Visualising Data Uncertainty

Given that good data quality is difficult to manage and maintain in organisations, considerations should be given to designing BI reports that either visualise data uncertainty or indicate the margin of error instead of displaying information as being 100% accurate. In this way, decision-makers can be made aware of the risks that they might be taking when acting upon a BI report and be in a position to plan better and manage expectations.

Data quality issues often begin at the point of data entry into a transactional system and find their way into BI applications. Some data quality issues can not be totally eliminated even with a good data quality programme. The main problem is that the operational priorities of the organisation are usually quite different to its analytical requirements. For example, to complete a transaction, a call centre operator might have to complete a form on a sales system, whilst talking to the customer on the telephone. Faced with a long set of fields to complete on the form, the operator is likely to pay careful attention to the fields that are important for the completion of the transaction, e.g. the name and the address of the customer, and not so much attention to the fields that are needed for later analysis of the data e.g. the location of the office in which the operator is based. A long list of options for office locations drops down on the form but the system is slow and the operator accidentally clicks on "Aberdeen" - the first from the drop-down list, as opposed to "Watford", the last option on the list that happens to be where s/he is actually based. The system gets a valid answer but a wrong one that if repeated over a period of time, will show the wrong totals for sales per call centre.

The important thing is to recognise that data quality can be less than perfect and not to delude the organisational-self that there is no possibility of errors in BI reports.

The Bank of England has been using fan charts for many years to indicate the degree of uncertainty in its various reports e.g. the GDP report shown below. It is interesting to note that the fan chart is used to indicate both past and predicted changes in GDP, but the trend is clear and given that, decisions can be made based on this data within the clearly indicated margins of error or levels of uncertainty.

This is a very visual way of indicating data uncertainty but other ways, e.g. simply stating the margin of error, could work too .

Search: Next Big Thing for Mobiles

The adoption of mobile phones to access information and digital content on the web is growing fast, but the mobile search user experience still leaves much to be desired. If mobile search doesn’t see dramatic improvement, it will hinder device sales and delay the advent of what could become a huge market for mobile advertising. Mobile search should be easy and give results focused on answers, not links.

Read the full article here: http://www.telecoms.com/15178/searching-for-the-next-big-thing

I Predict a Riot of Analytics in the Media Sector

There is much talk of next generation broadband - speeds of 50 Mb/second - becoming generally available. Fast broadband will bring growth in on-demand media downloads, TV viewing on the Internet, games, etc. thereby opening up more opportunities for analytics - a riot of analytics.

We already have Web site visitor data such as: number of visitors, clicks and downloads, traffic sources, map overlays, first time visitors and other visitor trending data such as loyalty, browser capability, language, and more.

In the future we are likely to see other data delivered by analytics applications. For TV and radio programmes that could mean more than just numbers of views/downloads. It could mean detailed data on viewer behaviour such as tuning out from a programme at an advert break, at a particular scene, at the point when an actor entered the scene, or when a presenter started to talk, following a comment, at specific points in the script etc.

This type of highly granular analytics would enable media companies to learn customer behaviour and likes and dislikes really well. It would enable the companies to promote and deliver personalised programmes of entertainment to returning viewers. The advertising could also be customised for the viewer’s taste, with special offers timed to last for short periods of time e.g. whilst a movie is being viewed.

This is just scratching the surface, the possibilities are huge.

Operational Efficiency in the Public Sector

Here are some of my initial thoughts on the OEP, extracted from a Butler Group OpinionWire published on 29 April 2009:


Last week, in the UK’s annual budget statement, the Chancellor of the Exchequer set out his plans for efficiency savings of UK£15 billion per year across the public sector. The plans include savings of UK£7.2 billion from annual IT and back-office costs. These are significant figures and are going to be taken into account by the Government in departmental settlements. On the face of it the pressure is on the sector to act quickly, but we question if that is likely given the sector’s ongoing commitments and existing programmes for efficiency savings.

Modernisation of public sector services through IT has been going on for some years. The e-Government programme of the early ‘noughties’ led the way, largely through modernisation of front-office systems. Since then the focus has shifted to shared services (Transformational Government) and to business processes such as reducing avoidable contact (The Varney Report and National Indicator 14), that is aimed at making the customer journey through public services a positive and efficient experience. The IT element of the Operational Efficiency Programme (OEP) announced alongside the budget last week takes modernisation to the next level; it increases focus on back-office systems with efficiency gains of UK£4 billion estimated in back-office operations such as Human Resources and Finance. Widespread benchmarking and targeted performance reviews across the public sector are set to drive out inefficiencies, such as a 300% variation in the cost of Human Resources per employee. A further UK£3.2 billion a year of savings is expected to be realised by strengthening IT project governance and transparent scrutiny. This is to be achieved through standardised benchmarks to drive up performance and improve IT procurement.

One of the recommendations of the report is that management information on IT expenditure is fully integrated into departmental processes to be collected on a regular, consistent, and auditable basis. This is to enable public sector IT efficiency to be compared and contrasted to derive efficiencies. However, with only a month to go to the start of the programme, some organisations will be forced into taking a manual, labour-intensive approach to OEP data collection and reporting. Although Performance Management (PM) software is not essential for this kind of reporting, it would be very advantageous in automating an otherwise manual process. Furthermore, without PM it would be very difficult to integrate performance requirements into departmental processes to derive efficiencies at granular levels. According to Datamonitor’s Technology Trends CIO Survey, as of 2008, only 33% of public sector bodies in the UK had invested in PM software. A further 38% were planning to invest in the software in the following six months. Assuming that all of those intentions turned into completed projects, we would still have 29% of organisations without PM software. Moreover, the Datamonitor survey showed that 50% of respondents were planning to invest in this type of software in the next six to 24 months. This implies that some public sector organisations are dissatisfied with existing PM deployments and would be currently engaged in projects to replace them. Therefore, for many, there is still work to be done in this area.

During the e-Government programme of the early 2000’s, local authorities received funding for the modernisation of systems and were required to make annual Implementing Electronic Government (IEG) statements to the Government. That process started as a basic form-filling exercise but over the following couple of years developed into the Electronic Service Delivery Toolkit (esd-toolkit) that semi-automated the process and enabled local authorities to complete their IEG statements online.The esd-toolkit was used to develop standardised service definitions that has further simplified the work of reporting by local authorities and has evolved into a shared services vocabulary. The Cabinet Office is working on standardised metadata that can be used by the whole of the public sector. However, this is work in progress and metadata has to be published and widely adopted before it starts to be of use – and so the automation of reporting across the whole of the public sector has to wait for a while yet.


Another IT issue that the sector faces in the wake of the OEP report is that many have already committed to IT projects that are driven by other Government mandates: e.g., the single view of customer for the NI14 requirement, the National Programme for IT in the National Health Service, and the ID Cards programme in the Home Office. The OEP is a well thought-out and much needed programme but it comes during a tough economic period and after a number of other similar requirements and initiatives (e.g. Transformational Government, the Varney Report, and CSR07).


There is work to be done in preparing for the OEP and achieving its objectives in the way that they are intended but given the public sector’s ongoing commitments, existing efficiency drives, and the proximity of a general election, the OEP is likely to end up largely as a numbers exercise – that is for the short term anyway.

Metadata Standards for Public Services

Public sector services typically involve inter-organisational collaboration to deliver complete outcomes to customers. So much so that what would appear as a simple request from the customer’s perspective could be a complex process involving a large amount of information crossing organisational boundaries. The processes are seldom automated and often involve the customer having to give the same information to different government bodeis a number of times. In the UK, there is currently a lack of standardised information about services, and accessing information and securely processing related data causes, delays, and inefficiencies. This has led to a new initiative by the Government to eradicate avoidable contact – that is contact with the public sector that adds no value to the service that is being sought by the customer. The move is throwing new light on the importance of semantic clarity for inter-organisational processes. Semantic clarity is needed to help customers find the right information as well as to help the public sector deliver joined-up services . The objectives can be achieved through the standardisation of metadata, data that defines information and services. The metadata would help bring semantic clarity across the board to establish a common service vocabulary to be used in all collaborative service delivery scenarios. Furthermore, metadata is closely linked to processes and where inter-organisational processes are concerned, efficiencies can be gained by integrating and automating those processes. Therefore a standardised metadata can  help the public sector with process integration too.

Many public sector bodies have already developed their own metadata standards to use as part of Document and Records Management or Enterprise Search and Retrieval deployments. Therefore, they are familiar with the benefits that standardised metadata can deliver. A common standard for all the public sector would extend the range of benefits; for example, it would enable content to be accurately classified and linked to the standardised metadata so that information could be found more easily and automatically through search and discovery processes across not just one but all government Web sites, delivering a vertical search capability. 

It would allow a customer’s request for information, to be linked to detailed electronic content structured around the metadata. Moreover, it would allow the most relevant service being requested by the customer, to be automatically identified regardless of the terminology used and linked to other appropriate services. For example, a request for information on registering a child at a school in the area could be processed irrespective of the wording of the request. It could also automatically link school registration information from the education authority to the free school meal service offered by another local authority such as the district council. Additionally, highly relevant content can be structured around the metadata for use in Frequently Asked Questions, electronic forms, and contact details.

There are already several metadata standards in operation in the UK public sector: e.g. eGovernment Metadata Standard (eGMS) and its subset, Integrated Public Sector Vocabulary (IPSV). These are in need of updating and are not adequately supported by the sector. Furthermore, they do not have a governance structure, and have gaps and duplications in the definitions. There is also the Local Government Services List (LGSL) that is well supported and widely used and which has communities of practice too, but it is limited to Local Government. The Cabinet Office in the UK is working on developing a broad metadata standard to optimise the use of IT resources in the sector, and to align organisation and IT strategies as part of the Transformational Government agenda. The work is being led by the Chief Technology Officer (CTO) Council. The CTO Council has representatives from various central and local government bodies who are jointly working on defining a Cross-Government Enterprise Architecture Reference Model (xGEA). xGEA consists of a number of domains and the metadata work is being carried out as part of the Information Domain by the Metadata and Vocabularies Working Group (MVWG). The group is working on developing the metadata standard, its governance, support, and long-term maintenance. It is also looking at doing proof of concepts on transit formats and mechanisms for sharing information and metadata to determine the easiest ways for interoperability to be achieved across boundaries. However, the metadata development process itself is not well publicised as yet, with the MVWG only just starting to formulate a communication strategy. We look forward to the group going public with its work.

First published in Butler Group Review on 04/06/2008. 

Twitter – the New TV?

Twitter is  known for its capability as a public relations and networking channel but the increase in the number of users makes me think that may be there is something bigger afoot. When I saw the latest follower figures for celebrity Twitterati such as Britney Spears (512,712 followers) and Stephen Fry (328,193 followers), it struck me that they have almost as many followers as some TV shows have viewers. To find out how the follower numbers compared with viewer numbers I did a search on the Internet and found some interesting data. For example, on October 2^nd 2008 according to the Guardian newspaper, BBC4, the digital TV channel, had a very successful evening in terms audience figures. It ran a series of programmes on a theme of "the golden age of steam" and pulled in an average audience of 390,000. This is already lower than the number of Britney Spears’ followers and Stephen Fry’s is not far behind. BBC 2’s Never Mind the Buzzcocks, the music panel show (and a cooler programme than anything to do with the railways) on the same evening had 2.2 million viewers which is a lot more than the most followed Twitterati, but this is only the beginning for Twitter. The following charts on Twitter Grader show how the number of Britney Spears’ and Stphen Fry’s followers have increased over the past few months and as you will see if you click on the links, the graphs indicate fast rates of increase for both:

http://twitter.grader.com/history/britneyspears

http://twitter.grader.com/history/stephenfry

If these continue it will not be long before followers of Twitter elite outnumber viewers of mainstream TV programmes. This will increase the pressure on TV channels to compete with Twitter and I expect to see more TV programmes establish a presence on Twitter for publicity. Other developments are likely e.g. we might even see characters from soaps perform on Twitter to compete with the real life drama of its ordinary users. That would have some interesting effects on ratings and advertising.  

Despite its success, two uncertainties hang over Twitter: The first is regarding scalability - can it scale up sufficiently to cope with the increase in demand for its services. The “Twitter is over capacity” and “something is technically wrong” messages are starting to appear a little too often again after a smoother run in the last few months. Secondly, the longevity of Twitter is not guaranteed given the market that it operates in has a steady flow of disruptive technologies and concepts.

A year can be a long time on the Web.

Defining IT Strategy in the Public Sector

This post covers one of my other areas of interest: Public Sector IT. The following is a copy of an article that I wrote for Butler Group Review last year.

The pressures on the Public Sector to do more with less and deliver further efficiency savings are mounting by the day. Many organisations in this sector are grappling with major projects that require software systems and cutting-edge technology just to respond to Government targets, let alone their own on-going business needs that require support and enablement from IT. The need for an IT strategy that tightly aligns the organisational drivers and IT capabilities has never been greater in the public sector than it is today. However, defining an effective strategy that builds a sound foundation for execution of objectives with flexibility can itself be challenging. A well defined and clearly communicated IT strategy would enable the organisation to respond to changing drivers fast and efficiently.Through a good strategy, an organisation would be able to develop its IT infrastructure and business processes to automate core capabilities, from which it can become more flexible and therefore responsive to business requirements. The IT strategy must take into account the aggregate demand on the IT function, including all drivers – internal and external, the current situation (work in progress and on-going commitments), and proposed actions in support of the strategy.This must be factored into the total of available resources and the current programme of work. In this article I will examine each of these areas . 

To set the scene we will first look at some of the many drivers that affect the UK public sector, as an example, and thence the need for strategic planning for IT. In terms of funding the sector is required to do more with less, and deliver further efficiency savings. It is also expected to become much more citizen and business centric and implement data sharing on an unprecedented scale to make the customer journey through public services as short and positive an experience as possible. This requires major work on data quality and integration as well as process integration to allow disparate organisations within the sector to deal with the citizen’s requirement in a way that minimises “avoidable contact”.The data sharing requirement has come at a time when a number of high-profile data loss scandals have made the public wary of how the sector treats personal data, thereby pushing data security onto the investment agenda for this sector.There are further pressures from the Transformational Government programme to share services to deliver more efficiency savings, from employees requiring flexible and mobile ways of working, and citizens and businesses demanding improved electronic service delivery channels, and better connectivity. Finally there is a requirement from the Government for professionalism in IT, in skills, project, and programme management. The aggregate of all these requirements is indeed a tall order for any organisation and it is essential that when the objectives are driven from this long list of drivers, they are clearly prioritised and matched to resources and timescales.

The IT strategy should address the following areas:

The vision – This is the aspirations of the IT division for a specific period of time. The vision outlines the role that IT is to play in delivering the corporate objectives, the division’s own overall high-level objectives, and expected levels of performance and efficiency to be attained including Service Level Agreements (SLAs).  

The high-level objectives – It is important to be clear about the business objectives for IT and the role that the business expects IT to play – the vital aspect is that the expectation is defined and understood at all levels, so that there are clear objectives against which the IT function can deliver.These would be in support of the vision and are in response to internal and external drivers. Examples of the former include the needs of business units, internal customer relationships, and bottom-up drivers arising out of technological advances. Examples of the latter have already been supplied. Often there are overlapping requirements between the drivers. Therefore, the objectives can be defined in a way that links those overlapping drivers and works towards satisfying several of them at the same time, e.g., the external driver to become more customer centric and to deal with customers efficiently can be linked to employees’ demand for flexible and mobile ways of working. Therefore, one objective could be to increase the availability of handheld devices and connectivity for field workers e.g. social workers, so that they could complete their case work in-situ, in the customer’s home rather than both of them having to go to the relevant office to complete the task.

Current situation – This highlights the work in progress and on-going commitments of the IT division e.g. existing SLAs, resource levels, and budgets. It also defines known issues and bottlenecks to highlight the need for action and change, e.g. too many departmental software solutions that add little value but cost the IT Management organisation a lot of time and maintenance effort. 

Proposed actions in support of the strategy – These are the key actions, in the form of projects or programmes, that the organisation needs to take in response to the identified and prioritised objectives, e.g. decommission a legacy application in favour of a modern vanilla package, build on existing infrastructure for capacity for increased data storage and/or network traffic, and initiate a data quality and integration programme.When defining the actions, it is important to take into consideration how they affect each other and existing work and resource levels. It is also important to plan with the future in mind and take into account scalability and flexibility. These actions would be managed and controlled using a high-level governance structure and procedures, and a measurement, monitoring, and reporting plan. The strategy should examine at a high level the expected impact and resulting requirements for change, how different parts of IT will be affected by the strategy, e.g. procurement, existing supplier relationships, maintenance and upgrades, and IT operations.

Supporting strategies – The IT strategy defines the high-level goals for all of IT. When documented, it would provide a top-level plan which can then be broken down into more detailed component strategies that address specific areas. An upside down tree structure is produced with the IT strategy at the top and others branching off from it.The lower down the branches you go, the more detail would be provided, until finally leading to project- and programme-related documents that define the implementation of the IT strategy. The areas to be addressed in supporting component strategies include but are not limited to:

• desktop; communication and connectivity;
• infrastructure; Information Management (IM);
• Procurement and Supplier Strategy; Integration
• Strategy; and in some organisations,
• Geographical Information Systems (GIS) Strategy. 

Some of these break into further strategies e.g. the IM strategy leads to the data management strategy. To summarise, the IT strategy must tightly align the organisational drivers and IT capabilities. It is important to be clear about the business objectives for IT and the role that the business expects IT to play. It should help the organisation develop a ‘foundation for execution’  that gives the organisation a flexible platform to deliver new business projects rapidly.The IT function should also feed into business strategy based on current and future technologies.

To summarise, the IT strategy must tightly align the organisational drivers and IT capabilities. It is important to be clear about the business objectives for IT and the role that the business expects IT to play. It should help the organisation develop a ‘foundation for execution’ that gives the organisation a flexible platform to deliver new business projects rapidly.The IT function should also feed into business strategy based on current and future technologies.