Lessons on what makes successful new startup services

I spoke at the Converged Services Congress in Amsterdam yesterday, which I shall write up later - but initially I wanted to blog one session which I thought was very useful for the Web 2.0 startip scene. It was by Daniel Holle, who is researching for his PhD at the University of Regensburg, and waqs based on what Telcos and Media c o's have found mekes for succesful new launches in the converged service space.

(I've remarked before that I have a foot in the Web 2.0. the Telco, and Media Co camps and though they are all doing much the same stuff, they are often calling it different names, and wearing different clothes )

Anyway, the key points of Daniel's research so far are that success is driven by:

Understand user Needs – stick to the user, not the technology

- Seek repeated feedback and integrate into service
- Understand the usage scenarios – watch what the users are actually doing
- Don’t get blindsided by technical possibilities

Make the choice on features – and integrate the systems to deliver these, rather than have a number of uni-services existing side by side

- Focus on what actually adds value from what already exists
- “Known” functions must be as good or better than “good enoughs” already available
- Give users a selection choice

Get Usability right, early – and reduce the complexity

- Use similar usability to existing “knowns” initially
- Seamless transfer of “knowns” is required to migrate customers to the new service (eg easily set up friends on a new social network, migrate email addresses to a new client etc)

Don’t forget about the overall ecosystem – need to test in an end to end environment.

- Daniel differentiated between two ecosystems – the early stage and late stage ones:

Early Stage Ecosystem – keep it closed and simple and ensure:

- high quality of operation
- high attention to User Experience – iterate like mad to get it right
- Small number of service options / tariff options initially

Later Stage Ecosystem

- Open up Ecosystem to 3rd parties when it is stabilised
- Avoid functionality that you can’t manage reliably
- Increase personalisation options

I asked Daniel if there were any lessons on scaling – ie whether to design for scale early, and risk over-egging the pudding, or grab customers first and worry about scaling later if you are successful. His view was that its far easier to scale a system when its simple, and his view was that thos companies that scaled well had got the product right first, kept it simple, then re-designed where necessary for scale .