Wendy Zhang, global data & analytics leader at a biotech company shares insights into addressing data quality challenges and building a process for better data by investing backward.
Wendy Zhang, global data & analytics leader at a biotech company shares insights into addressing data quality challenges and building a process for better data by investing backward.
Wendy Zhang, global data & analytics leader at a biotech company shares insights into addressing data quality challenges and building a process for better data by investing backward.
Wendy Zhang, global data & analytics leader at a biotech company shares insights into addressing data quality challenges and building a process for better data by investing backward.
D.D.: Wendy, you have spent many years leading and driving analytics initiatives at scale. From your experience, what common data quality challenges have you identified?
W.Z.: I think there are two important data quality challenges that people experience, and they are definitely related:
What I hear a lot in the industry is that there are a lot of buzzwords thrown around and some people are tempted to think that adopting a new piece of technology or AI/ML will solve all their problems, without realizing that there are a lot of data quality steps you have to take before to get yourself set up. This is one of the main reasons why many digital transformation initiatives don’t work.
Everybody knows the phrase garbage in, garbage out. I would actually argue that it’s even worse than that: garbage can lead to a disaster. It’s similar to building a house: if the foundation is wobbly and unstable, it's not going to last. Another analogy I use: everybody wants the penthouse, but no one wants to build the groundwork.
Getting the data foundation right can take a lot of time. It is a detailed and oftentimes tedious or unglamorous task.
More importantly, it’s not a one-off project, it’s a process. It doesn't have a certain end date - as your business grows, you have more data, and your company is forever evolving. As we collect more and more data, it will have a trickle-down effect on how to manage it and keep your business rules up to date.
So, the question arises: how do you monitor data that when you're supposed to do something new? Is it going to be completely revamped or is it going to be periodic monitoring adapting to the changing world?
At some point, the data quantity changes are going to lead to something qualitative and of sizeable substance.
W.Z.: Well, it's not a situation that will go away or improve on its own. And no technology is going to solve that problem alone for you. We have to do it now because the more data gets piled on, the greater the challenge becomes. If you’re not solving the root cause problem, all data quality challenges are only going to be amplified down the road, and it will be harder and harder to address them later. It grows multifold, especially as the scale and the complexity of data grows exponentially. This will impact your business processes, your people, and your bottom line. The longer you wait, the longer and harder it will take to address those issues, and the more profound will the business impact be.
The longer you wait, the longer and harder it will take to address data quality issues, and the more profound will the business impact be.
D.D.: In the light of the pandemic, many organizations have accelerated digital transformation. However, new technologies like AI and machine learning hinge on the ability to harness data of the right quality. Why is data quality important for successfully leveraging these new technologies?
W.Z.: Data quality is definitely a very important factor for accelerating digital transformation.
During the pandemic, we all found new ways of working and had to adapt to keep the business running. At the same time, most of us witnessed an awakening moment: the old ways that have worked well before, are not working so well anymore.
We have experienced digitalization in the digital age first-hand. In turn, this development brought some data deficiencies and inefficient systems into the spotlight.
As a result, we are now seeing a wave of companies trying to do digital transformation. And this involves
Data quality unfortunately still comes as an afterthought. It’s not part of the digital transformation process from the onset. We're always surprised by how messy our data is and how data of questionable quality slows down processes and leads to wasted financial resources time and time again.
Imagine you're well on your way from a technology standpoint, from the system and architecture standpoint, but your data is not ready. When you’re e.g. migrating data, you cannot use the same old processes of lifting and shifting data in the cloud. If you’re using the same processes, you’re going to deal with the same challenges.
Data quality unfortunately still comes as an afterthought. It’s not part of the digital transformation process from the onset. We're always surprised by how messy our data is and how data of questionable quality slows down processes and leads to wasted financial resources time and time again.
W.Z.: People definitely play a key role in the digital transformation investment. You need the right people with institutional knowledge to help you actually make that digital transition. Given that there is a talent war going on, we need to start building our talent pool to really do digital transformation. We need to find people with the right skills and experience, and that can be a challenge.
In the light of the ongoing pandemic, many marketing and analytics professionals have been reassessing their priorities, careers, and lives. As a result, it tightens the eligible talent pool in the market and it becomes really hard to find the right talent.
Once you find them, a company needs to help them grow and continue learning, so that they are engaged and motivated to drive the digital transformation process successfully.
D.D.: Indeed. So, for a digital transformation process to be successful, you need the following 3 elements: the right people, the right technology, the right data foundation.
W.Z.: Precisely. Technology is the enabler. It needs to work as intended: efficiently, effectively and economically feasible. People are the ones driving everything: how you design and redesign your processes, how you treat your data, and how you implement your technology. And data is the foundation holding everything together.
D.D.: Reaching 100% data quality is not a realistic goal. From your experience, what layers of data quality (e.g. accuracy, completeness, relevance, timeliness) would you recommend focusing on and why?
W.Z.: Of course, in a perfect world, they're all important. If we could achieve all the dimensions of data quality, that would be ideal.
I started my data career in the Air Force, and we were also always trained to make sure that data was current, complete and accurate. In that context, those three aspects are very very important.
When I moved to the financial services industry, our mantra was that data should be fit for purpose. This was a combination of having data that is current, complete and accurate.
In my mind, relevancy is the most important aspect of data quality. You can have accurate and timely data, but if it’s not relevant for what you're doing or fit for purpose, then it’s not really usable. You cannot really leverage it as it doesn’t serve your goals. So, I would say that both relevance and finding the data you need for the purpose are probably the most important aspects.
You can have accurate and timely data, but if it’s not relevant for what you're doing or fit for purpose, then it’s not really usable. You cannot really leverage it as it doesn’t serve your goals.
Are you struggling with exploding cost, rapidly growing headcount, long and tedious processes, and frustrated stakeholders? Learn how to turn your digital analytics from cost to profit driver, from strategy to implementation.
D.D.: You’ve identified an effective approach to improving data quality: investing backward. What is it and how does it work?
W.Z.: When we think about it, most projects are always forward-looking, no matter the industry. At the same time, there are also many legacy systems running in the background, unless the company is completely new. This leads to a lot of misalignment, disparate data systems, and challenges with data integration.
I often call this data debt: we might not be completely driven by data, but it’s influencing our processes and the people working with data.
What I often find is that people are reluctant to deal with the historical buildup of data and they are more inclined to start something new. There’s a risk about embarking on a new data project that can take years to be complete and reap the benefits. Getting the data ready is just the first step. There is still substantial work to be done when it comes advanced analytics with that data.
Investing backward ties into the mindset change that you need to address the root cause to win. You need to address some of the underlying data issues to get the short-term quick wins that you want.
The challenge that arises here: what can you do in the meantime to utilize some of the benefits right now, not years from now? One solution could be investing backward. This strategy helps you take advantage of some of the benefits that are already on the table by going back and addressing and it might solve the historical data quality challenges that you have identified.
This process often gets overlooked, as it’s not as glamorous or exciting as starting something new. It involves a lot of effort, and it might not get you a promotion. However, it is necessary.
Investing backward ties into the mindset change that you need to address the root cause to win. You need to address some of the underlying data quality issues to get the short-term quick wins that you want.
Wendy Zhang is the global data & analytics leader at a biotech company. She has spent the last 15 years helping Fortune 500 companies and Federal agencies deliver business results through building and leading global analytics teams and driving end-to-end business-critical analytics initiatives at scale. She is passionate about leading enterprises to become truly data-driven and drive actionable insights by fostering data as an asset and an analytics-enabling capability.
