The Power of Data in Times of Crisis

Liberating data has long been the request of many of our clients and prospects. Right now this matters more than ever but getting started on this to show value in days and weeks, not months also requires a different mindset to what has gone before.

One thing is for sure. The tools you need right now may not exist in your organisation and creating them could involve a complex and wide array of information sources that have previously not been considered for this purpose. Access to data may require cutting across organisational boundaries. Deriving insights may require stitching together disparate silos of data or tapping into otherwise dark data which is normally created or captured purely as a result of ongoing operations.

So how do you get started? Thinking about about the following areas is key:

At Kin + Carta our Agile strategy approach can help you define your key performance and risk indicators. Known as your Minimum Enterprise Data Set (MEDS) this is a focused, integrated set of strategically critical quantitative and qualitative data and insight. In this you should focus on what is important right now, in this crisis. This is made up of three key components:

1. Tracking market, competitive and organisational dynamics
2. Lagging and leading performance indicators, but also “anticipatory” measures.
3. Driving operations. An organisation is typically answering four questions to drive its operations which involve anticipation, control, improvement and communication.

Crisis management requires a more integrated view of your data. A single place for real-time decisioning. The data you need to feed your crisis dataset will be structured and unstructured, likely spread across multiple disparate data sources and technologies. Without worrying about how you might technically solve the problem of access and insight, the first thing to do is identify which data might be useful to feed the indicators you want to measure.

• List all of your sources of data that might be useful. Where are they, who are the business domain experts and technology experts. How can you gain rapid access to this data?
  
  
• Create hypotheses about how these datasets can inform what you’re measuring and who needs those insights. Pay particular attention to the opportunity to connect the dots across existing silos of data where technology can help create stories in the data you would not otherwise know about. What are the key entities that you could be using to tie this data together? For example, customer, market, geography or product. What about temporal alignment in real time?
  
  
• Think about personas and visualisations. How would these be different for management, specific departments, your employees, your customers and suppliers. Do you need a new tool or could you consider the downstream dissemination of insights into existing systems to help put information into the context of existing workflows? For new visualisations, in Visualize, Understand and Act: 5 Tenets for Visualizing Data to Compel Critical Actions in a Crisis my colleague Dave Clark discusses how the importance of visual aids in effective decision making. As a powerful example the phrase “Flatten the Curve” has captured the moment as a single visual to drive governmental decision-making and public awareness aid.
  
  

A rapid and can-do approach to data strategy will unlock the value you seek. The temptation is to think that this cannot be done because that is how it has always been, it will all require too much meticulous planning. But it can.

NHSX in the UK have engaged in an initiative to develop a data platform that will provide those national organisations responsible for coordinating the response to COVID-19 with secure, reliable and timely data; bringing together multiple data sources which already exist to provide a single source of the truth.

The days of working it all out before you start with data are long gone and that is in particularly sharp focus right now. Given the C-level directive to “make this happen” all relevant data sources can be unlocked and contents extracted using a variety of connector or real-time event-based technology approaches. Cloud-based data platforms and native technologies today are an incredible resource which offer all the scale, functionality and agility required to power any of your decision-making needs right now. Things you might be considering:

• Serverless, highly scalable, and cost-effective cloud data platforms designed for structured data that are easy to set up and manage and don’t require a database administrator, such as Google’s Big Query.
• The use of Google’s Natural Language Processing and Document AI API’s to rapidly extract value from unstructured data sources and also extract the entities and sentiment you may need to connect data together.
• The use of automated knowledge graphs to make connections across otherwise disparate data sources. By collecting all of your otherwise disconnected unstructured data in one place Elasticsearch can do this automatically and also provide immediate visualisations which can identify the most relevant connections across these data sources. This data can then be used to drive other downstream applications.
• The use of automated machine learning to identify anomalies or anticipatory patterns in time-series data such as those found in Google’s BigQuery ML or Elasticsearch’s Machine Learning module.
• Rapid Dashboarding tools to create decision-making visual aids which sit on these data platforms such as Google Data Studio or Elastic’s Kibana or Canvas.

Business can follow the lead of NHSX and also advice given by The World Economic Forum to rapidly leverage data, cloud computing and AI in times like these.

Don’t let the gatekeepers of your data or organisational boundaries stop you from leveraging the power of all of your data in this crisis. A quick-start data strategy approach can make this work. Cloud data platforms, technology and tools are ready to help you. Getting ahead of this crisis now with a real-time data platform will also give you the ability to put new hypotheses to your new dataset to anticipate what to do next.