Light purple  solid background

COVID-19: The data deficit

COVID-19: The data deficit

COVID-19: The data deficit

Sophie Heading | Expert,

How long will this last? Has the government of [Country X] gone too far? Have they not gone far enough? Are you a social pariah for going out for your morning coffee? How far should you go to #FlattenTheCurve?

All valid questions, and around the (virtual) dinner table, my personal views are closer to this article from STAT, “A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data.”

But I am not a health expert, and I don’t want to wade into the debate around the economic value (or cost) of a human life. For business though, I think two of the fundamental points outlined in the article hold true: first, we are making decisions without reliable data; second, better information is needed to guide decisions and to monitor their impact.

We all know this in theory, but in practice we have a tendency to rely on intuition over data; 71 percent of respondents to our 2019 CEO Outlook survey indicated they have overlooked insights provided by data and analytics in the past three years because they contradicted their intuition. Which means we can fall prey to any number of behavioral biases and heuristics in our decision-making – like overestimating the probability of some relevant event (overconfidence bias), or estimating the probability of an event by the ease with which instances of that event come to mind (availability bias).

Even more relevant to our current environment is that our judgement of risk is not based solely on what we think, but also how we feel. Which plays out in several ways: although risk and benefit are often positively correlated in the real world, in our heads, they tend to be inversely related (i.e. high risk means low benefit). When we express information in terms of relative frequency (e.g. 10 in 100), we judge it as more ‘risky’ than when it is expressed as a probability (10 percent). We also become more sensitive to possibility, rather than probability, which means very small probabilities carry far greater weight.

Decision-making can still certainly benefit from ‘gut’ feel, and there is no doubt that a degree of human empathy needs to be taken into account when preparing your response. But it is data that will provide more certainty amidst an uncertain and fast-changing environment. So, what do you need to measure and monitor?

As usual, this depends on your geography and your sector, so this list can’t attempt to be exhaustive, but it can be helpful to frame your thinking into three key categories:

  1. Forward-looking: this is a time where predictive analytics can truly ‘make or break’ a company. The most important analysis you can undertake - at what point (if any) under which scenarios is the continuation of operations no longer viable, and suspending business activities is preferable?

    For many private enterprises, detailed analyses of future working capital and cash-flow requirements are being conducted – to inform the consideration of how to: optimize the availability of tax incentives; identify alternative financing sources; ensure access to all available government incentives and programs; and recognize expense-reduction opportunities.

  2. (Near) Real-time: inventory data, customer demand, supplier capability, and system performance category alerts. In the retail sector, the ability to manage demand has never been more important; some retailers are seeing demand fall away and customers shift channels, others are facing unprecedented spikes in demand.

    Simulations can be run swiftly to identify “sweet spots” between apparently conflicting objectives, based on real-time data. Increasingly enabled by AI and automation, these scenarios can help prescribe rather than just predict.

  3. Rear-view mirror: COVID-19 is a perfect example of a ‘structural break’ - an unexpected shift in time-series data when patterns among historical variables change. Leading to the unreliability of the model and significant forecasting errors, and fundamentally impacting assumptions and planning (the phrase ‘blindsided’ comes to mind here).

So, historical data needs to be taken with a grain of salt at times. But it can still inform future decisions. Monitoring and measuring how you come through this crisis will be critical to ensuring your eventual rebound and embedding longer-term resilience. Take the chance to measure and evaluate business (in)efficiencies stemming from alternate models of work.