October 6, 2023
Every organization -- big and small -- now fully understands that data drives their business. How could they not? It’s being labeled as the “new oil,” and the “new gold,” and it is a precious commodity. In fact, many companies are looking toward the day when their data assets are included in their balance sheets, alongside revenue, inventory, property, plants and equipment.
This almost greedy quest for data has given rise to new technologies, springing up from well-sites of data. Sensors and Internet of Things (IoT) devices are gathering real-time location data to track performance on the shop floor, the driving patterns of automobiles, and energy usage patterns, to name just a few sources. Additionally, enterprise systems, such as CRM and ERP, are gathering, storing, and analyzing data.
Most recently, the rise of AI is not only adding urgency to the need for lots of data, but AI is also translating that data into meaningful insights. It is the receiver and giver of data at the same time.
Yet, everyone is so caught up with data as the fuel to effective AI and the idea that lots of data represents success, that they’re looking at quantity over quality. Below are five key data misconceptions that can impact how effectively data is used:
Your business will be judged by how much data you collect. Ever since data became a business imperative, companies have been scrambling to gather as much of it as possible, without a coherent strategy. Yet, no amount of data can solve a business problem if it’s not the right kind of data. Companies will not be judged on the amount of data they have, but in the business decisions that were derived from it. For this reason, before going out and harvesting data – the more the merrier -- companies should first identify their business problem and then set out to collect the data that can help them solve it.
All data has to be your own. It’s true that data you collect about customers, stakeholders, partners and products is unique to your company, and very likely to provide highly specific insights about your business. However, it doesn’t have to be your own internal data that informs sound decision-making, or fuels AI algorithms. Once a business problem is identified, it’s important to take a data audit to gather all the information that could be relevant, and then supplement it with synthetic data, data that is artificially generated rather than produced from real-world situations. Synthetic data not only helps you get the information you need, but it also eliminates privacy concerns that can arise when personal or confidential data is used for training data.
Data is objective. Companies that espouse to be data-driven because it’s scientific and unbiased and factual may need to think again.
Not all data is unbiased. As data scientists cull training datasets for AI solutions, bias can creep in based on the data collected. For example, when a solution is being trained to determine the eligibility of a person for a mortgage, how diverse is the data it is being trained upon? Or, if data is being used to train a facial recognition system, how diverse are the images it is using? Data needs to be looked at through a lens of unbiased diversity, since even the very act of collecting data (or failing to collect certain types of data) can enable bias to creep in unknowingly.
Data comes from text. When we think of data we often think of facts and numbers: lists of customers and interaction dates, quarterly earnings figures, product parts, numbers of cars, streetlights or whatever information is needed to inform business decisions. But today’s data is much richer. Data comes in the form of images, often beamed down from satellites, or captured from graphics and photos, or it even may be video and audio files. All these types of data work together to tell a story that ultimately solves a business problem by revealing insights.
Data is the domain of IT. As companies work to remove the data siloes between departments and create a single source of truth in a centralized data repository, the question arises as to who really “owns” the job. Since data is the fuel for a myriad of technology solutions, including AI, it might seem like the IT department should be the responsible party.
Yet as data becomes a corporate strategic asset and the lifeblood to effective decision-making, it needs to be driven on a more horizontal, enterprise-wide level. It’s not only about the IT platforms that will support the data, but a need to establish data governance -- who has access to it, what type of data is off-limits and how is privacy ensured? Many companies are beginning to adopt chief data officers who come at data from a business perspective and whose sole job is to make data a key corporate asset that drives decision making.
Data is really nothing new. Thanks to new data sources, the rise of AI and the quest for the truth, we just have more of it. Yet, before businesses frantically work on piling up their data assets like money in the bank, they need to identify what they’re hoping to achieve with it. As Mark Twain once said, “Data is like garbage. You'd better know what you are going to do with it before you collect it.”
About the Author(s)
You May Also Like