Sunday, May 22, 2011

Managed Service or Enterprise Software

Cloud computing is really big. Not a day goes by without some article or tweet talking about the advantages of running your applications somewhere else. Truth is that I cannot think what life was like prior to Dropbox. So I suppose it makes sense to use services that are shared and where one can tap into the economies of scale.

It is interesting that many small companies that I know still run their own ERP systems. Deployed behind their firewalls and specifically customised for your unique requirements, these systems are one instance of the software running on dedicated equipment. Surely it would make much more sense to move all this repetitive work onto a shared service?

We have a number of customers that started mobile money on shared solutions - software deployed on shared infrastructure (a hub) and then moved off it to install their own dedicated deployment. Why would they do this when it is clearly a more expensive alternative. The answer is purely a function of uniqueness. The more unique and different your requirements are, the more likely it is that you would rather deploy your own dedicated enterprise solution... managed by yourself and configured to meet your unique needs.

To what degree is it possible to offer a managed service for mobile money that is generic enough to solve most of the problems? That is the question.


BigChief said...

"To what degree is it possible to offer a managed service for mobile money that is generic enough to solve most of the problems? That is the question"

I think it depends on the level of maturity of the market really. It is harder to offer a generic service that solves all problems and provides all solutions in the emerging markets. The bar has to keep being raised gradually as the threshold requirements change with competition and availability of technology and infrastructure.

The emerging market is a different animal and not at all homogeneous. I keep telling everyone that a "one size fits all" approach will never work and one country could be as different from the other as night is from day.

As a change management practitioner, I see things from the perspective of the end-user. The End-user ultimately decides the way forward.

There is an interesting blog post titled "A List of Completely Wrong Assumptions About Technology Use in Emerging Economies" I reccomend to anyone

It is all really about the users and what they "need" not about the bells and whistles of the technology. The limitations of bandwidth and skill limitation are also major challenges that cant be wished away by trying to reduce cost.

There are also different definitions of "Managed Service" It does not always have to be a service provided remotely in the cloud. You can have a "Local Managed Service Model" where the solution deployed locally but all deployment, support and changes are outsourced to a 3rd party. I believe this is what really works for most of Africa

For this to work properly, the infrastructure and application has to be scalable and flexible enough for changes to be made by the local managed services provider.

There has to be a paradigm shift for this to happen effectively. Every deployment does not have to be like MPESA.

I have seen solutions provided by different technology companies based on wrong assumptions and some unneccessarily complicated. This model should change and should be a generic flexible base which will act as a platform for local developers to build on.

Technology companies should provide a platform for local partners to customize and adapt. To think that one technology company can solve all mobile money requirements in all countries and locales will be foolish. The model has to change

Ultimately the success of a deployment depends on up-time, availability and meeting the end user's requirements. The failure of the "hub model" to address all this can is mainly because of differing end-user requirements. Mobile Money is personal and it is about the end-user. The operator that provides more to the subscriber wins the game. The subscriber is not bothered if it runs as a shared service or locally.

A stable generic base platform but localised configuration and support is the best solution.

Infrastructure issues will also not disappear overnight. The bar has to be raised gradually but it is not yet time for the emerging markets.

Operators expectations have to be properly managed and change management should be a priority. Mobile Money is not airtime or VAS sales.

BigChief said...

To add to my earlier post I borrow a line from your intro to this blog

"This must be one of the biggest social changes that humans were ever submitted to"

The emphasis should be on the word "Change" and it has to be managed properly