Friday, 11 April 2008

Small Businesses and IT

One thing that I am 100% convinced of, is that over the course of the second half of 2008 a new class of application will begin to emerge in anger. Rather than being based on an underlying Single Machine Operating System, it will be composed of “cloud” services.

So where EC2 and S3 provide compute units and storage in the cloud, the next stage is to compose higher level functionality into cogent solutions, and no longer worry, from a development platform, what Single Machine Operating System lives underneath. The platforms to support these are currently divergent – Force.com , Google AppEngine, the platform that MSFT are clearly about to launch – these allow you to create generic functionality on fully hosted platforms, but from a 50,000 foot view they provide automatically scaled out, redundant platforms, generally with storage, data, and UI components.

Add a bit of Google Gears, or Microsoft FeedSync, two different techniques for occasionally connected apps (and, the piece often forgotten, occasionally unavailable Platforms-In-The-Cloud!)

Mix in a bit of Hosted Exchange, which you can buy from all over the place, or Google Calendar.

Throw in some oAuth and OpenID for decentralised delegation and authentication.

Add some hosted IVR or SMS from solutions like that from my employer, and you too can have CTI like the big players.

And it looks like the applications that SMBs will buy in 2008/2009 will look very different to the ones they bought in 2007.

This leads me to the conclusion that both sales and reward, and software development processes for small ISVs will start to look very different from today's models. Instead of on-premise based software, or even ISV hosted solutions, generic, composable mashups will be assembled from cloud based primitives through an oDesk type model (no more staffing issues!), and then “mashups of mashups” will probably get sold through a network of affiliates who do highly localised (both geographically and vertically) marketing and the small amount of custom integration work – which will be largely graphical – required for an end customer. Reward models could be based on subscription payments from end customers, and revenue share to the affiliate channels

The problem is working out which generic functionality, and which SMB verticals to address first!

Are you an SMB? Would you like me to come and build you some software? What's the itch you have that needs scratching? Which of your staff could be replaced by a clever machine? How can you empower your customers to interact with your business in ways that benefit both you and them?

Trust me, small business software is easier, quicker, and less risky then it used to be. If you want to know more, I can be reached, in the first instance through this blog.

2 comments:

Josh Breinlinger said...

Hey Tim,

Glad you think oDesk is a good enabler to this type of application development for SMBs. We're constantly striving to make it easier for SMBs to find top-quality resources in a rapid on-demand way.

fyi - we just recently launched some cool pages to show programming trends. http://www.odesk.com/trends.

-Josh
oDesk

DE said...

The bane of the cloud is bottleneck dependencies. Now when EC2 has a hiccough, large numbers of apps go down.

But the obvious benefit of cloud apps and cloud app development cannot be doubted. Being able to write a Rails app (on Heroku) to target FireEagle without a file being made on my hard disk has been a real eye opener.

Tim Stevens

Tim Stevens
Work
Consume
Obey
Be Silent
Die