Quantitative and Statistical Consulting for Businesses
Precision Consulting has helped hundreds of corporate clients find solutions to complex business problems through the use of advanced statistical modeling techniques. More often than not, firms contact us to provide them with the solution to a particular obstacle or problem, as is the case at standard management consulting firms. Many firms have enormous amounts of extremely valuable data at their disposal, yet they lack the in house resources to optimally analyze their data.
What sets us apart is a core specialty in solving business problems that require statistical methodologies and data analysis techniques which are often beyond the scope of the client’s internal staff. While this is sometimes due to the problem requiring the manipulation of huge and hard to manage data sets, most of our clients come to us simply because the methods needed to solve the problem are complex and unknown to them. Engaging Precision Consulting for statistics consulting implies that you want a boutique firm that will provide you with demonstrable results.
Our solutions are data-driven and analysis-based, and we’ve yet to turn down an engagement due to complexity. Our clients include many Fortune 500 firms, small and midsized businesses, Government entities, and universities.
- Statistical Methodology and Data Analysis at the Forefront
- Our Capabilities
- A Small Sample of Our Clients
- Academic and Business Credentials of Our Team
- Our Culture
- Why Do Firms Retain Us?
- Examples of Problems That We’ve Solved For Our Clients
Statistical Methodology and Data Analysis at the Forefront
Precision Consulting’s core specialty is developing sound statistical methodologies to analyze business problems, and then carrying out the requisite data analyses to solve those problems. Ordinarily, firms come to us requiring the answer to a specific issue. Our first step is always to fully understand the problem and the data available to us in solving that problem. Given the type and amount of data available, we then determine the optimal statistical methodology to solve the problem. Alternatively, in the case of the firm having no relevant data, our first step would be to create the optimal methodology to collect the needed data.
In many business problems, the data available can be incomplete, disorganized, or unwieldy. In these cases, we often need to find a clever way to extract useful info out of the available data, and/or to supplement that data with additional data. At other times, we are presented with a vaguely-posed business problem, and our job is to specify the problem statistically, and to find a robust and practical solution that doesn’t succumb to data-snooping pitfalls.
In all cases, our consultancy differs from typical management consultancies in that our methods and solutions are always empirical and objective.
Many statistics consulting projects that we take on are brought to us after a traditional management consulting firm was unable to specify how to optimally solve the problem(s), or just generally unable to handle the quantitative rigor of the solution(s). We have a broad range of experience providing solutions through the application of a vast array of statistical and numerical methods, including survival and event history analysis, ARCH and GARCH models, neural nets, bootstrapping, Thurstone/Shapley analyses, structural equation modeling and multi-dimensional scaling, and Monte Carlo simulations.
Our comfort with virtually all higher-level quantitative methodologies and our expertise in applying these tools in a corporate setting enables us to assist with virtually any business problem. Here is a sample of some of the analytical tools with which we are familiar:
- Various forms of Regression Analysis, including Non-Linear Regression and Ridge Regression
- Survival Analysis including Kaplan-Meier Analysis
- Time Series Analysis, including Vector Autoregression (VAR), Vector Error-correction Models (VECM), and GARCH and its variants (NGARCH, EGARCH, etc.)
- Panel Analysis
- Survey Development and Reliability/Validity testing of existing surveys
- Statistical Power Analysis for Sample Size determination
- Multivariate Analysis (with multiple outcome variables), such as MANOVA and MANCOVA.
- Qualitative analytical methods including Phenomenological Analysis and Grounded Theory
- Optimization methods such as Linear and Nonlinear Programming, Genetic Search, and Simulated Annealing
- Markov Chain Monte Carlo and similar methods
- Nonparametric Methods
- Zero-Inflated Count Models
- TURF, Thurstone Scaling, and Shapley Values (used often in marketing-related projects)
- Conjoint Analysis, Choice Modeling, and Maximum Difference Scaling (MaxDiff)
- Structural Equation Modeling (SEM), Confirmatory and Exploratory Factor Analysis (CFA and EFA), Multidimensional Scaling, and Path Analysis
- Neural Nets
- Machine Learning and Artificial Intelligence (AI)
- Parallel Computing to handle very large datasets and for problems that would otherwise take excessive computing time
- Various Bootstrapping and Jackknife techniques
- Spatial Analysis
- All statistical software packages including R, Matlab, SAS, Stata, MPLUS, LISREL, EQS, PASS, Maple, Mathematica, and SPLUS
- Data Mining with Weka and RapidMiner
A Small Sample of Our Clients
Precision Consulting’s philosophy is to build long-term statistical consulting relationships by delivering quality work on time for a reasonable price and in language that you can understand. Our clients include many Fortune 500 firms, small and midsized businesses, Government entities, and universities. While much of our work is with large firms, we also take on projects for smaller or midsized businesses, as our schedule permits. Among the larger firms we’ve worked with over the past years are:
- Johnson and Johnson
- Clayton, Dubilier, and Rice
- JC Penney
- Waste Management
- Ecko Unlimited
Academic and Business Credentials of Our Team
Precision Consulting is committed to maintaining the most talented and diverse team in the statistical consulting industry. By principle, our team is comprised solely of members with outstanding contributions and demonstrable excellence in both academia and the private sector. On some days, our team may look like a boutique of academics that collaborate with each other as if in a think tank, while on others we deal primarily with real world analysis.
All of our principals have Phd or Masters degrees in quantitative fields at leading Universities including Cal Tech, Harvard, Yale, Cambridge, Columbia, and Brown. Although most of us specialize in theoretical or applied statistics, our staff also includes experts in pure mathematics, econometrics, mathematical finance, psychometrics, qualitative analysis, and applied mathematics. We share a primary interest in the application of quantitative analysis to solve business problems, and our team has extensive experience testifying and providing related statistical analyses in jury trials, arbitration hearings, and bench trials.
Our principals have excelled in the business world as well as in academia, and this is one of our biggest strengths. Most of us have extensive Wall Street or Government experience prior to joining Precision, and our cloud-based business structure enables us to recruit experts from all over the world to join our team. As such, it would not be unusual at all for a new client of our boutique Firm to work with a team comprised of a former associate professor of portfolio theory from MIT’s Sloan school, a bond trader who spent years on the fixed income arbitrage desk at Goldman Sachs, a former economist from the U.S. Department of Labor, or an astrophysicist who gazed at the cosmos before bringing his physicists toolbox to Precision Consulting. We’re not just academic statisticians, but rather we can speak to you in your language, and we can communicate our results in that language.
We also pride ourselves on our ability to clearly communicate our findings, and regularly present the results of our analyses to partners and executives. We can specify and understand the issues facing your business, because we’ve faced many of them ourselves.
We don’t strive to have the largest Firm in the consulting industry. We do strive to have the smartest Firm in the consulting industry. Our perspective is that demanding this background is imperative for us to be successful in the types of projects that our clients bring to us.
The rigor of the quantitative methods needed for most of our engagements often requires that the problem be worked on by someone at the absolute cutting edge in their academic field. In addition, the ability to seamlessly communicate with management at our client companies is a vital skill. This is particularly true when the underlying quantitative methodologies used to solve the problem optimally are not something that management is familiar or comfortable with.
As with most academic fields, the body of knowledge in statistics is continually growing. To offer competent statistical advice requires staying abreast of all new developments in statistical thinking. Communicating and collaborating with academic statisticians conducting research on new ideas is the most important part of that process.
To that end, we leverage our relationships at Cal Tech and Brown University by retaining several prominent academic statisticians and applied mathematicians as outside advisors during their sabbaticals, enabling us to create ‘dream teams’ to tackle some of the more complex problems that are brought to us.
Precision maintains a collegial, cloud-based atmosphere that encourages extensive cooperation among our team. We love to solve problems, and the more complex the problem is, the more fun it is to find an elegant solution. Many of our clients consider us to be their quantitative hired guns, on call whenever needed, and from any corner of the globe.
We’ve been referred to in the press as “Quant Hired Guns”, and “Management Consultants on Steroids.” Most of us had a passion for problem solving at a young age, excelling in chess and other mathematics-based games, and have kept that love alive to this day. Our principals started Precision Consulting for precisely this reason. Our team atmosphere is beneficial to us in many ways, and helps us to find out of the box and creative solutions.
Why Do Firms Retain Us?
- Nothing Is Too Complex for us. We can perform any statistical analysis, regardless of the complexity, and have never turned down a project due to it being too complex.
- We take pride in our ability to communicate our results in an easy to understand manner. We don’t consider our job done until our clients completely understand our findings, what they mean, and the approach we took to get there. Therefore, we offer a substantial amount of client contact throughout the analytic process. We believe that our client relationships extend beyond the delivery of a finished product, and don’t consider an analysis complete if it isn’t intimately understood by our clients.
- Our principals all come from outstanding careers in the corporate world and/or in academia. The quality of the team working on your project is guaranteed to be at the absolute pinnacle of the management consulting industry.
- Our fees usually pale in comparison to the savings and/or additional profits that our work produces for our clients. Ordinarily, an investment in our services yields a return on investment that dwarfs that of other alternatives.
- We use a system of cloud-based servers and technologies, allowing us 24/7/365 access to our clients. Our cloud-based operations allow us to integrate our work into our lives, rather than separating it from our lives. We pass this advantage on to our clients.
- We passionately embrace new technologies, and are surely at the forefront of quantitative consulting firms in terms of our technological capabilities and infrastructure.
- Our boutique firm is growing by leaps and bounds, having used the economic contraction as an opportunity to expand and enter new markets. We were in the 2010 edition of The Inc 500, placing us as one of the 500 fastest growing companies in the United States.
- While much of our work is with large firms, we also often take on projects for smaller or midsized businesses, as our schedule permits.
- We are comfortable and experienced in working with hard to manage, unwieldy, and extremely large datasets.
- Over 80% of our clients request our assistance more than once, as our clients are almost universally happy with our different brand of consulting. As such, we have the opportunity to pick our future engagements carefully. We don’t accept all projects that are brought to us, and often explain to prospective clients that what they want to do is not statistically sound or plausible. We also tend to choose the engagements that we do accept based on whether the analytical areas interest us and fit within our core areas of knowledge.
Examples of Problems That We’ve Solved for Our Clients
- A large American Cable network approached us due to having displeasure with the methodology used by a major ratings agency in measuring their number of viewers, as these findings were negatively impacting their advertising sales. The ratings, based on relatively small sample sizes, showed extreme variance day to day (the primary cause of their low rating). Through a combination of resampling, simulation, and bootstrapping techniques, Precision showed that the ratings methodology used was grossly inaccurate. Based on our findings and report, the ratings agency changed the methodology used for networks of this type, and our client was able to increase their advertising prices back to previous levels.
- One of the largest Private Equity firms in the country contacted us to predict Fund inflows and outflows based on a suite of predictor variables. Our findings were highly significant, and our client completely revamped their client relations practices during certain economic times, in an attempt to decrease outflows.
- A prestigious Medical School contacted us for assistance on a large-scale study on the risks faced by organ transplant centers. Our analysis included a binary logistic regression to estimate the risk of transplant failure based on patient characteristics (such as age, gender, life support, etc.), followed by a Monte Carlo simulation to estimate the probability of closure of a transplant center due to failed transplants and insufficient transplant volume.
- A Fortune 500 firm providing analytics for the healthcare industry contacted us to determine how to optimally impute missing values in their data sets due to a major retailer stopping the reporting of detailed sales data. The main challenges of this project involved working with a multi-gigabyte dataset and designing an application for automatic model generation, to capture the unique features of tens of thousands of products. A custom script was created in R to create the models. Given the number of estimations required, cluster computing techniques were used to drastically reduce the amount of time needed to complete the task.
- An international architecture firm hired us to evaluate a questionnaire that they use for their clients. We conducted a reliability and validity analysis using MPLUS, optimized their sampling methodology, revised the process used to aggregate the survey data, and then conducted a qualitative analysis of approximately 20,000 responses to the survey using NVIVO.
- A large firm specializing in the placement of teachers contacted us to create an assignment algorithm to optimally place teachers in schools. Precision developed the outcome variable of teacher placement satisfaction, analyzed the drivers of that outcome variable, developed an algorithm to optimally place teachers given various constraints, and assessed the performance of the algorithm over repeated trials.
- An email marketing firm contacted us to develop a model that would allow them to predict the ‘permanence’ of subscribers to their service. Using demographic data and other data about each of their subscribers, Precision developed a model to predict the amount of time a subscriber would stay, and our client changed many factors of their client relations process for those subscribers deemed unlikely to be a permanent subscriber. We have done similar analyses for real estate companies looking for info on likely rent renewals, and also for firms looking to improve employee retention.
- Precision created a March Madness Bracket Optimizer for a major online and televised sport news network, for use on their website. We analyzed the selection criteria that users follow when picking their brackets, employing a machine learning algorithm to game theoretically account for possible biases among competitors. Our method is still the cutting-edge selection tool for this annual tournament.
- One of the world’s largest purchasers of domain names contacted us to build an algorithm to estimate the value of a given domain name prior to purchase. Using their huge data set from 10 years of purchase decisions (several billion of data points), we produced an automated valuation tool which they now use exclusively. The client is currently working with us on another project related to the application of concepts from Modern Portfolio Theory (MPT) to their holdings of domain names.
- The leading aircraft sales company in the United States contacted us to produce a time series model of aircraft prices and residual values for tax purposes. We created a comprehensive model that predicted the selling price and residual value of aircraft purchased over any given period of time from 1 to 50 years.
- A major water bottle manufacturing company hired us to analyze bottle weight data from a variety of product lines and compare to those of its competitors.We conducted the analysis and found a statistically significant difference in bottle weights, which our client used in a new advertising campaign.
- A large European hedge fund approached us to conduct a default probability analysis for a multibillion dollar project. Using a Monte Carlo based analytic approach in addition to a bootstrapping procedure, we found the probability of default to be within the Client’s acceptable range.
- A large producer of peanut butter contacted us to analyze the variance in the consistency of their product. After several visits to their factory, we designed an experiment to collect the needed data. We then used Principal Component Analysis to determine the likely sources of variation and their significance. Within 6 months, our findings framed a new initiative to increase the consistency of their product.
- A Fortune 500 credit card company hired us to determine the optimal bundle of benefits to offer (and at what times) to their customers. We conducted a Shapley Value analysis and provided them to the client, who then improved their benefits offerings in several ways.
- A leading research institute specializing in cancer cell biology contacted us after their internal statistics team had excessive difficulty working with a huge dataset, as well as working with the types of models that would be needed. We conducted a large-scale data analysis requiring Kaplan-Meier survival analysis, Cox Regression, and Linear Growth Curve Modeling.
- One of the world’s largest cellular service providers, about to combine various different online platforms into one unique platform, inquired as to the probability that users of various different online platforms within their system could be using the same user/password combination, hence being able to log into each other’s account. We used techniques from information theory to assess the likelihood of specific sets of user/password combos being used by two different customers.
- One of the world’s largest solar cell manufacturers contacted us to design experiments and analyze the resulting data to increase the client’s knowledge of sources of variation within its fabrication process. We visited their factory, designed a series of complex experiments, and analyzed the results to reveal the main sources of variation in their process.
- A global leader in the creation and placement of slot machines throughout the world retained us to conduct a comprehensive payout structure analysis of their entire portfolio of slot machines. We worked with the client to develop a proxy for player enjoyment based on volatility and expected loss, and maximized that latent variable with respect to various proposed payout structures.
- The leading primary research firm for the hedge fund industry approached us to consult on the methodology, structure, and analytic techniques to be used for a new comprehensive weekly survey for retailers, to be used for creating a granular index of consumer health. We provided the firm with questionnaire design, sample specification, a power analysis for sample size determination, and a survey frequency assessment. We also prepared the optimal data analysis method for the client, which was a combination of time series and MaxDiff procedures.
Call or email us to discuss your situation and how we may be of help to you.
Please click here to read our Data Security Policy.
The image at the top of the page shows the Mathematical Bridge, which is the popular name of a wooden foot bridge between two parts of Queens’ College, Cambridge University. The bridge was built by James Essex in 1749.