All text below licensed: as CC-BY (http://creativecommons.org/licenses/by/2.0/)

Colleagues in the Caribbean: a very good morning to you from the World Bank in Washington DC.

Thank you for inviting me to share my thoughts with you, it is my pleasure and privilege to join you at the start of your "Developing The Carribean" conference - Thank you for taking the effort to set up this remote connection. I'm sorry I can't be there in person for what promises to be a thoroughly exciting few days.

It might be different in the Caribbean but before the advent of Open Data in the United States and Great Britain - it was rare to see government officials, civil society organisations and technologists all gathered under the same roof with a common purpose: if you do nothing else, use your time to meet people and learn what you can do together. Enjoy it!

So, I'd like to share 4 things I've learned working in the world of open data over the last few years. Quickly put, the 4 ideas are these:

first: open data might mean different things to different people, make sure you know what it means to you and others.

second: open data is not an end in itself, it's a means to an end.

third: there's an ecosystem around open data and it's made up of materials, tools and people - it's OK to focus on one but don't forget the rest.

fourth: make sure you're meeting a demand or solving a real problem - make something that people would miss if it went away and most of all, talk about and share it.

I hope these 4 ideas will stay with you over the next few days and be helpful as you continue your work in the future.

So, let's see if we can get the first point right - are we all talking about the same thing? What is open data? What does open mean? What is data?

For me, if something is open, it means that it's both legally open and technically open. If something is legally open, you can freely access and use it, re-use it for a new purpose and redistribute either part or all of it. If it's technically open, then it's available  to you in an easily accessible, standard and editable electronic format.  So that's open - what about data?

For me, when talking about "government data" I mean all data and information produced or commissioned by a government. From statistical data and administrative information to political and policy information, and data from funded research. Government data might be anything from raw numbers and text to complete documents and maps along with various levels of analysis.

The reason I talk about this is simple: open is a powerful idea and a powerful brand. If you say you're using or producing open data, you've joined an amazing global community of individuals, organisations and governments doing the same. If what you're doing doesn't meet this community's definition of open, they're going tell you about it. And you, as part of the community have a responsibility to do the same: don't let someone call something open if it's not.

So here's my first key idea: open data is going to mean different things to different people. When you're talking about, advocating for or using it, make sure you know exactly what open data means to you. For me, it's really any information that can be considered technically and legally open and I don't let anybody call something open if it's not.

At the World Bank, we often say that we're open about what we know, and open about what we do. For example, our entire flagship statistical dataset, the World Development Indicators is free and open for commercial and non-commerical use and available online in open electronic formats at data.worldbank.org. With over 9 million users since launch, our data now reaches more people than ever before; users regularly take our data and incorporate into tools and services that make them money while adding value for others.

We're proud of the increased reach and potential impact our data can now have and I  believe that the best way to get value from data is to give it away.

So here's my second point: open data is not an end in itself, it's just a means to an end. I like to think of it as a necessary but not sufficient condition for what we really want to achieve. All the promises we hear about open data to do with increased transparency and accountability, innovation and efficiency are great, but these things won't happen automatically. They require effort. To be useful, data needs to be used and used by people with a purpose.

For the World Bank, the vision behind opening up our data, and indeed our entire organisation, was to "democratize development." It's about giving everyone free access to the tools and knowledge they need to help improve their own lives and those of others. Do I think the World Bank Open Data initiative has fulfilled our vision of democratizing development? Not yet. But it's getting there. It's a key part of a fundamental shift in the way we're thinking and working.

So what are the ends you're working towards in the Caribbean? What will open data help you to do and how can you help make this happen?

And here's my third point: people are part of a bigger open data ecosystem we have to remember. The Open Knowledge Foundation describes this as: materials, tools and people.

What are the materials? These are the data, documents, maps and information that governments and institutions supply and release.

What are the tools? These are the laws and policies that makes it possible to release data it in the first place, the software and computing infrastructure that makes it possible to publish, work with data, and the channels like radio, newspapers, television and the mobile phone.

And then there's the people: you. Skilled people who have and understand data, people who understand the challenges that data can help to address, and people who are motivated to turn ideas into action.

When these parts of the ecosystem are all working, some really interesting things can happen - here are two examples:

In Bazil the federal government ran an anti-corruption programme that audited and published the data on irregularities associated with fraud in procurements and public finances over the Internet and local media. This data, exposed for 373 municipalities had a dramatic effect on local politics and subsequent elections. In the areas where damning information was shared by radio, there was a 30% drop in chance of the same people being re-elected.  Open data, combined with analysis and dissemination by radio changed government.

On a different continent - Africa - the Kenyan government last year launched an open data initiative. As part of this, the ministry of education published the geographic locations and basic information about schools around the country. What they knew was that these locations and the information with them wasn't all accurate or complete. A local company, a telephone call centre, saw this data and got in touch with the ministry. They said - look, we can help you. We've got phone numbers for these schools. What we can do, is call them, get better information about their locations and other key data you want to collect, and then share this data back with you. The ministry was delighted - the company was able to use the skills and tools they have, at a lower cost than government, to help create better data for the public good.

There are hundreds of more examples like this one and I hope you'll soon start to share your own - what's common to them is this ecosystem: materials, tools and people, you need to work on all three to succeed.

And that's really my last point: as you carry on your work, be fully engaged and do something that really matters to you and to others. Find a problem and fix it. Understand demand.

If you're a government or an organisation wanting to release data, make it count - listen to demand and be driven by it, collaborate with and support the people who're using your data, their work with help you both.

If you're going to be using data in any way: make it count, find a problem and fix it. Develop your skills and understanding, make friends and connections in the community so you make things that matter.

My Caribbean colleagues, welcome to the global open data community: share, learn and show everyone what you can do.