Analytics modelling and the COVID-19 crisis

By Joanna England
According to a new study, analytics modelling has been rocked off its axis by the recent pandemic. We look at ways in which companies can right the bala...

Using data and insights to predict customer behaviour and market stability, has, for many years been common practice in the global business world. But the pandemic has wrought havoc with AI platforms that previously did an excellent job of predicting the future. The reason for this, a McKinsey study reports, has been the extraordinary behaviours that have affected the market. For example, huge demands for sanitization products, far less demands for retail clothing and petrol, consumers wanting far greater access to broadband services and so on. 

The usual trends analytical modelling and forecast software tracks, have essentially gone haywire. Systems that relied on predictable patterns to predict future trends, have been thrown entirely off course. 

New analytical modelling

Some companies are already looking at new ways to analyse data so that they can start working with predictive modelling based on the current, unusual climate. In data collected from the McKinsey study, analysts found that businesses most likely to succeed would be those that; identify the most critical analytics systems, and then perform a layered model-triage process that concentrates on analytics from those critical systems.

Ultimately though, creating new models that can successfully produce predictive trends, would be the best outcome. 

Building a new model

The more complicated a model is, the less transparent it becomes. According to McKinsey, “In cases where the simplest possible model is a complex one, explainability tools can make opaque models more transparent to reduce risk and enable humans to incorporate their expertise and judgment to address highly volatile elements.”

Then there are also issues related to which data should be used and analysed, which features are ranked – and why, and what assumptions were made during the model-build in the first place. For example, one company cited in the study, was an energy business that used a model which couldn’t predict data if the products (oil and gas) were places in negative values. 

Fixing the data source problem

While building new models helps companies and organizations in the long-term, businesses need accurate data now. There are several ways they can access information that will give them greater predictability, and using new or previously unused internal data sources, is one of them. Examining web page navigation and mobile app usage, as well as transaction data, might provide a much clearer picture of customer behaviour, rather than the more traditionally used credit scores. 

Current data, used differently

An example of this would be to see if your current data, might also offer insights into other customer trends and social behaviour. The McKinsey study cites an engine manufacturer, which used telematic data previously used to support maintenance work, to instead, predict and model traffic patterns as cites re-opened from lockdown.

Collect data more frequently

Sometimes, far more accurate trends are spotted when data is collected more frequently. For example, if information is collected once a week, one conclusion could be drawn. But collecting data daily, will very possibly show different information.

Use external data sources

Investigating open-source analytics such as those found in public health and specific locations, could be used to predict the behaviour of the workforce and therefore, supply chain difficulties and so on. However, this could prove costly as it may necessitate the buying of data. 

Agility and predictive data

One aspect of business management that has lent itself well to flexible data acquisition, is agility. Agile companies tend to advocate for faster decision-making. Therefore, the matter of data collection and analytics has been handled better within these organisations. 

The faster a company can respond to a shift in market patterns, the better the outcome for them. McKinsey references a telecommunications company as an example, saying: “One telecommunications provider uses agile practices to deploy analytics-driven micro-campaigns daily, evaluates the results immediately, and then fine-tunes campaigns the following day. This rapid-fire approach recently helped the company recognize quickly that one of the models informing the campaigns was not accounting for the spike in remote working. By updating its algorithmic models to account for this shift, it better predicted and responded to the need for additional products, including personal Wi-Fi hotspots—a product area for which it began capturing greater market share.”

Create a digital nerve centre

An ERP that tracks the data of every transaction, order, supply chain, inventory and so on, requires a team to monitor all changes and new trends. Organise a dynamic group that focuses entirely on the new market data. 

Embrace real-time data

Make sure the modelling system you employ has access to data in real time. With trends and global situations changing so frequently, it’s imperative to know exactly what changes are happening and when. The faster you notice the trends change, the faster you can respond to them.

Welcome agility

Restructuring your organisation and embracing an agile approach, has been proven to assist businesses in these challenging times. This is because agility enables a company to respond much faster to market changes and demands. 

Change with the world

The term, ‘the new normal’ is now one we must accept and respond positively to. Eventually, stability will be achieved. But until then, businesses must strive to achieve success against extraordinary odds. McKinsey states that when companies embrace; “interdisciplinary teams that include gender, ethnic, cultural, and geographical diversity, along with a diverse set of roles and perspectives, to develop new analytics capabilities for the business,” they are far better placed to survive in a post-COVID-19 environment.


Featured Articles

AI in SOC: Where Should Security Teams Look to Apply It?

As threats evolve, AI's continuous learning ensures robust protection that can prove invaluable for security operations centres

Swiss Re: Pharma, Not IT, to See Most Adverse Effects of AI

Swiss Re' AI report revealed surprising results showing pharmaceuticals stands to be the most adversely effected industry from the applications of AI

AI Safety Summit Seoul: Did it Meet Industry Expectations?

Before the summit, there were high hopes for meaningful outcomes - we see if industry leaders like EY's Beatriz Sanz Saiz thinks so

IBM's Salesforce Partnering Shows watsonx's Enterprise Reach

AI Strategy

Are Business and Government Diverging on AI Safety?

AI Strategy

Alteryx Industry-First AI Copilot Sees New Era of Analytics

AI Applications