Leveraging Data to Navigate Volatile Market Conditions

Looking back on the first six months of 2020, it’s clear New York is in the midst of unprecedented disruption. Disruption has taken several forms. While some industries have experienced temporary upside or downside spikes in demand due to social distancing requirements, others have seen business models turned on their head and rendered obsolete.

The speed and magnitude of COVID-19 induced disruption begs the question - will we be subjected to a higher frequency of ‘grey swan’ phenomena - semi-predictable outlier events that can dramatically impact prevailing market conditions?

The speed and magnitude of COVID-19 induced disruption begs the question - will we be subjected to a higher frequency of ‘grey swan’ phenomena - semi-predictable outlier events that can dramatically impact prevailing market conditions?

High performance firms generally operate business intelligence (BI) systems to enable data driven decision-making, theoretically providing increased competitive adaptability and resilience. (Click here for custom business intelligence tool development best practices.) Unfortunately, most business analytics approaches take static historical data and simply extrapolate it into the future. Such methodologies don’t enable sufficient agility and accuracy to make intelligent investment decisions in volatile business conditions similar to those facing decision-makers today.

Standard BI systems exhibit three potential failure points: data recency, data diversity and limited forecasting abilities. Solving for these three factors can unlock increased data expressiveness and data-derived business insights in highly volatile environments:

Standard BI systems exhibit three potential failure points: data recency, data diversity and limited forecasting abilities. Solving for these three factors can unlock increased data expressiveness and data-derived business insights in highly volatile environments

Data Diversity: Narrow data analysis limits peripheral vision within the competitive environment a firm operates in. Historically, BI systems incorporate only owned, first-party customer and product datasets. While useful in understanding historical performance, first-party data does not contextualize the factors driving disruption. In today’s environment, that could translate into understanding the geographic incidence of COVID-19 and its downstream business impacts like unemployment rate, mortgage forbearance rate, and eviction rate. When layered on CRM-level data, strategic go-to-market opportunities and exposures are much more easily identified.

Data Recency: Growing uncertainty correlates with increased volatility. Defined as an increased dispersion of measured values, heightened volatility underlines the need for higher recency datasets. High recency can help business analysts more quickly identify outlier business environments and enable firms to act more agilely in crafting upstream and downstream reactions to these conditions. Combined with more data diversity, increased data recency helps decision-makers identify and leverage competitive advantage from demand shifts and new solution adoption patterns. With the increasing availability of application programming interfaces (APIs) and streaming data infrastructure, it's increasingly than ever to support near real-time availability of owned data and hourly or daily availability of third-party partner data.

Forecasting Intelligence: While there is a continuum of forecasting methods available to business analysts, most BI solutions support simple mathematical operations on historical data. Statistical forecasting methodologies offer more promise. Techniques like linear regressions identify the best-fit line within historical time-series, delivering directional guidance without providing much predictive value, particularly in the context of outlier environments.

While sufficient in some industry verticals and business contexts, cutting-edge firms may opt for ‘machine learning’ approaches, including neural networks, that identify and ‘learn’ patterns provided by input data sets. Advanced neural networks can retain information over numerous time-steps and detect sub-patterns within patterns, anticipating all potential trendlines, while learning new ones as they emerge. Forecasting powered by ‘neural’ machine learning algorithms, particularly when working with large input data volumes, can equip decisionmakers with superior predictive insights and their corollary business advantages.

If relationships between variables are known or modelled, like expenses and revenues, analysts can execute advanced sensitivity analyses. Sensitivity analyses highlight the variance in a critical value or KPI given the percentage changes in input variables, highlighting the full range of possible outcomes, given the potential ranges of input values like cost of capital or goods.

Final Thoughts

While business intelligence, per se, does not require programmatic processing, and can be aggregated periodically for board meetings and strategic planning processes, building ‘always-on’ systems-of-record helps firms accelerate time to value. By making strategic investment decisions based on highly expressive, predictive data analysis, firms can increase their operational agility and both short and long-term competitive advantage.

We build tech-accelerated and enabled intelligence for customers. To learn more about our advisory services and how we help build and measure high-growth firms, contact us here.