At times, some companies can feel as though they are caught between the present day of preserving the legacy systems they have now – and the future when a transition into AI and machine learning will be fully realized. This state of limbo provides further pressure when employees who only understood the legacy systems retire, yet there are no questions that there is an acceleration toward machine learning.
All in on the future? Or is that being too aggressive?
First, we can say that most of our clients in the last 20 years at Roy Talman & Associates have not been using 40-year-old technology. One of the sectors we serve, high-frequency trading, only started growing significantly at the beginning of the 2000s. Every system built around that sector has been “footing the bill” since 2001 and after that. Many companies still have them.
Still, we can’t help but hear leaders and managers saying, “OK, we’ve got a 15-year-old system. It still works, but we’re always worried it will stop working. So we’re defining a new system that will do even more than the existing one, even though it could be a costly proposition.” It’s a tricky decision. Some businesses hope that as they build a new system, it will essentially overtake the old one, so they will never need to go from Version Two to Version Three. In reality, you must maintain both Version Two and Version Three. It’s not like flipping a switch from one to the other.
The Act Of System Replacement Never Stops
Over the years, we’ve witnessed many financial firms have one system that’s taken seven or eight years and hundreds of millions of dollars to build. Before long, the need to replace that system with a new one surfaces and could cost even more than the first system.
There is one crucial point to be made here: As the new system is being built, you cannot retire the whole of the prior system at once. For quite a while, parallel systems may be run. Slowly but surely, the new system takes over from the old system. By the time that happens, guess what?
Seven or eight years have gone by and it’s time to build the next system, which could now be not just hundreds of millions of dollars but more in the neighborhood of a billion dollars. The cycle repeats every seven or eight years until the next system is fully functional.
What a far cry from the apps we regularly update on our phones, right? Phone apps are being replaced continually. This is what other companies are coming to realize about the pattern of system replacement – the frequently updating apps model is one they need to adopt, which regularly calls for a different architecture for the system.
An Era Made For Microservices Architecture
Microservices architecture is comprised of multiple technologies that are very interchangeable. Suddenly, when a system has so many parts that can work together and be replaced easily, modern software can be developed in a very accelerated and exciting way. At any given time, a large company like Facebook could run thousands of different versions of its software.
At the same time, even if there are 100,000 microservices, the functionality doesn’t change and the user doesn’t need to re-learn anything.
It’s a massive difference from how developers wrote software 20 years ago. If you changed a system, the problem was that you might create changes in thousands of different areas all over the system.
These days, our systems more frequently work with constantly evolving iterations of the “best of the best” technology – those not embracing this ever-modifying and flexible systems architecture risk falling farther and farther behind.
Take PayPal, for example. Twenty years ago, PayPal was a tiny pipsqueak of a company compared to a giant banking leader such as Citibank. Return to the present day, where the market cap of PayPal is approaching 50% of Citibank. Finance should have a sophisticated technology setup, but many hang on to their 30-year-old mainframes.
So far, so good for those ancient mainframes today, perhaps. But eventually, those who use these technologies will see they’ve been failing for quite some time and not realize it until it’s too late.
Coinbase is another fine example. Coinbase has had a global market cap of $67 billion. By comparison, the same international market cap for Citibank is $147 billion. Yet, when many think about the debut of Coinbase, there were serious questions about how the company could create so much value out of thin air. That’s not a problem anymore.
We can say that the longer a company keeps its main legacy system while simultaneously accelerating the development of a new system and testing it, the best of both worlds from yesterday to tomorrow will be experienced.
The top technology demands the “top 1% of the top 1%” talent to work with it. That’s why so many in the tech space turn to Roy Talman & Associates. Our clients know their opportunities are rare and worth waiting for the unquestionable fit. So they consistently rely on our team’s superb experience and resources to help identify their next great hire through a series of tests that prove they’ve got what it takes to make an impact today. And lead the way to exciting transformations in the years to come. Settle for nothing. Talk To Talman First and gain everything.