While enterprises are making progress on managing data, much of it remains siloed and unstructured and needs to be better managed to help digital transformation efforts progress. Data fabrics can help, according to officials from Boomi, a provider of integration platform as a service (iPaaS), speaking during a session at Gartner’s Data & Analytics Summit Wednesday.
“We’re still in this world of significantly siloed stores of data in this digital revolution that’s happening of new apps and use cases being spun up,” said Sean Keenan, director of data product management at Boomi. Every enterprise is wrestling with how to collate data and bring it together, he said.
Gartner has defined data fabric as a design concept that informs and automates the design, deployment and use of integrated and reusable data objects regardless of deployment platforms or architectural approaches. Data fabrics let users build a flexible, agile, scalable architecture that will be able to supply data to humans or machine users.
According to a figure Boomi cited from MIT’s Center for Information Systems Research, 51% of organizations have their data in silos. Further, among the findings of a Harvard Business Review study, less than half of an organization’s structured data is actively used in decision making and less than 1% of unstructured data is analyzed or used at all.
The right data is critical for digital transformations
The relationship between data and digital transformation remains a major challenge for CIOs, said Myles Suer, principal data project marketing manager at Boomi, who also spoke. The goal is to eliminate issues around moving data from source to target, and sharing and integrating everything so users will have a common view and be able to do things with data, he said.
“Even structured systems present a ton of challenges because there’s so many of them,” due to the explosion of SaaS apps that companies are running their businesses on today, Keenan said. Companies continue to wrestle with solving that problem, as well as getting any analytical outcomes on unstructured data.
SEE: Why your company needs a chief data officer (TechRepublic)
Eventually, Keenan said he believes unstructured data will have its place and progress can be made around harnessing that as well. But he noted another troubling finding from the HBR study was that more than 70% of employees have access to data they should not have access to.
The amount of time people spend wrangling data and data hygiene has given rise to the expression that it might be more appropriate to call data scientists data plumbers, Suer said. All of these issues are impacting the ability of organizations to transform.
Better data can be used to achieve transformational objectives that include customer experience, business ecosystems, optimized processes and new business models, Suer and Keenan said.
When Keenan talks to customers, he said every organization wants to deliver better products and services—and new products and services and better customer experiences.
“Frankly, customers and consumers are demanding that in this digital world we all live in. Data is the driver of delivering those things,” which should motivate companies, Keenan said.
The traditional data warehouse approach lets people extract whatever data they had. The paradigm of leveraging data before it was modeled perfectly in some sort of dimensional or snowflake model “was great for its time, but lacked a lot of agility,” Keenan said.
SEE: Analytics: Turning big data science into business strategy (ZDNet/TechRepublic special feature) | Download the free PDF version (TechRepublic)
Now, “what we’re seeing organizations do is getting people closer to the raw data faster and allowing them to iterate and understand what data is and [apply] some context upfront and then let them make some quick decisions on how they want to leverage it.”
Getting to that data in a more agile fashion and allowing users to iterate on it “and see what nuggets of gold are in there is definitely a trend we’re seeing,” Keenan said.
Chief data officers should create a single source of truth and improve data hygiene, Suer said. They should also look for platforms that offer end-to-end capabilities to manage data from discovery to outcomes, he said. Data fabrics enable a set of data ops services.
A data fabric platform should help:
- reduce the time to discover and evaluate data
- reduce the time needed to standardize the data and create data governance
- create a single correct view into company data
- simplify, consolidate and enrich the data
- facilitate better data stewardship processes
There should also be the ability to create APIs so future users can use the data as well, Suer said. If a CDO has set up stewardship, they should know the rules around PII and make sure it is governed.
Enterprises will find value in having one platform with a suite of integrated, end-to-end services, Keenan echoed.