8 tips for streamlining legacy IT

Despite the acceleration of digital roadmaps and the continuing push to modernize, many CIOs still have old applications and infrastructure in their IT stacks.

Organizations have a few key reasons for holding onto legacy technologies: They’re stable, they’re inherently efficient, and they’re paid for, says Abhi Bhatnagar, a partner at management consulting firm McKinsey & Co. He worked with one company that had opted to retain a massive mainframe state, noting that “it’s well-run, highly homogenized, and a hugely customized, fully depreciated state.”

At the same time, CIOs may find that replacing some legacy technologies won’t yield high enough returns to make undertaking that modernization work a priority.

“There’s a broad brush that it all needs to be moved to the cloud or SaaS. But the reality is that not all of your portfolio needs to be modernized, especially if it’s stable and efficient as it is,” Bhatnagar says.

Whatever the reason, many CIOs still retain legacy applications and infrastructure.

One of the challenges then is to streamline those legacy technologies, to make them work at peak performance, to deliver optimum benefits at the lowest costs and with the lowest risks. Here are eight tips for doing that.

Adding automation

Automating pieces within a legacy environment is one way to bring efficiencies by reducing manual efforts, limiting or eliminating errors, improving quality, increasing data consistency, and upping productivity.

“Whether Unix, Linux, mainframe, or three-tier [architecture], there is room for automation to come into it,” says Juan Orlandini, chief architect at Insight Enterprises, an IT services firm.

As he explains, many of the business processes supported by legacy systems contain segments that are highly structured, repetitive tasks. Similarly, there are pieces of the IT work itself — the blocking and tackling–type tasks such as storage provisioning and backup — that feature that same repetitiveness.

Orlandini says it’s important to assess the processes targeted for automation and streamline those before automating them; as he points out, automating bad processes won’t bring the hoped-for benefits.

He also cautions IT leaders on expectations, saying that automation will bring speed to legacy systems and the processes they support, but it’s not a panacea nor a replacement for full-scale modernization.

Borrow principles from cloud environments

Orlandini says experience has taught him that simply moving legacy systems as-is to the cloud — the “lift-and-shift” approach — does not typically pay off.

“It can be an option, there are times when it’s appropriate, but our data and experience show that it’s typically more expensive [than leaving it in place],” he says. “So if what you’re trying to do is optimize your infrastructure, it’s better to do other work to get a good return for your efforts.”

One such option is to take some of the principles of managing cloud and apply them to managing legacy. Take storage provisioning, for example. Cloud has predefined allocations to streamline the process; administrators can do the same when working with on-premises resources.

“On the face of it, it seems like it might be a waste of [storage] resources. Yes, you might waste storage space. But in the end, it saves you so much human time because you’re taking some of that processing out of the equation. And storage today is cheaper than human resources, so you actually gain,” he says.

Implement a product-oriented approach

CIOs could similarly borrow from modern management techniques and adopt a product-oriented approach to legacy tech, Bhatnagar says.

This helps IT and the business keep focused on business outcomes — whether it’s speed, efficiency, better responsiveness, or innovation. With that focus, IT can zero in on how to structure and optimize legacy technologies for those outcomes, he explains.

A product-oriented or portfolio approach also gives the technologists who support the legacy technologies better visibility into problems impacting those desired outcomes.

“They have a good sense of the environment, what are the risks — whether it’s a compliance risk or speed — as well as the business value that the platform drives,” Bhatnagar says.

Consequently, IT teams can then identify and advocate for the best ways to add the capabilities necessary to achieve desired results.

“That’s where the mindset of product management comes through,” Bhatnagar adds.

Furthermore, this approach helps IT and the business work together in articulating the value that modernization will deliver, when the time comes for that move.

Peel off pieces

Organizations that still depend on legacy technologies often do so because they handle so many core business processes, making modernization complex and risky.

But those organizations can still improve IT performance by peeling off pieces, says Balaji Raghavan, a principal consultant for banking, financial services, and insurance at Tata Consultancy Services.

“All of this can be done in a tactical timeframe,” he adds.

For example, a company with a legacy financial system supporting multiple functions can keep it in place to run those pieces that are very stable, such as accounting. But it can take out pricing and billing, both of which can quickly change due to market and customer dynamics.

This approach, Raghavan says, moves companies further forward on their modernization journeys while also reducing complexity within the legacy that remains.

Kathy Kay, CIO at Principal Financial Group, took a similar approach around systems supporting processes in her company’s life insurance division. She says consolidating them in a single modern system wouldn’t bring a return that would justify the costs, but she still saw areas around how users interacted with the systems that needed improvement. So she targeted those user-facing components, using APIs to connect those new capabilities to the legacy system.

“Where we could automate, we automated, and we reduced paper processes, but at the core there is still that legacy. They’ll run their useful life and as those books of business dwindle, we’ll just sunset those systems,” Kay adds.

Have a talent strategy

In April 2020 New Jersey Gov. Phil Murphy made a plea for technologists who knew COBOL to work on the state’s legacy unemployment system that was overloaded as the pandemic and the resulting job losses took hold.

Don’t scoff: New Jersey is by no means the only entity relying on the 60-year-old programming language. COBOL is pervasive, with estimates of more than 200 billion lines of COBOL code still in use — and not going anywhere fast. Micro Focus, a software and IT company, found in its 2020 COBOL Survey that “70% of enterprises favor modernization as an approach for implementing strategic change as compared to the replacing/retiring of key COBOL applications as it continues to offer a low-risk, and effective means of transforming IT to support digital business initiatives.”

With COBOL and other legacy technologies sticking around, CIOs must have people on their team with the skills needed to keep them working and in tip-top shape, Orlandini says.

“Increasingly what we’re seeing is the older technology, which is still highly important to the business, is having a brain drain because the staff wants to do new fun stuff and the new workers don’t know the old skills. But it’s important to encourage some of your staff to maintain skills in [legacy technologies]. So don’t make that look like a bad thing or career suicide,” he adds. “You need someone who knows how to keep it humming, a well-trained person who knows how to keep it well-oiled and can fix it when it breaks.”

Apply modern approaches to old technology

At the same time, though, Kay says CIOs should use contemporary methodologies, skills, and thinking to find ways to improve legacy infrastructure and systems.

She has upskilled her employees who work on legacy tech so they’re not left behind as everyone, including her own IT shop, eventually moves onto more modern IT environments. But then, as a bonus to the company, these workers who now are skilled in both legacy and modern technologies can see options for streamlining older IT that they hadn’t thought of before.

“When you have people who have been imbedded in the legacy technology, they know the business well. And when they upskill, they can look at things differently and can quickly come up with different ways of doing things,” she says.

Kay’s company recently acquired another business, a move that required Kay’s team to migrate customer data from the acquired company into its own legacy system. Her engineers leaned on their new skills to devise more modern integration capabilities that proved much more effective, efficient, and scalable than the designs they’ve used in the past.

Move out the data

Like most CIOs, Kay is working with her executive colleagues on strategies to improve the customer experience. That requires the ability to access and analyze the right data at the right time, so that the company can deliver more personalized experiences and make recommendations for services based on each customer’s own needs.

But some of the needed data rests in core systems. Kay says she plans to modernize those systems but it’s a multiyear journey.

So, in the interim, the company extracts the needed data from its core legacy system and then moves it to the cloud where it’s more readily accessible for use in the modern analytics capabilities that support the customer experience initiatives.

“That reduces our dependence on the old system and makes it easier to sunset that legacy system when it’s time,” Kay adds.

Shut it off

Johna Till Johnson, CEO and founder of the research advisory firm Nemertes, has seen companies hold onto legacy applications and technologies because they support a business process; but when companies examine the situation, they often find that the business process remains only because the technology is in place.

“CIOs are still too close to the technology to say, ‘It’s the business process that needs to be rethought here.’ But when they do, 99 times out of 100 there is a business process that only exists because of the technology,” Johnson says.

She cites a simple but illustrative example of a business that retains a fleet of printers to support its requirement for physical signatures on documents. The business function leaders may not see an immediate need to change in that case. But a CIO trying to improve IT operations has a big incentive to streamline those legacy printers even if they’re not on the modernization priority list.

“Today there are ways to get the same benefit of wet signatures without needing printers anymore. So CIOs might as well give the business a little bit of a shove and get rid of them sooner than later,” she says.

She notes that it’s a simple example but “one everyone can see, where you don’t streamline the legacy technology, you streamline the process.”

She adds: “The simple recommendation I have is just pull the plug. Don’t waste any more time and energy on these legacy technologies [supporting legacy processes]. If it’s so far down the priority list that it’s not worth modernizing, then turn it off and see if anyone notices.”