Build, Buy Or Rent!!

The very distinguished industry watcher Erik Keller writes that build is back inside enterprises. He sees that when the enterprise software industry speaks of growth, it is growth only of cash-oriented expenditures (SaaS, maintenance payments, training, servicing bug fixes) and non-license fee capital expenditures (services to modify existing packages, custom application extensions, and so on). As software license growth stagnated during the last five years, custom applications rather than pre-packaged ones have represented the lion's share of enterprise software growth.
As I see it, we are now in a world where the economic dynamics have changed substantially - offshoring and contracting have made maintaining and extending applications a much more viable option. The packaged software vendors are in many ways charging their customers far too high – particularly the maintenance cost – it still baffles me that the walking dead, small, big or pathbreaking software company – all alike charge maintenance in a similar band. The so called rapid implementation of packages look more like a brochureware(though I should admit based on empirical evidence that the rollout times have come significantly in the last five years), but the number of funcationalities that need to be implemented probably offsets any great advantage. Vertical solutions prophesises by package vendors more often than not fall short of expectations. Packaged software is getting more and more complex to manage. Enterprises may not have a choice but to buy for core transactional application but all extensions/add-ons would find build an attractive option . Needless to say strategic applications would get built with lot of custom code –in some cases the entire lot, in some cases atleast the icing in the cake. I see that banking & financial services, one of the most prolific spender on IT services atleast – sees a lot of custom development – the trend is increasing.

Erik keller parades impressive statistsics in support of his argument – I think some more discussion may be needed over the impact opensource or SOA has on the application landscape that we see today but offshoring is definitely helping to bring down costs. Some classification data may also need some discussion – say the case of iLog implementation that Erik highlights – where to classify deployment efforts – build or buy – given that the kernel – in this case the rules engine is a prebuilt product – if we are to take a view that extending/integrating this need to be classified as build solution then by definition a majority of packaged applications may get shifter there – in these instances my argument is look at what adds strategic value and classify accordingly – here Ilog’s scheduling algorithm brings more value than the java code used in building the applications. Nonetheless an impressive article.
I think that buy,build & rent all shall co-exist and enterprises would have to make appropriate choices based on their need. The package industry's future would be determined by its own conduct and not by any new build wave. But clearly the build movement shall continue to retain/slightly expand in its share of business volume. I shall come out with a follow up post on this in the near future. As an aside, Boeing connexion is not exactly the service to rely upon for involved blogging!!