Table of Contents
In times of crisis, organizations look for things to hold on to that give them a sense of certainty. After all, uncertain times mean tough going for business as usual, and also mean that you’re likely navigating in unknown waters. One of the first places an organization will turn to for answers is its data, and by extension, the person in charge of it.
The problem is, a crisis is usually a hard reboot, making all your old data irrelevant to the new reality, and thus making your life harder. It may not seem like a big problem on the surface, but consider that Gartner estimates “69% of organizations do not measure the financial cost of poor-quality data,” and that’s in non-crisis times. When you need to provide clarity and guidance, you’ll have to find external data that offers answers, which is easier said than done.
The pressure of finding fresh data
Even in non-stressful situations, external data acquisition in data science is time-consuming and complex. The process may go something like this:
- Understanding the business issue you need to be resolved, which is essential to understand the external data sources that may help improve your machine learning models
- Researching to find the right type of data and the right partner to provide it
- Performing your due diligence to make sure that it both meets your requirements, as well as compliance, latency, and coverage — all of which can make even a great dataset useless
- Testing new data — a process that can take anywhere from days to months when you factor in legal discussions, POCs, and commercial terms
- Calculating the ROI, which may not always be positive
Of course, when the situation becomes stressful, this process, which, again, can take anywhere from weeks to months, becomes simply illogical. The problem is compounded when you consider that this only accounts for a single data source, not multiple ones that you’ll likely need to find the right answers. The question is, how can you bypass this process of data acquisition? By finding data discovery tools that can connect you to thousands of external data sources, instantly.
Re-routing away from roadblocks
It’s hard to determine your destination if you don’t know where you’re coming from. You need to understand how your organization makes decisions so you can build a strategy to drive your colleagues toward a data science-driven approach. This may not be impossible if your organization already uses data to a degree. For example, if your organization already has an extensive BI infrastructure, it may already be using high-quality data and understand its importance. You can take it to the next logical step and demonstrate why adding predictive capabilities is crucial, and why it must rely not just on internal data, but on external signals about the landscape.
On the other hand, if your executive team likes to rely on their “gut instincts” and see analytics as a way to support or inform their assumptions, you may have to present data science in a different light. Instead of pitching it as a new model of decision-making, you can present it as a way of accounting for a broader range of outcomes they couldn’t have seen before. In this scenario, data science is a compliment, not a replacement, to their decision-making style. They aren’t invalidated, but empowered, by machine learning.
Understand your audience, and speak their language
It’s important to keep in mind that when you’re talking about data acquisition or changing the way you make decisions, you’re really introducing unknowns into your organization. Communicating the value of data clearly, and understanding your organizations’ needs and preferred data consumption models better will help you acquire the right data, and focus on providing answers faster. More importantly, make sure you have the tools in place to speed up the acquisition process so that you can focus on getting results faster, and more actionable insights.
