The Silver Lining in Your Cloud Migration: A Chance to Clean House
Any vendor who helps your company move data to the cloud will promise Quick and Easy. Taking the headaches out of a potentially massive migration is one of the main selling points for every technology provider involved in the process. Who would suggest adding steps, costs, time, or complexity to the process of putting your data in the cloud? Well, I would.
The easiest way to get your data into the cloud is to “lift and shift.” Take what you’ve got on premises and put it in the cloud, exactly as it exists today on your servers—not only the data, but your mappings, your processes, all the code that helps you define, process, and move that data. But that means that you’ll also export all the inefficient code, sub-optimal short cuts, and outdated practices that have built up over the years and that have been quietly burning up more of your operating budget. Why enshrine old mistakes on a new platform?
It’s another case—as I discussed in this post on cloud transparency—of the cloud providing an opportunity to identify and rectify inefficiencies. In the case of data migration, you’re going to have to test everything once you get it to the cloud, anyway. That’s a significant portion of the time and cost it would take to reengineer for efficiency, and you’re already committed to it. So this is the best opportunity, with the smallest addition of time and resources, to get the job done right.
Every IT organization struggles with the (im)balance of innovation and maintenance. A classic estimate is that at least 70 percent of any IT budget goes to supporting the existing infrastructure rather than doing anything new. And that’s not a static number. Over time, inefficiencies multiply and the task of supporting your sprawl consumes more resources.
It starts small. Few IT orgs are ruthless about enforcing coding standards, so over the years, a dozen different developers using their own preferred shortcuts code a dozen different error-handling routines that disappear among 10,000 other units of code you’re supporting. When one of those routines fail, your support analyst has no single best practice as a point of reference for understanding and correcting the problem. Every fix takes longer, requiring more staff time and more budget dollars.
And I’m talking about real money. When I was running a huge enterprise data warehouse at Pfizer, I reduced support costs by 20 percent every single year by building things more efficiently—and every saved cent went to innovation. That’s vital to enterprise IT today, because business leaders are more insistent than ever that IT innovate more, yet IT budgets rarely keep up with a company’s ambitions. There’s no way to succeed without reducing maintenance costs.
To the degree that using your cloud migration to do a full checkup on your code will increase time and costs, build a business case, because I learned time and again that there’s a significant ROI argument: “Yes, we’ll invest 20 percent more on the front end to reengineer our ETL processes during this migration, but we’ll achieve an annual savings in support costs of percent, which gives us ROI in under 18 months—and that ongoing savings will go directly into (key digital transformation initiative).”
And there are more opportunities than just what’s in your data center. Most large enterprises have a shadow IT problem, with business units using their own budgets to maintain ad hoc IT services. More than once, I went to the business unit and said, “Give us your budget on this, and my team will take this over for you. We’ll migrate the data, maintain the service, and split the savings between our teams.” The BU has one less headache, and one more pot of money to spend on better things. IT has more control of its infrastructure, better data quality… and a little more money to spend on innovation.
Always talk to your internal customers, and never overlook an opportunity. Beyond roping in rogue IT projects, find out how users actually use the software you provide them. How many licenses are going to waste? Which functionality are you paying for that no one uses? What can you add that would significantly boost success? The answers to those questions will be a mix of efficiency gains and innovation—and often one will pay for the other.
The value of cleaning house
IT people, especially at the infrastructure level, have an appreciation of efficiency and elegant design for its own sake. Well-designed processes and clean code are pleasures in and of themselves. But they also have business value, especially in an era when IT agility is paramount.
Using a planned data migration into the cloud to go that extra few steps to really clean up and reengineer the code around your data will deliver value, and the excuse of the migration helps you win approval for a project that would probably never get a green light on its own.
If you’ve got war stories about finding the silver linings in your cloud migrations—or other ways to drive efficient IT—let me know in the comments. And come back in a couple of weeks for my next post, which will look at best approaches for getting value from a cloud data lake.