For Jo Abernathy, CIO at Blue Cross Blue Shield of North Carolina (BCBSNC), it was time to just walk away from the healthcare company\u2019s aging IBM Db2 databases running on AIX. \u201cWe decided to prioritize the elimination of some pervasive technologies that have become too expensive relative to comparable products, and where support was lacking,\u201d she said. Db2 has been expensive to license. Also, identity and access management has been problematic, and support for AIX today is a specialized skill set. Most importantly, they needed good support. \u201cWe haven\u2019t gotten that,\u201d she says. \u201cWe needed to do something different.\u201d\n\nThe key to deciding how long to keep a technology lies in figuring out where you get the best return on investment, and while moving off Db2 might sound like a no-brainer, there\u2019s nothing easy about the decision process. \u201cIt\u2019s not like you can jam all the factors into a machine and it pops out the answers,\u201d she says. \u201cIt\u2019s very situation-specific.\u201d\n\nLinda Ivy-Rosser, VP and research director at Forrester, knows all about dilemmas like these. \u201cMost CIOs are managing cost by maintaining the core \u2018as is,\u2019 just trying to hold on and stay within that \u2018if it ain\u2019t broke don\u2019t fix it\u2019 mantra,\u201d she says. \u201cBut even if you decide to retain legacy systems, that doesn\u2019t mean you do nothing. To bring growth in, some level of modernization has to occur, and you still have to pay down technical debt.\u201d\n\nHere\u2019s how other IT executives tackle the problem.\n\nThe weight of tech debt at UC Riverside\n\nReducing technical debt was CIO Matthew Gunkel\u2019s objective when his team decided to migrate the University of California Riverside\u2019s ERP system to Oracle Fusion Cloud Financials. \u201cWe had 90 custom applications that supported our finance organization,\u201d he says, that were originally built using PL\/SQL and the Grails Framework. \u201cWe spent a lot of time supporting those applications on a legacy codebase. It was decades old, badly commented, and didn\u2019t follow modern coding and development paradigms.\u201d Developers willing to work on the system were hard to come by, and when someone new did come on board, it would take six months to get them up to speed. \u201cWe didn\u2019t have enough people or enough time,\u201d he said. Today, as part of the migration, his team reduced the number of supported applications to just 30.\n\nThe scarcity of talent that can support business-critical legacy systems can block growth, and is a good reason to consider retiring aging technology, says Ivy-Rosser. \u201cWe see a lot of that with our banking, insurance, and actuary clients.\u201d If migration isn\u2019t feasible, another option is to work with a service provider that can provide a large pool of legacy talent.\n\nPart of the code review process was getting developers to think differently about how they worked in the old paradigm at UC Riverside and how they might work differently in the new one. For example, the team initially estimated they would need a month to rewrite an application. But after carefully reviewing the code and streamlining the application, they ended up doing it in a week. \u201cThey were able to eliminate a middleware piece that was adding complexity,\u201d he says.\n\nAny opportunity to pay down debt, repurpose, and make strategic investment decisions that give the highest yields should be looked into, says Ivy-Rosser. But Gunkel says other legacy systems aren\u2019t worth the risks to update. For example, researchers built industry-leading climate models in Fortran. \u201cIt\u2019s very cost ineffective to rebuild that, and you run the risk of causing some change that shifts the model or trust in the model,\u201d he says. In cases like that, he adds, you hold onto it.\n\nEven when you have a business case for replacement, you may be at a disadvantage in terms of disentangling your systems from a technology if you didn\u2019t plan for the technology\u2019s eventual replacement up front, Gunkel says. \u201cYou need to understand the exit strategy from the beginning,\u201d he says, particularly with a SaaS engagement. Abernathy agrees. Ideally, your separation agreement and transition support were negotiated up front, she says.\n\nRegular reviews are critical to assess any changes in terms and conditions. For example, when a storage contract with Box went from unlimited storage to a capped amount, the scale and cost no longer made sense\u2014especially when Gunkel reviewed what competitors were offering. \u201cYou have to constantly analyze and understand where and how the competitive landscape is changing,\u201d he says.\n\nThe University is now moving toward annual reviews of all IT systems and software. \u201cIt\u2019s about how we can make everything work together more effectively, and where we can leverage commodity systems,\u201d he says.\n\nAnd Ivy-Rosser adds, \u201cAs you revisit the technology\u2019s capabilities, consider how it maps to desired strategic outcomes and business value, and whether the technology is incompatible with modern capabilities.\u201d\n\nBCBSNC Keeps It Simple\n\nEven when a technology needs replacement, convincing the business can be challenging. There\u2019s a tension between retiring old assets and focusing on new applications that bring additional value, Abernathy says. \u201cBusiness leaders often underestimate the complexity that technology leaders deal with,\u201d she says. \u201cIt\u2019s harder than they think, and being able to convey that without sounding like you\u2019re exaggerating or whining, it\u2019s tough. That\u2019s why we strive for simplicity.\u201d\n\nWhen reviewing products and services, Abernathy considers whether a technology still fits into requirements for simplicity of geographies, designs, platforms, applications, and equipment. \u201cDriving for simplicity is of paramount importance because it increases quality, stability, value, agility, talent engagement and security,\u201d she says. Other red flags for replacement include point solutions, duplicative solutions, or technologies that become very challenging because of unreasonable pricing models, inadequate support or instability.\n\nIn some ways, moving to SaaS-based applications makes the review process simpler because decisions as to whether and when to update and refactor are up to the provider, Ivy-Rosser says. But while technology change decisions are the responsibility of the provider, if you\u2019re modernizing in a hybrid world, you need to make sure your data is ready to move and that any changes don\u2019t create privacy issues. With SaaS, the review should take a hard look at the issues surrounding ownership and control. \u201cAnd as new services are released at a faster pace, you need to make sure if you\u2019re in a hybrid world that your operations, architecture, and infrastructure are optimized to leverage those new features,\u201d she says.\n\nAnother challenge for Abernathy is deciding between maximizing an existing technology versus getting too locked into one vendor. For example, BCBSNC uses Salesforce, which Abernathy describes as a low-code platform that makes it fast and easy to get business applications out the door. \u201cOn the one hand we need to leverage the heck out of it,\u201d she says. \u201cOn the other, you\u2019re getting into vendor lock-in, and there might be a less costly point solution that, say, AWS is rolling out. Salesforce is more expensive, but you have it all in one platform. It comes down to who\u2019s going to be the best partner. And then there\u2019s the simplicity factor. There are always trade offs.\u201d\n\nAbernathy also developed an annual budget just for refreshes and retirements. \u201cWe have Technology Lifecycle Management, a bucket of money we get every year and we use it to keep our systems running, secure and cost effective,\u201d she says. Another strategy, she adds, is to time major refreshes and retirements with the business cycle, if you can. \u201cWhen companies are having a good year, it\u2019s a good time to say let\u2019s throw money at this legacy thing and start a migration,\u201d she says. \u201cBut there has to be a lower OpEx, and the ROI has to be there.\u201d\n\nOn the other hand, some technologies that aren\u2019t delivering may not be worth replacing right away. \u201cIf the expected return for replacing something isn\u2019t compelling, especially if we know it\u2019s on a limited timeline for existence, we may let it ride,\u201d Abernathy says. Even an end of life notification isn\u2019t always enough to make her fold. In some cases, she says, they\u2019ve negotiated best effort break-fix support, used third-party support, or even gone unsupported for a period of time.\n\nCosentino follows strategic criteria\n\nSometimes compelling new features in a major upgrade provide a strong incentive to migrate, which was one reason why surfaces manufacturer Cosentino is moving to SAP\u2019s S\/4HANA. \u201cIn most cases, we make decisions based on strategic criteria to determine the right time to make a change,\u201d says group CIO Jos\u00e9 Rodr\u00edguez. \u201cThis could be driven by our company\u2019s strategy, technological advancements, user experience, or economic factors. The next migration to SAP S\/4HANA addresses several of these factors.\u201d\n\nCosentino performs roadmap reviews every six months. \u201cThe criteria for these decisions is always around alignment with the company strategy,\u201d he says. And while many systems are refreshed on a regular schedule, IT infrastructure in the factories is replaced every 36 months. Even though that\u2019s aggressive, he says, \u201cwe are a factory that runs 24\/7. We can\u2019t allow a single minute of downtime. With newer equipment, the probability of failure is less.\u201d\n\nRodr\u00edguez has eight key criteria he uses when reviewing existing IT hardware, software and services:\n\n\u201cTo me, the most important criteria is business alignment,\u201d he adds. \u201cThat means alignment with our strategy, market, costs, and technology.\u201d\n\nBSH Home Appliance\n\nVice president of digital platform services Berke Menekli takes an analytical approach to regular legacy technology assessments at BSH Home Appliance. \u201cWe record the business capabilities into our enterprise architecture platform for every tool so we can run a portfolio optimization process over those\u201d to detect any overlap in capabilities, he says. \u201cAny organization of our size has 1,000 to 2,000 applications registered,\u201d he adds, so having that capability analysis is a key factor when deciding whether to retire or replace a technology.\n\nHe also considers alignment with current business strategy and where the product or service fits within the technology refreshment cycle. Factors include whether the product is at end of life, whether the asset is fully depreciated, and whether they\u2019ve fully leveraged the value of the investment. Everything is on a schedule: laptops and data center equipment are replaced every four years, smartphones every three to four years, and data collection devices in the factory every seven years.\n\nDuring the software review process, which takes place every two or three years, he considers the license contract, replacement cost, and ongoing, recurring costs. \u201cWe also look at IDC and Gartner research to understand the capabilities and costs of alternatives,\u201d he says. There\u2019s no standard timeline for replacement except for end of life announcements, he adds.\n\nMenekli\u2019s strategy for maximizing the value of IT investments has changed in recent years. \u201cPreviously we focused on optimizing yearly costs. Now we\u2019re focused on the end game: reducing our overall cost position.\u201d In some cases that\u2019s meant accelerating implementations so the company can retire older investments in order to reduce costs. But it\u2019s all about the finances, not bringing in the latest and greatest technology. \u201cIf it\u2019s fine to go a couple more years with the technology, we do.\u201d\n\nThat may be the case with Macbooks and Chromebooks, which Menekli says tend to have a longer lifespan than his Windows laptops. \u201cWe\u2019re looking to see if we have a business case if we can keep them for seven years,\u201d he says. \u201cWe\u2019re always investigating whether there\u2019s a better business case.\u201d\n\nDeciding factors\n\nThe most important consideration when deciding to replace information technology or not is whether it\u2019s still in alignment with the business strategy, says Rodr\u00edguez. Then look at what\u2019s happening in the market, with costs, and with the state of the technology.\n\nEvery organization should review legacy technologies during annual planning and budget cycles, says Ivy-Rosser. That\u2019s standard practice. \u201cBut every time you have to justify a CapEx or adjust an OpEx budget, you should do the review again\u2026and make technical debt discussions a transparent part of the decision-making process,\u201d she says.\n\nBut cleaning up technical debt shouldn\u2019t be your only consideration, says Menekli. You also need to consider the financials and whether a technical refresh is warranted.\n\nBe careful before placing your bets on a replacement technology, and realize that the grass is not always greener, adds UC Riverside\u2019s Gunkel. \u201cAll platforms have their problems, so be very intentional about what those problems are, and which ones your team can intentionally solve and can control through your staffing and team.\u201d If the technology you have helps solve your organization\u2019s weaknesses, holding on may be your best bet.\n\nWhen it\u2019s time to make a change, if you\u2019ve done your homework up front and have a separation agreement, you\u2019ll have an easier time of it, says Abernathy. Also, have a plan if the new technology falls short of expectations. \u201cDecide early on how long you want to invest if the value is slow to come,\u201d she says.