Gartner made a name starting in the mid-90s forecasting the estimated cumulative cost of Y2K remediation. I was there – and the big numbers it bandied about helped focus enterprises on the core problem. But it also led to hype, panic buying (and exaggerated market declines on the other side of the peak) and many, many poor IT investments. Since then Gartner has picked on many events – such as the introduction of the Euro between 1999 and 2002 - and piled up potential costs to come up with a single, usually scary, related aggregated IT project cost forecast.
The latest is what Gartner calls the IT Debt crisis:“the cost of clearing the backlog of maintenance that would be required to bring the corporate applications portfolio to a fully supported current release state.” It estimates it to be “$500 billion in 2010, with the potential to rise to $1 trillion by 2015.” At its Symposium next month, this should draw folks to the session "Building the Compelling Case for Application Overhaul".
The problem is there are many good reasons to NOT upgrade/modernize many applications, and I believe Gartner is out of line using words like “debt” which have guilt associated with them.
Here are some realities that need to be factored in the modernization analysis:
- Many applications deserve to be consolidated, not modernized: While lots of good work has gone into data center consolidation and virtualization of servers, applications have lagged in such scrutiny. There is duplicative or minimal “effective” functionality in more applications than you think. A portfolio analysis which rigorously inventories applications by complexity, criticality, stability, usage volumes etc will show 10 to 50% of applications could be retired without skipping a beat.
- The packaged software maintenance model needs to be radically revised: Those who have moved to third party or self-support of many packaged applications – and in the process chose to stay on older releases - deserve kudos not criticism. There is little incentive for them to upgrade to the latest release especially with software vendors who arrogantly expect you to pay for arrears - years of “missed maintenance”. Look at the track record of the software industry: Vista. Long delayed Oracle Fusion applications. Stream of Google products which never leave beta status. SAP ECC 6.0. See many companies voluntarily lining up? Usually it takes threats from the vendors (there is a term for it – “forced march” evoking brutal, hostile wartime conditions) to get the customer bases to move, when it should be business ROI. Shockingly, those who stayed with the original software publisher on older versions are expected to pay a premium for long atrophied versions. This for minimal support as vendors have moved to automated and self-help models, break fix from much cheaper offshore locations and developer community models. The packaged software maintenance model is broken. Don’t reward software vendors for it by staying on their upgrade treadmill at their arbitrary 18 or 22% a year of software “license value”, much of it in unused shelfware.
- Application/Business Process outsourcing deserves an overhaul: Starting with Y2K, and expanding over the last decade, application maintenance in many organizations is now done by the Accentures and TCSes. They have reason to perpetuate, not retire or modernize, applications. When they at the outset take over an application inventory, most go through the portfolio exercise described in a. During the course of their contract few of them offer to refresh their portfolio analysis. Example, many organizations went to a BPO model but continued to support their own Oracle or SAP applications. Why not let the BPO provider move to their own transaction engine and retire your own? Also expect more –most outsourcing firms have mostly provided “single tenant” support when SaaS models are showing huge productivity with multi-tenant, shared service models. Even if you have to keep them in single tenant mode, challenge their metrics. Effort per ticket or enhancement should have been steadily declining. No point bragging about CMM Level 5 or Six Sigma unless it shows in continuous productivity. Finally, make them reevaluate their delivery models – wage inflation, currency impact have made India, in particular far less compelling then it was a few years ago.
- Modernization projects need to become much more agile: We have moved into a world where SaaS vendors are doing upgrades in background and bringing systems down for just 5 minutes. In contrast many enterprise upgrade projects require long weekends to execute and months of planning/preparation/ and yes prayer. Don’t attempt to modernize applications until you challenge your IT staff (and your system integrators) to dramatically rethink your change management/upgrade processes and making them massively more agile.
- “Oldies can be goodies”: Finally, let me invoke a former Gartner analyst, Erik Keller. In my book on innovation which, of course, encourages all kinds of modernization, he provides pause as he now helps special – education children and cancer victims with gardening/horticulture projects:
“We make our kids understand that they are stewards of the environment and the success or failure of the garden is in large part due to how they take care of their plants and the soil. At the end of the season, you measure success by how much you have harvested and how well you have left the garden for next year’s farmers. In the technology business, we often don’t think about tomorrow’s impact of today’s actions. The term “legacy ” is considered a negative in technology whereas in farming it is a positive.”
By all means frequently review your application portfolio but plan your actions out of business payback and expect many efficiencies pointed out above.
But don’t do it out of guilt - you are not “indebted” to anyone.
Dennis Moore adds his perspective - may be the vendor owes you...