected dynamics. If anything is open or has emergent behaviors such as congestion, a standard mathematical model would not be able to solve. Thirdly, process driven or discrete event modeling is not advantageous due to its inability to represent a facility naturally. This led Brian Heath and Cardinal Health to explore alternative analysis options.

The Solution:
Agent Based Modeling (ABM) with The Simulation and Modeling software gave Cardinal Health the device required to tackle many distribution warehouse issues without the restrictions of traditional tools. ABM represents abstractions of distributed autonomous entities that can interact with each other and their environment through space and time, allowing Cardinal Health to capture work time allocation, congestion wait time, cycle times, distance traveled, worker variability and other important metrics.

The model built was ultimately concerned with the activities of employees and the interaction with each other during the day, making it necessary to import data such as picking time and performance standards into the model. Now, Cardinal Health can gather congestion wait time data and see how much of a problem it is causing in the warehouse since “agents” are modeled as individuals with special relationships to each other. Additional parameters included in the model are several worker speeds, worker behavior, learning curves, cycle times, product turn-around and distance covered walking or driving.

The ability to import Excel files was also imperative as Cardinal Health has numerous warehouses, and it is mandatory to test multiple layouts. Using The Simulation, if a change is needed, it’s as simple as updating the Excel file, importing it into the model and running the model again.

The Outcome:
The Agent Based Model built with The Simulation software allows Cardinal Health to compare layouts, picking technology and product slotting strategies. In addition, they can evaluate different methods of picking to update staffing models and for on-the-floor support if a workload changes as orders vary on a day to day basis. Statistics is also gathered such as tact time, how many batches are completed in an hour, truck unloading time, and sequencing of events.
Besides the clarity given through the above metrics, the model revealed a problem due to the random distribution of work. Each employee’s work load was uneven making one faster and one slower. By balancing the workload, employees began working at a similar pace and congestion decreased dramatically.

By minimizing congestion using The Simulation software, Cardinal Health was able to decrease the average shift length from 10.5 hours to 7.25 hours and increase the amount employee capacity. Cardinal Health saves over $3 Million annually using Agent Based Modeling with The Simulation technology.