Zayo unlocks a trove of data to deliver business value

Zayo, a leading communications infrastructure provider, has saved millions of dollars since implementing a new data strategy 18 months ago. In an interview with TM Forum Inform, Zayo Chief Data Officer David Sedlock shared how the operator is building a foundation for delivering value from data and why it is essential to get the entire organization involved in the process.

Zayo Group was founded 17 years ago and has made 49 acquisitions to create a fiber network spanning 18 million miles, providing connectivity services to telcos and enterprises in North America, Europe and Africa. The company went public in 2014 and was taken private in 2020 when it was acquired by investors Digital Bridge and EQT for $14.3 billion.

The operator’s new management soon saw a glaring oversight – they had no data strategy.

“We put everything into Salesforce… We had everybody accessing anything they want, whenever they want, and there was no single version of truth that people could report on,” said Sedlock.

When Sedlock joined Zayo from AT&T in January 2023, he had a mandate from the boardroom to build a data strategy from the ground up, including the architecture, platform, models, processes and governance. He described the program as “greenfield in a brownfield environment.”

All together now

A basic principle that underpins Zayo’s approach to data is that every part of the organization needs to participate.

“This is not a data transformation. This is an overall business transformation. You have to have ownership of the data within the business, and the business has to come on this journey with us,” said Sedlock. Otherwise, the data framework will look good but won’t be used effectively.

He likened it to building a house but without putting any thought into hooking up the electricity, gas and water. It is just an empty construction that can’t be lived in.

Sedlock opted for a hybrid centralized/decentralized model for Zayo, because “it’s the only one that is really successful.” His team acts as a center of excellence that creates the foundation and delivers the tools and capabilities, and then “people within the business utilize the data in the most effective and efficient ways.”

The building blocks

To get the best data and ensure buy-in across the group, Zayo had each department nominate a “data steward” to work with Sedlock’s team. The data stewards are responsible for identifying, defining and owning their “critical data elements”.

These are the pieces of information that are required to run the business, such as customer name or job ID. “If you do not complete it accurately, completely and timely, the person downstream from you cannot do their job and we cannot run our business,” he explained.

Zayo has so far identified more than 800 critical data elements. They are catalogued in “data libraries”, which is a way to manage source data coming onto the Snowflake data platform so that it can be reused.

Here, Zayo uses TM Forum’s Business Process Framework (eTOM) and Information Framework (SID) standards.

“We decided to fully adopt the TM Forum data model … as opposed to making something unique to us,” said Sedlock, noting it was important to use industry standards to “normalize” data across the company especially since it had grown through many acquisitions.

Ultimately, the critical data elements are source data and are combined to create “business ready data sets” or what Sedlock calls Zayo’s “gold data sets” within the enterprise-wide Snowflake environment, which will be used for everything, including the operator’s dashboards, reporting, data science, analytics, artificial intelligence (AI), and generative AI (GenAI).

“Everything will be going against it, because we know that we’ve gone through a standard ingestion process into data libraries; we’ve identified the critical data elements; we put the controls in place to make sure all the data is clean; it lands in these data sets and that’s what we use. It creates a single version of truth of our data in Snowflake for each one of these data sets,” he said.

Data quick strikes drive early results

Tapping data is already delivering cost savings and revenue recognition for Zayo. While the data platform is still being built, four members of Sedlock’s team have been working in parallel on “data quick strikes” to tackle some of the operator’s biggest challenges.

Zayo’s quick strikes have one to three weeks to “see a pathway to value,” and if not, the team moves on to the next one, he explained. The team has run quick strikes on churn, revenue assurance, and service delivery.

For churn, the team developed a predictive model to replace manual spreadsheet work that has resulted in millions of dollars in cost savings. A revenue assurance quick strike enabled the operator to realize millions more of revenue.

“We can deliver to the business so that we can have a return on investment, that turns [my department] from a pure cost to a profit center, while we’re delivering the core capabilities that we have never had before,” he said.

No more “Mr. No” to AI

When it comes to AI, Sedlock said he has been “Mr. No” until recently because the operator wasn’t ready. But now that the foundational capabilities are built, the operator has “a trusted data source” on the centralized Snowflake platform and it can “point AI and GenAI capabilities against those trusted gold data sets.”

Zayo has “numerous” AI proof of concepts and has connected to two marketplaces to leverage various large language models. But he said the operator is “going slow to go fast.”

Looking ahead, Sedlock said he is focused on stabilizing and running the data platform, delivering as many gold data sets as fast as he can, and leveraging those data sets for data science, AI and GenAI.

“That’s the cool stuff, but it’s not the real stuff. The real stuff is how we use those gold data sets to enable the business. Cool tech doesn’t run a business. It has to be applicable to supporting the business,” he said.