Sergey Nivens - Fotolia
Iron Mountain CTO Fidelma Russo is a veteran digital transformation leader who not only works with enterprise clients on their transformation journeys, but identifies trends and opportunities for her own company's digital business.
Russo, who joined Iron Mountain in March 2017, is also executive vice president of global technology and operations. She previously served as the company's CIO until May 2019.
This broad professional experience has given her insight into those factors that can help ensure success as organizations remake themselves in this digital era. For Russo, it's all about the data in digital transformation and she explains her perspective below. The interview has been edited for length and clarity.
Of all the elements that go into digital transformation, why is data critical?
Fidelma Russo: Over the past few decades, companies have been amassing large amounts of data in many formats throughout their organizations and in lots of data lakes. When you start a digital transformation [initiative], you have a lot of decisions to make around which processes should be transformed, how they should be transformed, what should be retired [and] the risk of not transforming.
In my opinion, you need good data [in digital transformation] to make good decisions about the extent of opportunities you have in your business. And you have to continue to use that data to make decisions on where to pivot during your transformation. It's all about data driving decisions to inform you about what aspects of digital transformation are appropriate for your business and then measuring it to get the results you need. That's why it's so high up on the list of issues.
More specifically, you think one of the biggest issues CIOs face is deciding what data in digital transformation to consider and what to discard. Why is that?
Russo: It's best to talk about an example. [What] many companies grapple with today is finding that single 365-degree view of their customers. But today, the data is scattered across the enterprise in systems and physical documents that aren't linked together. And, many times, they have conflicting pieces of data.
The customer data -- if that can be cleansed and [is] of high quality -- can be used for very valuable things like understanding the customer's buying patterns and what the customer's overall sentiment [is] towards the company.
But a lot of this data is dirty -- it has conflicts, it's incomplete, it's aged. So before you start to use [data in digital transformation], it's important to cleanse the data and establish a going-forward taxonomy.
It's also just as important to talk not just about how I clean and homogenize it, but also what data I don't need. Some of it will be discarded because it's wrong. Once you do that cleaning exercise, which is a big effort, and build that single view of the customer, you have to have a data governance program in place or you end up back where you started.
It's all about cleaning, streamlining, deleting data, putting in place your process and then putting in your data governance so new data follows the data architecture, structure, and processes and procedures that govern that data all through its lifecycle.
Regarding data in digital transformation, should companies start by just discarding incorrect or obsolete data or should they be more selective about the data the organization holds in general?
Russo: It's about all those things. The tendency over the past decade was to keep everything because you weren't sure if you would need it again. But the privacy regulations are forcing companies to really confront [the questions of] 'What shouldn't I keep?,' 'What should I keep?' and 'How long should I keep it for?'
If it's inaccurate, it's obvious you shouldn't keep it. But each enterprise will have its own opinion on whether a piece of data is important on not. That leads you to have your governance done. [It's] a cross-functional exercise where you have a dialogue about what data an enterprise should keep and what it should delete.
Isn't that one of the hardest parts of data governance -- determining the real and potential value of data?
Russo: I think it is a harder piece -- understanding what's valuable and what's not valuable -- than determining if something is inaccurate or duplicative. A number of years ago this used to be the purview of the IT organization, but it's really a business problem and it has to be informed by the business.
We [at Iron Mountain] have instituted a data governance program where we have data ambassadors from each part of the business who come up with the rules around what information is valuable and what isn't. We also have regulatory rules we have to comply with, [and] customers who have policies and procedures that, in certain cases, we have to comply with. But what data is valuable and what data will [give] your organization a competitive advantage? That has to be informed by the business.
But are businesses taking the initiative to do that work?
Russo: This is where everybody is struggling right now. Some of it is [that] you have to have an impetus for change. A couple of things are going on that are driving people to change. The regulations, [for instance], are forcing people to change. Second, the move to cloud and this hybrid world of on-prem and off-prem is forcing people to think about what to move, what to keep and what to discard. It forces them to think about how long [to] keep this data, where [to] keep it, when was the last time [it was] touched and is it still relevant?
It's an evolution. For a long time, it was easy to just keep everything.This is a much more nuanced decision. Sometimes you might make a mistake, but I think systems and people begin to learn and they'll get a lot smarter about the areas where they want to have long retention rates and where they can delete data after 30, 60 or 90 days.
Shouldn't an organization keep data it has collected even if it doesn't see its value yet? Perhaps it will identify the value in the data at some point.
Russo: It's a little bit like your house -- you can't keep everything even if you want to. If you think [you might] lose a piece of data that might be critical for a breakthrough, you have to ask 'Is this data elsewhere?' [For example], if you think about AI and machine learning, it's highly unlikely if you get rid of data that there's no other piece out there that will help you train an AI model. I think you're talking about the exception rather than the rule.
I think the risk of deleting that one piece of data that's going to turn your business on its head is low. It's understanding risk, understanding whether it's a unique piece of data in your organization or it's just like everybody else's data, [in which case] you don't need to keep it all the time.
But isn't storage cheap enough to keep more data than needed?
Russo: Yes, storage is cheap. But to use that information and process and govern [it] -- all of that is quite expensive. AI and machine learning, using APIs from the big vendors with the different platforms -- there's a cost component. Data at rest is cheap, but data in use is not. Making sure you have a well-managed data architecture is a critical piece of digital transformation. And the CIO plays a critical role pointing out the cost and the effectiveness and what platform needs to be put in place.
Does 'discard' just mean delete?
Russo: It means delete [or] securely destroyed. It's secure deletion. You may have some [data] that you're not really sure about and an option might be to send it to offline tape storage and set the timer. If you don't touch it in a couple of years, you get rid of it.
Who drives this data in digital transformation conversation?
Russo: I think it all comes from the top. It comes from the CEO who wants to digitally transform and then it pipes down to the leadership team who understands that data and data management are important parts of that transformation.