There have been many predictions for 2012 and one common theme is that Business Intelligence and Analytics will continue to be a big area of growth for CIO's. This resonates strongly with the ethos at 5point9, where we want to help customers maximise the value from their information. Big Data is also being heralded again as a discipline or area inside this. In this post we take a look at Big Data and the likely developments / issues.
In this post we take a look at Big Data and focus on the building blocks or foundations of the platform and considerations, this is a topic we are going to be looking at in more detail over the common weeks.
Doing it because you could rather than because you should:
Big issue with Big Data. This is something that is common place and this is usually indicative of the disconnect between the business and the technical team. There are many moving parts inside Business Intelligence / Analytics programs and this is intensified in the big data concept. The result of this is that you have a solution which is fantastic but is disconnected from the business.
There are many techniques to counter this but one way which we have used with success with many customers is to have a Steering Group which is based on teams from the various parts of the organisation who come together as equals with senior buy in. This is not a guarantee of success and this must be time boxed and have a strong chair. The teams involved should be made up of the technical elements (Data Owners, Master Data Management, Infrastructure etc) and from the business there should be representatives of the line of business and users. The key benefit of this is that everyone is aligned from the beginning and there is broad agreement of the vison for the project - what it will achieve, why it is happening, the expected ROI and justifications.
Visualisation techniques:
Despite this being a hot topic for several years this is still an area of neglect. A lot of effort is put into the data and this is absolutely essentially. However without a platform for business users to simply and effectively interact with data the solution will become a white elephant.
There needs to be an interface which allows users of all levels of capability to use and gain insights. The consensus that senior managers need a high level and information workers need to be able to access the details is too restrictive. Dashboards are not the answer for the execs and low level detail is not always the answer for the information workers. Everyone in the organisation is a consumer of the information and therefore should be able to traverse through several layers of detail and complexity simply and still understanding the context. There needs to be a story to set the scene and make it simple for users to understand what is being looked at. The old mentality of assuming that senior people are not tech savvy or only want to see a high level is flawed.
At 5point9 we want to enable all users (with the right security) to be able to understand the 35,000 ft view but also get in amongst the reeds of their business. The social network angle is still woefully in adequate in organisations but this is beginning to grow with the use of collaboration tools and the integration of these with the data and information. Large organisations are now embracing instant messaging but are stil lagging behind on the social developments - we are developing a post on this soon, watch this space.
Actionable information:
We have a rule at 5point9 with customers for Big Data and BI - if we cannot succinctly describe the actions which will result from something we do not include it. This can be a heated activity but one which customers delight in. Once this ethos is distilled we can see real change within the customer and a much improved focus. We want to be a sniper rather than shooting from the hip.
One outcome of this approach is that there is absolute business clarity from the data right the way through to the knowledge used by the business. It can often highlight areas of missing information or additional details that are required for data to become actionable.
Security and Audit:
Do you have a handle on who can see what data and how is this being managed? Is the decision maker / team deciding this correctly placed? Has the business being consultated and what is the process for users to change what they can see? This is a big area and one which can easily frustrate users.
There is a balance between restricting users to the correct level and also allowing them enough access to see the information they need to do their job effectively. This is also massively influenced by regulations and industry specific standards. As well as security access the audit-ability of what users have seen needs to be considered and as big data is more prevalent this becomes key to be able to report on this.
Data Governance:
Master Data and Data evolves over time as does security and users requirements (visualisation etc). We need to ensure that vehicles are put in place to handle this. Big Data programs which do not iterate or allow for change will quickly become dormant and redundant.
Performance:
If the performance is slow or the tool is cumbersome or the users have issues, they will create silos or extractions from the repository. You users are resourceful and smart - which is why you have hired them. Is this happens you quickly revert to having pots of information without governance and control or consistency. You lose the ability to ensure that data is being used across the organisation consistently. The simple way to do this is to ensure the performance of the system and ensure that users have the correct level of access in time and accurate.
Data availability must also be processed within acceptable time limits. We helped a large retail outlet who had expanded at unprecedented rates. There data could not be processed within the windows available to them because of the logistics of the manufacturing and retail teams covering the globes. These challenges need to be addressed and highlight the need for the steering group to be involved early and have a broad coverage of the moving parts.
What do you think?
Thanks for reading,
John
In this post we take a look at Big Data and focus on the building blocks or foundations of the platform and considerations, this is a topic we are going to be looking at in more detail over the common weeks.
Doing it because you could rather than because you should:
Big issue with Big Data. This is something that is common place and this is usually indicative of the disconnect between the business and the technical team. There are many moving parts inside Business Intelligence / Analytics programs and this is intensified in the big data concept. The result of this is that you have a solution which is fantastic but is disconnected from the business.
There are many techniques to counter this but one way which we have used with success with many customers is to have a Steering Group which is based on teams from the various parts of the organisation who come together as equals with senior buy in. This is not a guarantee of success and this must be time boxed and have a strong chair. The teams involved should be made up of the technical elements (Data Owners, Master Data Management, Infrastructure etc) and from the business there should be representatives of the line of business and users. The key benefit of this is that everyone is aligned from the beginning and there is broad agreement of the vison for the project - what it will achieve, why it is happening, the expected ROI and justifications.
Visualisation techniques:
Despite this being a hot topic for several years this is still an area of neglect. A lot of effort is put into the data and this is absolutely essentially. However without a platform for business users to simply and effectively interact with data the solution will become a white elephant.
There needs to be an interface which allows users of all levels of capability to use and gain insights. The consensus that senior managers need a high level and information workers need to be able to access the details is too restrictive. Dashboards are not the answer for the execs and low level detail is not always the answer for the information workers. Everyone in the organisation is a consumer of the information and therefore should be able to traverse through several layers of detail and complexity simply and still understanding the context. There needs to be a story to set the scene and make it simple for users to understand what is being looked at. The old mentality of assuming that senior people are not tech savvy or only want to see a high level is flawed.
At 5point9 we want to enable all users (with the right security) to be able to understand the 35,000 ft view but also get in amongst the reeds of their business. The social network angle is still woefully in adequate in organisations but this is beginning to grow with the use of collaboration tools and the integration of these with the data and information. Large organisations are now embracing instant messaging but are stil lagging behind on the social developments - we are developing a post on this soon, watch this space.
Actionable information:
We have a rule at 5point9 with customers for Big Data and BI - if we cannot succinctly describe the actions which will result from something we do not include it. This can be a heated activity but one which customers delight in. Once this ethos is distilled we can see real change within the customer and a much improved focus. We want to be a sniper rather than shooting from the hip.
One outcome of this approach is that there is absolute business clarity from the data right the way through to the knowledge used by the business. It can often highlight areas of missing information or additional details that are required for data to become actionable.
Security and Audit:
Do you have a handle on who can see what data and how is this being managed? Is the decision maker / team deciding this correctly placed? Has the business being consultated and what is the process for users to change what they can see? This is a big area and one which can easily frustrate users.
There is a balance between restricting users to the correct level and also allowing them enough access to see the information they need to do their job effectively. This is also massively influenced by regulations and industry specific standards. As well as security access the audit-ability of what users have seen needs to be considered and as big data is more prevalent this becomes key to be able to report on this.
Data Governance:
Master Data and Data evolves over time as does security and users requirements (visualisation etc). We need to ensure that vehicles are put in place to handle this. Big Data programs which do not iterate or allow for change will quickly become dormant and redundant.
Performance:
If the performance is slow or the tool is cumbersome or the users have issues, they will create silos or extractions from the repository. You users are resourceful and smart - which is why you have hired them. Is this happens you quickly revert to having pots of information without governance and control or consistency. You lose the ability to ensure that data is being used across the organisation consistently. The simple way to do this is to ensure the performance of the system and ensure that users have the correct level of access in time and accurate.
Data availability must also be processed within acceptable time limits. We helped a large retail outlet who had expanded at unprecedented rates. There data could not be processed within the windows available to them because of the logistics of the manufacturing and retail teams covering the globes. These challenges need to be addressed and highlight the need for the steering group to be involved early and have a broad coverage of the moving parts.
What do you think?
Thanks for reading,
John
No comments:
Post a Comment