Talk Is Cheap, Action Is Costly
February 1, 2021 Timothy Prickett Morgan
Companies pick the platforms for their mission critical systems so they can run them for decades. And it is a very, very big deal when they decide to change those platforms. And, furthermore, it is a hell of a lot harder to change these platforms than many people think, and that is because most of the value of those platforms is locked up in the applications and databases that run atop the servers, operating systems, and middleware.
It is with this in mind that we always take any survey data that asks IT executives if and when they are going to move off one platform and onto another one with several grains of salt. It is not that we think people are telling untruths or fooling themselves – we think they are giving an honest opinion. But it is rather that such moves are filled with risk and in many cases the moves are more risky than keeping the legacy platforms in use running those legacy applications. That is why modernization is such a force on all applications, including not only IBM i but also IBM z/OS and Microsoft Windows Server platforms, and maybe even a few Linux platforms here and there.
As I usually do in January, I recently participated in the webinar going through the results in the seventh annual IBM i Marketplace Survey performed by HelpSystems. If you missed it, you can check it out and hear what Tom Huntington, the executive vice president of technical solutions at HelpSystems; Alison Butterill, IBM i product offering manager; Ian Jarman, business unit executive for Power Systems Lab Services; and Brandon Pederson, worldwide IBM i product marketing manager; and myself all had to say about the survey results. One of the slides referred to the following question: What are your plans for the IBM i platform?
Here is what the respondents to the survey had to say:
This chart has wiggled around here and there over the seven years we have seen this survey, but the shape does not really change all that much. That 42 percent of those surveyed said they had no plans to change anything and that 25 percent said that they planned to increase their IBM i footprint is not a surprise, 12 percent said they were not sure of their plans and 12 percent said they would migrate some of their applications off the IBM i platform. Of the 9 percent who said that they were going to move all of their applications to a new platform, 3.5 percent of those (or 39 percent of that 9 percent) said they would move within the next two years, and if you assume that half move in one year and half move in the second year, that implies an attrition rate of 1.75 percent. The rest of those who said they were leaving the platform are going to take longer. And for some, it will take a decade or more. Or perhaps forever.
An attrition of 9 percent a year would be dramatic, and would be much higher than the average attrition rate of around 3 percent per year that we saw between 2000 and 2018. In the early years, we were losing maybe 15,000 customers per year on a peak base of 275,000 unique customers, and that rate slowed as the diehards stuck with the platform, and maybe now we are losing 2,100 to 2,600 customers per year (against a much smaller base of somewhere between 120,000 and 150,000). And the base probably adds a few hundred customers per year as well, which helps even if it doesn’t balance it out. At an even amount of 9 percent of the current base per year or 10,8000 customers, it would only take 11 years to vaporize the IBM i base. If it is 9 percent of the base incremented down per year as it declines, it will take 40 years to reduce it from 120,000 to 3,000 customers. Neither of these scenarios seems likely given the current commitments that Big Blue is making in the Power server line and the IBM i platform. We think that without too much investment, IBM could keep the IBM i platform going at least to 2035. (That is just using Power11 chips.) We think the actual attrition rate is probably closer to 1,000 customers a year against maybe 300 new customers, but that is just a guess.
Here is the fun bit as far as I am concerned, and this thought just occurred to me during the U.S. webinar that I did with HelpSystems and IBM, and it is that if you asked the same questions of any other legacy platform that is still viable – meaning Windows Server and System z pretty much since Solaris/Sparc and HP-UX/Itanium and OpenVMS/Itanium are all dead – the customers would probably say the same thing. And would probably not move off those Windows Server or System z platforms at the expected rate, either. They definitely are abandoning the other platforms. And Linux, as we remind everyone during the survey webinar, is the only operating system platform that has been growing for the past several years.
The other neat part of this chart is where they say they are going, and this year we added SaaS services as a response alongside of the other platforms because this is the real way people are thinking. And when they go to a SaaS platform, there is a better than even chance it is going to be Linux instances running on a cloud or Windows Server instances running on a cloud, but among the IBM i base, there are a fair number of application vendors who are building a successful business by moving their IBM i applications to their own or someone else’s IBM i-based clouds. I do not really believe that any IBM i shop will move to z/OS and mainframes, and I believe very few are really looking at AIX as strategic in the way that Linux clearly is over the long haul.
Intent is not the same thing as action, and if it were, I would have found the money and time to build the deck and the hot tub out back. Instead, I work in the garden and build my wood pile and tend my chickens.