Organizations from all industries are migrating to the cloud to leverage their digital strategies. The cloud offers nearly limitless storage and compute, spurring a move of applications and data to platforms like AWS, Google Cloud and Microsoft Azure from on-premise systems.
To help companies make the most of a migration, Evren Eryurek, Director of Product Management at Google Cloud, joined Tamr Chief Product Officer Anthony Deighton for a discussion on how to successfully navigate a migration to the cloud. We’ll highlight some of their key takeaways in this blog. You can watch the entire conversation below.
Incorporate business objectives into your cloud migration
A migration offers the opportunity to address the legacy data problems (think siloed data, incomplete records, and duplicate records) that impact data quality and prevent organizations achieving their business objectives.
“For data to live up to its true potential, we have to think about improved data quality as we’re moving to cloud. If you’re going to put a bunch of data in the cloud without understanding it’s value or what it contains then you’re going to end up having the same problem in the cloud that you had on-prem, lots of disconnected data that is leading to inaccurate insights,” Eryurek said.
Ultimately, organizations want to achieve business outcomes, like obtaining better customer insight to boost sales or driving savings by optimizing their supply chain. Identifying those objectives, what data you have to support them, and how the migration can address any data gaps around those reaching those objectives aligns your move with your business goals and leads to a faster return on your investment.
“If you don’t align your migration to the business, you’ll have misguided decision making, which is detrimental to entire business units. Marketing will target the wrong people and you won’t have the right sales forecasts. A good data strategy is first and foremost to helping solve these problems,” Eryurek said.
Leverage the cloud and machine learning to dramatically improve data quality
Relying on traditional master data management (MDM) solutions that require extensive rule writing made sense when this technology debuted since data was primarily stored on-premise, there wasn’t a lot of it, and the cloud didn’t exist yet, said Deighton. But this antiquated approach to managing data won’t scale to accommodate the substantial volume of data companies now handle. This is why modern MDM takes a machine learning-first approach to data mastering.
“What’s powerful about a machine learning-based approach is that it takes advantage of this large amount of compute and also scales gracefully. So as data changes, the model adapts and learns. With machine learning, the more data you throw at a solution, the more efficient it becomes. People are really excellent at looking at, for example, customer records and saying, ‘Yep, these are the same customer’, or ‘No, these are different.’ But they’re not good at doing this at scale,” he said.
Humans will still be in the loop, he added and will train the machine learning model by providing feedback, which will have the greatest impact on improving data quality.
Eryurek also sees people playing a role in handling more valuable tasks like improving model performance.
“The smartness will still come from us [humans]. Machines can do a lot more crunching than we do, but they can only do that. Our roles are crucial as you interact with these machine learning algorithms,” he said.
You’ve migrated to the cloud. Now what?
Make a note – moving data to the cloud isn’t a set-it-and-forget-it situation. Companies are continuously adding data, making continuously improving data quality a fundamental component of reaping the benefits of a cloud migration.
“It isn’t a one-time move. What’s more likely happening is that more data is showing up all the time. You’re making acquisitions, starting new product lines, onboarding operational systems,” Eryurek said. “And also the habits don’t die fast. Even if you’re on the cloud, you will continue to operate the way you did for the past 30 years. That means you might not put data in the right place or label it properly. That’s why improving data quality should be a continuous process and part of your data strategy.”
Continuously improving data quality is also a good strategy for companies that have already moved to the cloud.“There are a lot of chief digital officers that I deal with who say, ‘I need to figure out a way to bring insight from the data. I don’t even know where my data is, what I have in my data. And all my data is in there either in the cloud or in a hybrid model. There are tools like that can help with this and Tamr plays a significant role in helping [Google] provide a solution,” Eryurek said.
Don’t let uncertainties prevent your company from embarking on its cloud journey. Cloud providers embarked on the migration path over a decade ago with their customers and are happy to share the knowledge gained from those experiences.
“The journey might have twists and turns, but don’t be afraid. Let’s begin the journey with your data. There’s no wrong way of starting. It’s not going to happen overnight, and we will make mistakes, but we are here to share all the lessons learned,” Eryurek said.