Slide 10 1 Chapter 10 Managing Information Quality

Slide 10. 1 Chapter 10 Managing Information Quality Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 2 Managing information quality • After this lecture, you will be able to: – Assess approaches for managing information quality. – Understand the relationship between data quality, information quality and knowledge quality. – Define the purpose and process of an information audit. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 3 Learning outcomes Typical questions facing managers related to this topic: – How do we highlight the importance of information quality to employees? – What approaches should we use for managing information quality? – Who is responsible for information quality? Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 4 Information quality – does it matter • Is anyone bothered about information quality? • No, but information quality essential to … – improve business performance (financial, operations, marketing metrics) – Inform decisions about all aspects of business – Understand the fast-moving business environment – Understand customer needs, wants and behaviour = Customer intelligence managers – Knowledge management Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 5 A process for improving information quality Figure 10. 1 A process for improving information quality Source: BIM Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 6 What is quality? • Lillrank (2003) reviews the definition of information quality in detail. • He says there are four types of quality definitions: – conformity to requirements – meeting or exceeding customer requirements, – excellence – relative quality – value for money Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 7 The ‘DIKAR’ model of the transformation of Data to Information to Knowledge leading to Actions and Results Figure 10. 2 The ‘DIKAR’ model of the transformation of Data to Information to Knowledge leading to Actions and Results Source: Adapted from Murray (2000) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 8 Different types of quality • Information quality is dependent on data quality assessed through attributes such as accuracy, completeness, validity and consistency as described later in this section. • Knowledge quality is dependent on information quality assessed through attributes such as relevance, presentation and timeliness. • Knowledge quality is also dependent on the experience of the manager and their access to tacit and explicit knowledge. The quality of actions and results is dependent on the capabilities of managers to create, analyse and action performance management metrics. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 9 Problems due to poor quality data Figure 10. 3 Problems due to poor quality data Source: Data Warehousing Institute (2002) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 10 Data quality • For English (1999), data quality is equivalent to inherent information quality. • He defines this as the degree to which data accurately reflects the real-world object that the data represents, whether this is information about a person, business event or other type of objects. • Data quality or inherent information quality refers to the quality of source data that is used to build information resources referenced by managers. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 11 Data quality – some examples 1. Accuracy means that the data correctly defines the event or object which it describes. For a post-code – accuracy means does the postcode actually describe where the person lives? Data could be inaccurate if the customer has moved address since the data was collected, i. e. the data is out of date. 2. Completeness refers to whether all the data is present. A company could profile its post-code or Zip code field for completeness. It may find that these are only recorded for 80% of customers. This means that is difficult to identify where a fifth of customers live which is important for targeting relevant communications to them which are often based on the lifestyle of people living in a particular area. 3. Validity means that the data falls between acceptable ranges defined by the business. For a post-code there will be a standard format in every country that is required for valid data. Validity does not necessarily mean the data is accurate. For example customer birth dates must be within a defined range such as from 1900 to the present day, but are inaccurate if the birthdates are not correct for some reason. Validity is sometimes defined as congruence with business rules. 4. Consistency means that the data elements are consistently defined and understood. With a UK post-code for example, some people entering the post code ‘DE 22 1 GB’ may not realise that a full post-code is required and may have entered the data as ‘DE 22’ – the postal area. Consistency is particularly important where different sources of information are used in an analysis. For example, are customer enquiries measured in a similar way in different countries? Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 12 5 types of customer-related data 1. Demographic. Characteristics of an individual customer such as age, sex, date of birth, address and phone number. This is largely static, but when a customer moves contact details will require updating, this is a major challenge in customer data quality. 2. Transactional. Information about business events or interactions involving customers. This includes purchases and contacts via phone or e-mail. 3. Behavioural. Information about how a customer interacts with the company. This includes clickstream data obtained from web and e-mail. 4. Relationship. For consumers it may be useful to know relationships between family members such as husband wife, mother and daughter. For business customers, it is useful to know the relationship between different people in the business. 5. Derived. Data collected by analysis tools such as customer profitability or lifetime value, customer growth potential and propensity to purchase. Derived data is strictly information since it produced by information processing. This data is of particular concern in data protection since it may be necessary to make customers aware of it. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 13 Mini-case – Abbey and Avellino • Abbey Bank uses a software tool known from data integrator Avellino whose company slogan is ‘Know you data, Know your Business. ’ • Their software ‘Discovery’ was used by Abbey Bank to better understand its data and ensure marketing, customer service and other operations are supported with high-quality, accurate and complete customer information and business intelligence from a central enterprise data warehouse. • Discovery enables the bank to better understand the data structure within the warehouse and other databases and to pinpoint issues for resolution. • By doing so, the bank has been able to cost-effectively raise data quality levels. It has also cut the time it takes to analyse the accuracy of customer data by 90 per cent, easing pressure on IT resources and improving marketing efficiency since more relevant communications can be delivered to customers. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 14 Mini-case – Abbey and Avellino • Abbey Bank breaks down the estimated Net Present Value of savings assisted by Avellino Discovery (so far) as follows: – More efficient integration of Abbey Bank data into the central data architecture delivered £ 1 million (US$1. 5 m) NPV. – Migrating databases into the central data architecture can now be achieved faster and with more accuracy. The total benefit is estimated at £ 2. 5 million ($4 m) NPV. – Data-cleansing activities, assisted by analysis provided by Avellino Discovery, has netted the bank £ 1. 7 million ($2. 7 m) in savings. – More accurate bad debt provisioning and credit scoring has led to well over £ 10 million ($16 m) in NPV benefits. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 15 Information quality • Pragmatic information quality. This is described by English (1999) as the degree of usefulness and value data has to support the enterprise processes. He notes that information that is of high inherent information quality could be of low value pragmatic information quality. Information quality = f(Data quality + Definition + Presentation) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 16 Attributes of pragmatic information quality 1. Relevance. A vital aspect of information quality – it must directly support a decision that needs to be taken or a task that needs to be performed. Relevance relates closely to knowledge quality. Identifying relevant data is a key method of countering information overload. 2. Presentation. The data must be presented in a form that makes it easily understandable. This is dependent on the type of information. For geographical-related information, presentation of regional variations on maps may be most appropriate. For other applications tabulations, spreadsheet pivot tables or different forms of graphs may be the best format. The information must be of the correct level of detail as explained below. 3. Timeliness. Information needs to be up-to-date for most applications. How current the data needs to be depends on the problem. To analyse market share information across a country, data up to a year old might be acceptable. But to analyse data about promotions on flights it is important that the data is made available as soon as possible so that learnings can be incorporated into future marketing campaigns. If information takes a month to collate, then its relevance will decrease. Xu et al. (2003) note the conflict between timeliness and accuracy for accountants. They say that if only 5 or 6 days are available for producing end-of-month reports then accuracy decreases as estimates rather than exact figures may have to be provided. 4. Availability. Information that is of high quality in terms of relevance, presentation and timeliness may still be of low quality from the perspective of a manager who is unable to access it because it is held elsewhere in an organization or they do not have access to it. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 17 Real-time information about pages recently viewed on a web site Figure 10. 4 Real-time information about pages recently viewed on a web site Source: Marketing Online (www. marketing-online. co. uk) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 18 Aggregated or rolled up information on pages viewed on a web site Figure 10. 5 Aggregated or rolled up information on pages viewed on a web site Source: Marketing Online (www. marketing-online. co. uk) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 19 Drill-down to search key phrases used to reach a web site Figure 10. 6 Drill-down to search key phrases used to reach a web site Source: Marketing Online (www. marketing-online. co. uk) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 20 COBIT information quality attributes Information quality requirement Definition Example from Lo-cost Airline Company for route data introduced in chapter 1 1 Effectiveness Deals with information being relevant and pertinent to the business process as well as being delivered in a timely, correct, consistent and usable manner. The provision of flight capacity information for an individual route which can be used to improve the process of marketing and scheduling flights on a route. 2 Efficiency Concerns the provision of information through the optimal (most productive and economical) use of resources. The cost of collection and dissemination of the information above should be economic. 3 Confidentiality Concerns the protection of sensitive information from unauthorised disclosure. 4 Integrity The information above should be protected from release to outside sources. Relates to the accuracy and Refers to data quality of above information i. e. completeness of information as well is it correct (although confusingly this term is as to its validity in accordance with also used under effectiveness). business values and expectations. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 21 COBIT information quality attributes Information quality requirement Definition Example from Lo-cost Airline Company for route data introduced in chapter 1 5. Availability Relates to information being available when required by the business process now and in the future. It also concerns the safeguarding of necessary resources and associated capabilities. Is this information timely i. e. how long does it take for the reports on capacity to be produced? Note again, this is also referred to under effectiveness. 6. Compliance Deals with complying with those laws, regulations and contractual arrangements to which the business process is subject, i. e. , externally imposed business criteria. Since the company is registered in the US, flight data will have to comply with Sarbanes Oxley, i. e. it will have to be accurate. 7. Reliability Relates to the provision of appropriate information for management to operate the entity and for management exercise its financial and compliance reporting responsibilities. Requires that the information above is consistently of good quality, i. e. relevant, accurate, timely, secure, etc. This also overlaps with other categories. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 22 Knowledge quality • Like information quality, relevance is the key to knowledge quality. The knowledge must be relevant to analyse information, take a particular decision or inform actions. • Knowledge quality is not a concept that is as frequently used as data quality or information quality, but English (1999) has developed a similar ‘equation’ to that for information presented above, suggesting the relationship between knowledge and the factors that determine its quality: Knowledge quality = f(People + Information + Significance) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 23 Influences on knowledge quality The elements of the ASHEN model introduced in Chapter 5 are effectively elements of knowledge quality. They are: • Artefacts Any thing made by people, processes, documents, tools in which knowledge is imbedded • Skills Abilities that can be trained and measured without ambiguity, but remember the time issue • Heuristics Rules of thumb, the outcome of experience, the main repository of knowledge mostly unarticulated • Experience Accumulated experience of failure and success which allows the right pattern to be triggered in the right context • Natural talent Some people are just better at doing things than other people – and they are often not the people you expect Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 24 Where ASHEN can be applied • The ASHEN factors can be used to assess knowledge quality at different Knowledge Disclosure Points: • Decisions • Problems resolution • Solution creation • Judgement • Learning points Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 25 Performance Management systems • The quality of actions taken by managers and results delivered in an organization depends on the processes and systems used to manage and control the direction of the organization. • The processes and systems intended to monitor and improve the performance of an organization are known as performance management systems. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 26 Peformance Management defined • Andy Neely of Cranfield School of Management’s Centre for Business Performance defines Performance Management as: ‘The process of quantifying the efficiency and effectiveness of past actions through acquisition, collation, sorting, analysis, interpretation and dissemination of appropriate data. ’ (Neely, 1998) Performance management system A process used to evaluate and improve the efficiency and effectiveness of an organization and its processes. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 27 Efficiency and Effectiveness Meeting process objectives, delivering the required outputs and outcomes. “Doing the right thing” Efficiency Minimising resources or time needed to complete a process. “Doing the thing right” Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 28 Corporate or Business Performance Management • Corporate or Business Performance Management. Cognos, a vendor of applications to support corporate performance management describes it thus: • ‘The aim of Corporate Performance Management (CPM) is to integrate a number of hitherto discrete applications into a single environment that includes all the necessary elements of performance management. ’ • ‘These include: strategic and tactical planning; financial considerations such as budgeting and consolidation; the use of key performance indicators to support scorecard analytic applications; the ability to monitor events on a real-time basis and to notify managers accordingly; and, of course, conventional query and reporting capabilities. ’ Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 29 Business Peformance Management (BPM) • Business Peformance Management (BPM) is supported by the BPM standards group of analysts and vendors who are seeking to create standardised processes for process (see www. bpmstandardsgroup. org). • They define BPM thus: – BPM is a set of integrated, closed-loop management and analytic processes, supported by technology, that address financial as well as operational activities. – BPM is an enabler for businesses in defining strategic goals, and then measuring and managing performance against those goals. – Core BPM processes include financial and operational planning, consolidation and reporting, modelling, analysis, and monitoring of key performance indicators (KPIs) linked to organizational strategy. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 30 The role of information in controlling organizational performance Figure 10. 7 The role of information in controlling organizational performance Source: BIM Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 31 Ease of decision makers to access information Figure 10. 8 Response to question: ‘How easy is it for decision makers in your organization to combine operational and financial information, and to do so when measured against corporate objectives? ’ Source: Consultants Advisory (2003) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 32 Executive portal Figure 10. 9 Factiva. com executive information portal showing summary of external and internal business information Source: ww. factiva. com Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 33 An example of a performance measurement system for an e-commerce electrical goods retailer Figure 10. 10 An example of a performance measurement system for an ecommerce electrical goods retailer Source: Friedlein (2002) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 34 Six Sigma • Six Sigma is a label for a quality measure and improvement program originally developed at Motorola that focuses on the control of process to the point of ± six sigma (standard deviations) from a centerline, or 3. 4 defects per million items. • As the standard deviation increases towards six sigma, the number of faults decreases and quality of a process or service increases as shown in Table 10. 2. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 35 Six Sigma • Truscott (2003) describes its purpose as: ‘To provide a universal performance metric, or measure, that can be applied to any product, process, service regardless of its relative complexity; a world class performance benchmark and the marketing name for the Six Sigma improvement initiative. ’ Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 36 Six Sigma example Six Faults or events Percentage Sigma per million opportunities of faults or value events Time without electricity per month 1 691, 462 30. 85% 464 h 2 308, 538 69. 15% 207 h 3 66, 807 93. 32% 45 h 4 6, 210 99. 38% 4 h 5 233 99. 78% 9 min 6 3. 4 99. 99% 8 sec Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 37 SMART • Specific – Is the detail in the information sufficient to pinpoint problems or opportunities? • Measurable – Can a quantitative or qualitative attribute be applied to create a metric? • Actionable – Can the information be used to improve performance? (Achievable) • Relevant – Can the information be applied to the specific problem faced by the manager? • Time-related – Can the information be viewed through time to identify trends? Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 38 Ten measure design tests 1. The truth test. Are we really measuring what we set out to measure? 2. The focus test. Are we only measuring what we set out to measure? 3. The relevancy test. Is it the right measure of the performance measure we want to track? 4. The consistency test. Will the data always be collected in the same way whoever measures it? 5. The access test. Is it easy to locate and capture the data needed to make the measurement? 6. The clarity test. Is any ambiguity possible in interpreting the results? 7. The So-what test. Can and will the data be acted upon? 8. The timeliness test. Can the data be accessed rapidly and frequently enough for action? 9. The cost test. Is the measure worth the cost of measurement? 10. The gaming test. Is the measure likely to encourage undesirable or inappropriate behaviours? Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 39 Why measure - traditional? Traditional: 1. To track recent/ current actual performance against targets/predictions/history. 2. To track recent/ current performance against external regulations or internal policies. 3. To track perceptions of performance deficiencies and monitor their improvement. 4. To motivate managers and employees to achieve specific performance objectives. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 40 Why measure 2 Emerging: 1. To help predict future trends. 2. To validate or challenge existing assumptions. 3. To discover new insights (through data analysis). 4. To stimulate the creation of new initiatives, objectives and targets. Ultimately: 1. To aid decisions and substantiate improvement/investment recommendations. 2. To show the achievement/realisation of anticipated benefits resultant from actions. Source: Neely et al. (2002) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 41 The information audit An information audit has been defined by the Aslib Information Resources Management Network, referenced in Orna (1999), as: ‘A systematic examination of information use, resources and flows, with a verification by reference to both people and existing documents, in order to establish the extent to which they are contributing to an organization’s objectives. ’ Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 42 Applications of an IA Information audits may target a number of specific information management needs or applications which might include: 1. Creation or refinement of an intranet or extranet. 2. Improvement of library or information centre services. 3. Reduction in time spent searching by staff for internal or external information. 4. Corporate Performance Management. 5. An individual business process, department or data type, e. g. financial data, inventory data or customer data. 6. Competitive intelligence about an organization’s marketplace and competitors. 7. Knowledge Management (see chapter 6 which defines the process for a knowledge audit, which is one form of information audit). Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 43 Benefits of an IA (Orna, 1999) Short term benefits Long term benefits Attention to immediate threats, and exploitation of opportunities Quick financial and efficiency gains from making information more available and usable Increasing awareness of presence and location of all necessary information Enriched understanding of what information and knowledge mean throughout the organization Development of a strategy for managing knowledge and information Better use of information in supporting key business processes. Integrated management of the full range of information Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 44 More arguments for an IA • Nigel Oxbrow, managing director of TFPL a London and New York based company specializing in information audits sees the benefits of as information audit as ‘a database of information resources, improved understanding of information costs and value, and improved quality of information services. An increase in awareness about information, accompanied by changes in user expectation and patterns of sharing are also expected by products’ (Di Mattia and Blumenstein, 2000). • He describes the risks in terms of information and knowledge quality of not conducting the audit as: • ‘duplicate, incomplete or inaccurate resources and work; inefficient use of an intranet; innovative ideas that don’t get shared; and intellectual assets that are not fully utilised’. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 45 1. 2. 3. 4. 5. 6. 7. Orna’s questions about for an IA What are they? Details on company performance. Where are they? They are produced in paper form for distribution when requested. Also available online in PDF and HTML formats. Who is responsible for them? Responsibility for timely delivery of the investors’ relations information is the Head of Investor Relations. The accuracy of information is the responsibility of chief financial officer and as mentioned earlier is now subject to the Sarbanes–Oxley Act. What kind of information do they contain? Financial statements, quarterly company trading statements, annual reports and presentations by the management team. How do people who manage them define the users and the way they are used? This has not been thought through (see 7) since it is a legal requirement to produce them and if people such as investors require them they will be sought out. What do the users themselves say? The users would include potential and actual investors, the media and researchers including other companies. Key users such as potential investors could be asked to assess the data. Are there other people who could make good use of this resource who: 1. Don’t know about it? Newly joined managers may find this information useful in 2. induction, but they are not currently made aware of it. Know about it, but don’t have access to it? Not relevant since available on line. + Information lifecycle questions Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 46 Swash’s view Swash (1997) stresses these issues of information quality in her summary of questions needed in the information audit. She recommends: • ‘Audit questions should seek to identify what information is central to business need, what sources are actively used and how often. • These data should preferably be weighted in order of priority. • The problems arising through the non-availability of timely and accurate information should be recorded any discrepancies between perceived and actual needs quantified. • Further questions should be included to explore the use of primary contacts, both internal and external which will frequently be more highly valued than published sources. ’ Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 47 Information audit questions relating to an organization’s objectives Figure 10. 11 Information audit questions relating to an organization’s objectives Source: Orna (1999) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 48 Information policies • An information policy supports an information strategy by setting out the high-level principles of how an organization should use organization. • It refers to general aims such as using information to improve organizational performance, ambitions to share knowledge and the intention to secure information. • It will also typically refer to identifying staff responsibilities for information management. • What we believe is also needed in addition to this high-level policy is an information quality policy giving more detailed practical guidelines on managing information quality and data quality. • This includes procedures to evaluate and improve information quality. Information quality is controlled by staff actions and checks built into business applications and database. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 49 Managing information quality English (2000) stresses the need to manage information quality. He says that managing information quality involves: 1. Raising awareness of the problems caused by poor quality information (through education and training). 2. Defining processes for measuring information quality (the information audit is the first stage in this). 3. Defining processes for improving information quality. 4. Providing education and guidance on facilitating the improvement in these processors. 5. Measuring the cost savings and customer satisfaction that results from improved information quality. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 50 Information management controls Xu et al. (2003) distinguish between systems controls and human controls on information quality – these are the means by which processes are put in place to improve information quality. In research on accountants these authors found that IT people tend to think that systems controls are most important, while accountants tend to think that human -related factors are more important. The accountants believed that system control rules cannot be built in for every aspect of data quality since judgement is required. They argue that education and training are critical. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 51 Sources of data quality problems A Data Warehousing Institute (2002) survey which looked at data quality in all types of database system found that the main causes of data quality problems were: • Data entry by employees, for example when talking to a customer on the phone (76%) • Data entry by customers, for example when entering their details over the Internet (25%) • Changes to source systems i. e. changes to database definitions made by database administrators without thinking through information quality implications (53%) • Data migration or conversion projects (48%) • Mixed expectations of data quality by users (46%) • Errors introduced by the import of external data (34%) • System errors causing corruption of data (26%) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 52 Data validation approaches • Data type checking. – Numeric, character, date • Data range checking. – Sufficient digits • Input limits. – X > Value > Y • Restricted value checking. – Unique fields selected, e. g. job titles, valid postcodes • Unique value checking. – Key fields, e. g. customer id • Multiple field validation. – Consistency between fields • Referential integrity checking. – Checks links between different tables in a database Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 53 Data quality controls • Data quality auditing. This is a combination of systems control and human control. Data quality auditing is performed on actual records in the database. This can be done on an ad-hoc basis through performing database queries designed to test data quality. This is easier for some aspects of data quality than others. • Data profiling (Olson, 2002). This analyses the data quality within the context of the data dictionary (Chapter 9). Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 54 Data cleansing – de-duplication • There are two main forms of database cleaning which are known as de-duplication and hygiene processing. De-duplication is needed to reduce the occurrence of multiple records about the same customer. • The process of de-duplication or ‘de-duping’ involves an initial matching on certain fields such as name and address to identify potential duplicates and then deleting these duplicates while minimising overkill (to avoid deleting records which appear to be matched, but are different) and underkill (records that are not matched, but are, in fact duplicates). • Matchkey processing is most commonly used, this involves taking the first letter of the different components of name, position, company and address. Fuzzy logic algorithms are now used to identify similarities. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 55 Data cleansing – hygiene processing • Hygiene processing consists of 3 steps that are usually performed before ‘de-duping’: • Formatting – remove punctuation marks, extra blanks. • Parsing – identify main data elements, e. g. company names, street name, postcode, etc. which may be in combined fields. • Validation and enhancement – e. g. a postcode used to check address or the case correctly applied to first names and surnames. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 56 Data quality at Hilton Hotels 1 The daily data processing needed for a major international company such as Hilton Hotels is significant. Each day 60, 000 records are updated from the 3, 000 Hilton group hotels in 65 countries. Over a three-year period, Hilton has assembled a database of 14 million profiles each containing detailed information such as: – Names and addresses – Line of business – Spending history – Preferred payment method – Staying preferences from favourite rooms to favourite pillows Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 57 Data quality at Hilton Hotels 2 • Customer Information System (CIS). The system screens for data quality. On average, 58% of addresses are returned as unmailable and a further 30% are incomplete. A manual follow-up is then completed to correct details where appropriate. If these details are not accurate then some of the millions of dollars used for direct marketing by post, e -mail and the web site are wasted, either due to messages not getting through or the offers not being relevant. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 58 Data quality at Hilton Hotels 3 • • • The attention that Hilton International attaches to data quality is indicated by these additional measures to improve data quality: Hilton uses external partners to train local data controllers in each country Data quality targets Monthly HHonors team conference where data quality is on the agenda Quarterly CRM mini-boards for senior managers Weekly data quality steering groups Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 59 Data quality responsibilities Figure 10. 12 Data quality responsibilities Source: Molineux (2002) Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 60 Improving Information quality 1. Relevance – use approaches in next slide – audit. 2. Presentation – use of specialist tools or manually crafting spreadsheets. 3. Timeliness – requires targets for timeliness and integration of systems, e. g. closing books for month in accounting. 4. Accessibility – managing integration of different sources through data translation XML or sourcing integrated systems such as ERP. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 61 Controlling relevance automatically – an example • These types of service can provide relevant information about competitors and markets by e-mail alerts using a number of systems control techniques: • Keyword filtering – only sends news articles that include keywords such as company names or product names. • Range filtering – only sends information about a change in share price if it exceeds a certain percentage. • Grouping content – e. g. by geography, industry, competitors. • Aggregated alerts – options to receive individual alerts or digests for the entire day or week by different topics. Chaffey and Wood Business Information Management © Pearson Education Limited 2005

Slide 10. 62 Relationship between the main applications at the games retailer Figure 10. 13 Relationship between the main applications at the games retailer Source: BIM Chaffey and Wood Business Information Management © Pearson Education Limited 2005