In a world where terabytes of data are generated every second, companies are struggling with an important choice: data quality over data quantity! Data enthusiasts argue that the more data we collect, the greater the chances of discovering hidden patterns, predicting future trends, and making well-informed decisions. Yet, a growing chorus of skeptics contends that the sheer volume of data could, in fact, become a stumbling block rather than a stepping stone to progress.
So what should companies focus on? More or better data? We speak with Libba Stanford, a strategic data advisor with 20 years of experience. She shares with us her insights, experiences, and strategic steps companies can take to make their data more meaningful and purposeful!
Let’s dig in.
What Does Good Data or Quality Data Really Mean?
The answer to this depends on who you’re asking! If you’re talking to a data scientist or a data governance type of role, they’re gonna be focused more on the data meeting the internal standard. If you’re talking with the C Suite executive, then they may consider good data to be accurate and timely and relevant business intelligence that they can use to make important business decisions. And so with each role, good data is going to mean slightly different things, but overall good data is talking about cleansing that data and answering the main question that the data attempts to answer.
According to Libba, better data simply means understanding and fulfiling the context of what are you trying to accomplish. Libba warns against calling someone’s data dirty until we don’t understand the context. Oftentimes, there can be conflicts and disputes over what classifies as missing or unavailable data. For example, a data scientist working with Libba thought she had dirty data when one of her data sets did not contain data about a particular attribute. The data was simply not available due to external factors, however, the data scientist presumed it was “missing.” Understanding this difference is crucial to setting governance standards and creating a data quality strategy.
What Are the Common Misconceptions About the Notion that More Data is Better Data?
Libba answers this question about data quantity vs data quality with an interesting story about yield monitors in agriculture tech.
According to Libba, ‘some yield monitors on combines traditionally collect data at one-second intervals. That’s a lot of data. And so, you know, you’ve got a small field that you’re already collecting maybe 10,000 data points on and it’s dirty and not just from the dirt, it’s dirty data. In 2018, John Deere combines decided to start collecting and putting out data, at five Hertz, which means we went from one Hertz to five Hertz and 0.2 2nd data. Overnight, we went from 10,000 data points to 50,000 thousand data points and there was no immediate value in what we were getting! The biggest problem was there wasn’t any more quality control of that data in the collection. And as a matter of fact, it was skewing the results that we were seeing from the data for our context. So spatial integrity of that data was critical to us, and now we had data all over the place and it may be right or it may not be right. It caused a lot of bottlenecks and it wasn’t helpful.
As you see, more data (without control, context, and purpose) does not necessarily mean better data!
How Do We Go from More Data to Better Data and help people recognize context?
Companies push for more data without addressing problems. They simply don’t speak to frontline workers collecting the data, and sometimes forget about the whole context of the data.
Citing the example of farmers, Libba believes data collection doesn’t solve farmer problems. These farmers have to make plenty of decisions and while they get the data, they don’t know what to do for a season. The growers or farmers personally think about marketing, banking financing, chemical interaction, and many other business decisions! For example, they get data about losing 15 bushels in an area due to spots in the field holding water, but what do they do about it? The problem remains even though there is data!
Meanwhile, frontline workers collect and experience the data, and have that full picture of information that companies can question and pull answers from and better strategize and build out a plan. However, there is very little communication with frontline workers. Again, there is no problem-solving. The focus is merely on process or information improvement.
How to persuade key stakeholders to shift from more data to better data for problem-solving?
One of the biggest challenges to persuading key stakeholders lies in the fact that they are very disconnected from actual decision-makers. Anything that they are saying gets heavily filtered as it goes through those communication channels, leading to organizational issues and confusion. In order to persuade key stakeholders, you have to sit down with them to understand the context of their data before making any decision about it.
Libba shares her corporate experience where failing to communicate with key stakeholders in the marketing team led her to make decisions that created bottlenecks. Scenarios like this are hardly new. Data strategists, IT managers, and business managers such as the marketing and sales team share the same data set (for example, customer data), yet, they work in isolation! The result? Decisions that negatively affect the revenue and profitability of the organization!
Therefore, one of the ways to persuade key stakeholders is to cultivate a culture shift! Stakeholders need to communicate with each other and understand the implications of bad data. They first need to accept that if they’ve got data, it is bound to be bad, and they need to fix this bad data and maintain its quality. If not, bad data has a cascading impact – this can result in incorrect follow-up information being incorrect, lack of trust from employees, and potentially as a result, lack of trust from your customers.
What role does technology play in facilitating the shift from more data to better data?
Libba answers this question with an amazing analogy about cotton!
We highly recommend you watch the video at the 40:02 time stamp to hear how Libba talks about how technology facilitates this shift.
In essence, Libba urges companies to focus on the data itself first and what needs to happen. When companies focus on the technology first, they end up with problems like multiple customer databases that do not share any cohesive value!
Wrapping it Up
The summary of this amazing webinar in five key points:
✅ Understand the context of the data
✅Use the data to solve a problem (don’t just collect data)
✅Prioritize communication with key stakeholders specifically when it comes to data quality
✅Talk to frontline workers to understand their experience with data
✅Use technology to assist with data goals instead of just creating/collecting data assets
We hope this webinar helps! If you’d like to know how we can help you fix your data, get in touch 👇