What if Your Data isn’t Working for You?

what-if-your-data-isnt-working-for-you
Photo by energepic.com from Pexels

Data collection, data analytics and the application of AI should enable companies to make better decisions.

Why then, despite the daily production of 2.5 quintillion bytes of digital data, is this not the case? Salesforce reports that for 1700 leading companies surveyed, generating good quality data to improve decision-making is the number one challenge. According to IBM, poor quality data is now so ubiquitous it costs the global economy upwards of $3 trillion annually.

The problem is two-fold:

  1. The type of data that is made available.

  2. The manner in which data is presented.

The Type of Data

Key stakeholders within companies who consume data are often not involved in determining the type of data they collect. This practice is like a chef trying to run a kitchen without any control over the raw ingredients being purchased. Most companies do not have established processes to involve decision-makers. As a result, managers lack access to the specific type of data they need to make better decisions. Good data is the essential pre-requisite for artificial intelligence technologies such as machine learning, which require a foundation of high-quality, structured data sets to work. Without useful data, any investment in AI, therefore, will likely not achieve the desired returns.

The Presentation of Data

Most employees in a typical company do not have a technology background, and even fewer companies invest adequately into training their employees in data analytics. Companies often generate troves of data without any comprehensive plan to organise or use it effectively. Inconsistent or disorganised data results in companies failing to gain a competitive advantage despite their efforts in data collection.

Leaders need to ask: what data is most relevant to capacitate the strategic objectives of their organisation?

To answer this question, a company must evaluate their department’s key functions, the interactions between them and their contribution towards the advancement of the firm’s strategic objectives. Then, they need to identify their data needs and create a strategy based on these parameters.

This process will likely show that current systems of data collection are useless or data is not presented in a format that translates to bottom-line value. According to Forrester Research, as much as 75 per cent of all data collected is never successfully used for a strategic purpose.

Creating a progressive corporate data strategy should focus on these three steps:

  1. Data democratisation, that is, all key stakeholders should have access to all relevant data.

  2. Analytics and AI training should be made available to management to enable them to interpret data for decision-making. Also, data needs to be presented in a way that does not require the ongoing contributions of data scientists to gain tangible insight.

  3. A data-driven culture should permeate the entire organisation. Businesses should incorporate and analyse data on a daily basis, as well as collaborate with data scientists to shape the ongoing, iterative evolution of how data can be utilised.

Some firms have already started this process.

Airbnb recently launched a Data University, encompassing more than 30 data science modules tailored to the needs and skill levels of various work functions. The results have been positive, with more than 2000 employees trained in the last 18 months with weekly usage of the internal Airbnb data platform rising by 50 per cent within the same time frame.

Another encouraging example is Unilever with their internal insight engine, People World. This platform allows global marketers to pose natural language questions and consolidate organisation-wide data mining documents and social media insights rapidly. Through queries sent to People World by Unilever employees, the company can see internal information demand in real-time. The algorithms underpinning People World can correlate that information with external data like macroeconomic factors and competitive analysis. In this way, consumer trends might be better anticipated.

So how can your organisation ensure the most relevant data is available to the stakeholders? How can you use this data to achieve long-term strategic goals?

Focusing on patterns and trying to understand cause and effect does not always account for the qualitative factors that make up a critical component of decision-making.

Companies should consider a different approach: employees must think about how they can empower themselves to make better decisions about the business.

  1. Form working groups within every department for employees to share the metrics they include in their decision-making process.

>The aim is to reach a consensus regarding the collection and use of data.

  1. These working groups will benefit from having a data scientist and AI expert present to answer any questions about what is technologically feasible.

>The greater the understanding of technology, the sooner employees will be able to extract better insights from data.

Involving the workforce carries tremendous advantages. If the stakeholders requiring data are deciding what data should be collected and how it should be made accessible to them, any technological platform that an organisation chooses to deploy will be more adaptable.

The development of new technologies should prioritise the needs of those requiring data every day. Organisations need to remember that data scientists can provide companies with a competitive advantage only once the type and methodology of data collection accurately reflect the requirements of the business. If stakeholders are involved in data analytics, they can determine these requirements, set parameters and develop algorithms to deliver genuinely relevant insights. This will ensure that the data strategy advances the objectives of the business.