Our Story

CRD supports a large national flood insurance provider with a variety of analytics use-cases, including fraud monitoring, claims processing, geo-analytics, and personalized marketing. Insurance is a serious and detail-oriented industry, and the CRD team processes large volumes of government regulatory data to reverse engineer the FEMA National Flood Insurance Program (NFIP) rater. CRD helped augment the existing risk modeling process of its insurance client by processing more granular elevation and hydrology data, and incorporating these new features into the client’s production system.

CRD supports a regional medical spa and injectables clinic through data pipeline work connecting the ERP systems and customer identities across multiple stores into a single unified data environment. CRD creates operational dashboards that are consumed daily by the client management team, and contributes strategic insights to inform future growth via de-novo (new) locations and acquisitions. CRD insights on employee level performance and customer-level loyalty and product replenishment prediction have driven incremental revenue for this expanding chain, and the return on project investment grows with scale as new locations come online.

CRD helped a regional construction management company to create a procurement optimization workflow helping it save on raw materials acquisition. The construction company received quotes for different lumber types from dozens of vendors, and the comparison of SKU level data across vendors was arduous and led to inaccuracies. The CRD solution ingests item price data across multiple formats received from the different vendors, and calculates the optimal quantities to purchase from each vendor to meet project requirements and minimize cost.

CRD helped a biotech company optimize the yield of a manufacturing process to create a highly in demand pharmacological input. By applying computer vision to transcribe the written notes hidden in hundreds of batch lab reports, CRD extracted the relevant parameters on which to build an ML model to increase throughput. Parameters such as refrigeration temperature, centrifuge speed, and ambient factory temperature, were optimized, all the while adhering to strict quality control rules.

CRD helped a tire manufacturer create an algorithm to optimize the maintenance schedules for a surface mining operation. The team collated data from tire temperature and pressure sensors, maintenance records, and local weather to create a reliable prediction of failure likelihood. By quantifying the amount of wear (ton kilometer per hour - TKPH) on each tire, the team created a real-time dashboard & a maintenance schedule based on objectives. The algorithmic insight improved productivity and reduced downtime for the manufacturer, and increased demand for the client’s pressure monitoring system.

CRD often works with private equity clients to build proof of concept, MVP and ultimately fully deployed data driven management tools for portfolio companies. One PE-backed client operating a regional chain of clinics for aestheticians worked with CRD to create a data pipeline to transform transaction, claims, and gift card data from their numerous disparate EHRs into a strategic operational dashboard. As the company obtained additional capacity to expand and double in size through an acquisition, the ability to “lift and shift” code written for one entity to a new one using the same systems factored significantly into estimates of project performance and exit multipliers.

CRD team members helped a global CPG manufacturer predict the volume impact of promotional activity, controlling for seasonality, weather and marketing spend. The pipeline CRD developed pulled weather and Nielsen ratings data and merged this with internal sales records for a major ice cream brand. The model developed unveiled insights on the optimal timing of promotional and marketing activities after incorporating all relevant data inputs / hypotheses.

CRD team members helped a national luxury retail / e-commerce group to kick start its personalized program. The team developed a customer data unification pipeline across various unlinked web browse, email click, and purchases / returns data sources, and we used this to create a feature set of historical training variables for each customer. CRD then developed a model to predict likelihood of conversion and the next best product to offer each customer, and ran a series of experiments in coordination with the client's email marketing team. The experiments showed significant lift over historic baselines and led to expansion of the personalization program + hiring of a dedicated marketing data team to support it.

CRD team members helped a PE-backed restaurant holding company build a demand forecast model and implement it within a managerial labor planning dashboard. The pipeline CRD developed pulled weather and sporting events data and merged this with internal sales and staffing records to create a rolling 4-wk forward forecast. This forecast showed significant performance gains (in terms of MAPE) vs. the prior naive baseline, so the team ran a pilot using an automated email and link to Tableau forecast view, and conducted info sessions and office hours to improve adherence.