| Definitions and Overview | | | | when implementing a new BI/ DWH solution. Trained |
| Business Performance Management (BPM) establishes | | | | users are 60% more successful in realising the |
| a framework to improve business performance by | | | | benefits of BI than untrained users. But this training |
| measuring key business characteristics which can be | | | | needs to consider specific data analysis techniques as |
| used to feedback into the decision process and guide | | | | well as how to use the BI tools. In the words of |
| operations in an attempt to improve strategic | | | | Gartner, "it is more critical to train users on how to |
| organisational performance. Other popular terms for | | | | analyse the data." Gartner goes on to say "... that |
| this include; Enterprise PM (EPM), Corporate PM (CPM) | | | | focusing only on BI tool training can triple the |
| Enterprise Information Systems (EIS), Decision | | | | workload of the IT help desk and result in user |
| Support Systems (DSS), Management Information | | | | disillusionment. A user who is trained on the BI tool |
| Systems (MIS). | | | | but does not know how to use it in the context of |
| BPM: Cycle of setting objectives, monitoring | | | | his or her BI/DWH environment will not be able to |
| performance and feeding back to new objectives. | | | | get the analytical results he or she needs...". Hence |
| Business Intelligence (BI) can be defined as the set | | | | bespoke user training on your BI system and data is |
| of tools which allows end-users easy access to | | | | essential. |
| relevant information and the facility to analyse this to | | | | Careful planning of the training needs and making the |
| aid decision making. More widely the 'intelligence' is the | | | | best use of the different training mediums now |
| insight which is derived from this analysis (eg. trends | | | | available can overcome this issue. Look for training |
| and correlations). | | | | options such as: Structured classroom (on or off |
| BI: Tools to Access & Analyse Data | | | | site), web based e-learning (CBT), on the job training |
| Key Performance Indicators (KPIs) are strategically | | | | & skills transfer, bespoke training around your |
| aligned corporate measures that are used to monitor, | | | | solution & data. |
| predict and anticipate the performance of the | | | | Technical Overview |
| organisation. They form the basis of any the BPM | | | | Information Portal: This allows users to manage |
| solution and in an ideal world it should be possible to | | | | & access reports and other information via a |
| relate strategic KPIs to actual operational | | | | corporate web portal. As users create & |
| performance within the BI application. | | | | demand more reports the ability to easily find, |
| KPIs provide a quick indication on the health of the | | | | manage & distribute them is becoming more |
| organisation and guide management to the | | | | important. |
| operational areas affecting performance. | | | | Collaboration: The ability for the Information Portal to |
| In many companies analysis of data is complicated by | | | | support communication between relevant people |
| the fact that data is fragmented within the business. | | | | centred around the information in the portal. This |
| This causes problems of duplication, inconsistent | | | | could be discussion threads attached to reports or |
| definitions, inconsistency, inaccuracy and wasted | | | | workflow around strategic goal performance. |
| effort. | | | | Guided Analysis: The system guides users where to |
| Silos of Data: Fragmented, Departmental Data Stores, | | | | look next during data analysis. Taking knowledge |
| often aligned with specific business areas. | | | | from people's heads and placing it in the BI system. |
| Data Warehousing (DWH) is often the first step | | | | Security: Access to system functionality and data |
| towards BI. A Data Warehouse is a centralised pool | | | | (both rows and columns) can be controlled down to |
| of data structured to facilitate access and analysis. | | | | user level and based on your network logon. |
| DWH: Centralised/Consolidated Data Store | | | | Dashboards & Scorecards: |
| The DWH will be populated from various sources | | | | Providing management with a high level, graphical |
| (heterogeneous) using an ETL (Extract, Transform | | | | view of their business performance (KPIs) with easy |
| & Load) or data integration tool. This update | | | | drill down to the underlying operational detail. |
| may be done in regular periodic batches, as a one off | | | | Ad-hoc Reporting and Data Analysis: End users can |
| load or even synchronised with the source data (real | | | | easily extract data, analyse it (slice, dice & drill) |
| time). | | | | and formally present it in reports & distribute |
| ETL: The process of extracting data from a source | | | | them. |
| system, transforming (or validating) it and loading it | | | | Formatted/ Standard Reports: Pre-defined, pixel |
| into a structured database. | | | | perfect, often complex reports created by IT. The |
| A reporting (or BI) layer can then be used to analyse | | | | power of end user reporting tools and data |
| the consolidated data and create dashboards and | | | | warehousing is now making this type of report |
| user defined reports. A modelling layer can be used | | | | writing less technical and more business focussed. |
| to integrate budgets and forecasting. | | | | Tight MS Office integration: More users depend on |
| As these solutions get more complex, the definitions | | | | MS Office software, therefore the BI tool needs to |
| of the systems and what they are doing becomes | | | | seamlessly link into these tools. |
| more important. This is known as metadata and | | | | Write Back: The BI portal should provide access to |
| represents the data defining the actual data and its | | | | write back to the database to maintain: reference |
| manipulation. Each part of the system has its own | | | | data, targets, forecasts, workflow. |
| metadata defining what it is doing. Good | | | | Business Modelling/ Alerting: around centrally |
| management & use of metadata reduces | | | | maintained data with pre-defined, end user |
| development time, makes ongoing maintenance | | | | maintained, business rules. |
| simpler and provides users with information about the | | | | Real Time: As the source data changes it is instantly |
| source of the data, increasing their trust and | | | | passed through to the user. Often via message |
| understanding of it. | | | | queues. |
| Metadata: Data about data, describing how and | | | | Near Real Time: Source data changes are batched up |
| where it is being used, where it came from and what | | | | and sent through on a short time period, say every |
| changes have been made to it. | | | | few minutes - this requires special ETL techniques. |
| Commercial Justifications | | | | Batch Processing: Source Data is captured in bulk, say |
| There is clear commercial justification to improve the | | | | overnight, whilst the BI system is offline. |
| quality of information used for decision making. A | | | | Relational Database Vs OLAP (cubes, slice & |
| survey conducted by IDC found that the mean | | | | dice, pivot) |
| payback of BI implementation was 1.6 years and that | | | | This is a complex argument, but put simply most |
| 54% of businesses had a 5 year ROI of >101% | | | | things performed in an OLAP cube can be achieved in |
| and 20% had ROI > 1000%. | | | | the relational world but may be slower both to |
| ROI on BI > 1000% from 20% of organisations | | | | execute and develop. As a rule of thumb, if you |
| There are now also regulatory requirements to be | | | | already work in a relational database environment, |
| considered. Sarbanes-Oxley requires that US listed | | | | OLAP should only be necessary where analysis |
| companies disclose and monitor key risks and | | | | performance is an issue or you require specialist |
| relevant performance indicators - both financial and | | | | functionality, such as budgeting, forecasting or 'what |
| non financial in their annual reports. A robust reporting | | | | if' modelling. The leading BI tools seamlessly provide |
| infrastructure is essential for achieving this. | | | | access to data in either relational or OLAP form, |
| SarbOx requires disclosure of financial & | | | | making this primarily a technology decision rather than |
| non-financial KPIs | | | | a business one. |
| Poor data quality is a common barrier to accurate | | | | Top Down or Bottom Up Approach? |
| reporting and informed decision making. A good data | | | | The top down approach focuses on strategic goals |
| quality strategy, encompassing non system issues | | | | and the business processes and organisational |
| such as user training and procedures can have a large | | | | structure to support them. This may produce the |
| impact. Consolidating data into a DWH can help | | | | ideal company processes but existing systems are |
| ensure consistency and correct poor data, but it also | | | | unlikely to support them or provide the data |
| provides an accurate measure of data quality allowing | | | | necessary to measure them. This can lead to a |
| it to be managed more pro-actively. | | | | strategy that is never adopted because there is no |
| Data Quality is vital and a formal data quality strategy | | | | physical delivery and strategic goals cannot be |
| is essential to continually manage and improve it. | | | | measured. |
| Recent research (PMP Research) asked a broad | | | | The bottom up approach takes the existing systems |
| cross section of organisations their opinion of their | | | | and data and presents it to the business for them to |
| data quality before and after a DWH implementation. | | | | measure & analyse. This may not produce the |
| - "Don't know" responses decreased from 17% to | | | | best strategic information due to the limited data |
| 7% | | | | available and data quality. |
| - "Bad" or "Very Bad" decreased from 40% to 9% | | | | We recommend a compromise of both approaches: |
| - Satisfactory (or better) increased from 43% to | | | | Build the pragmatic bottom up solution as a means to |
| 84% | | | | get accurate measures of the business and a better |
| DWH implementations improve Data Quality. | | | | understanding of current processes, whilst performing |
| Tools Market Overview | | | | a top down analysis to understand what the business |
| At present BI is seen as a significant IT growth area | | | | needs strategically. The gap analysis of what can be |
| and as such everyone is trying to get onto the BI | | | | achieved today and what is desired strategically will |
| bandwagon: | | | | then provide the future direction for the solution and |
| ERP tools have BI solutions e.g SAP BW, Oracle Apps | | | | if the solution has been designed with change in mind, |
| CRM tools are doing it: Siebel Analytics, | | | | this should be relatively straight forward, building upon |
| ETL vendors are adding BI capabilities: Informatica | | | | the system foundations already in place. |
| BI vendors are adding ETL tools: Business Objects | | | | Advanced Business Intelligence |
| (BO) Data Integrator (DI), Cognos Decision Stream | | | | The following describes some advanced BI |
| Database vendors are extending their BI & ETL | | | | requirements that some organisations may want to |
| tools: | | | | consider: Delivering an integrated BPM solution which |
| Oracle: Oracle Warehouse Builder, EPM | | | | has business rules and workflow built in allowing the |
| Microsoft: SQL 2005, Integration Services, Reporting | | | | system to quickly guide the decision maker to the |
| Services, Analytical Services | | | | relevant information. |
| Improved Tools | | | | Collaboration and Guided Analysis to help manage the |
| Like all maturing markets, consolidation has taken | | | | action required as a result of the information |
| place whereby fewer suppliers now cover more | | | | obtained. |
| functionality. This is good for customers as more | | | | More user friendly Data Mining and Predictive |
| standardisation, better use of metadata and | | | | Analytics, where the system finds correlations |
| improved functionality is now easily available. BI tools | | | | between un-related data sets in order to find the |
| today can now satisfy the most demanding | | | | 'golden nugget' of information. |
| customer's requirements for information. | | | | More integration of BI information into the Front |
| Thinking and tools have moved on - we can now | | | | Office Systems e.g. a gold rated customer gets VIP |
| build rapid, business focussed solutions in small chunks | | | | treatment when they call in, data profiling to suggest |
| - allowing business to see data, store knowledge, | | | | this customer may churn, hence offer them an |
| learn capabilities of new tools and refine their | | | | incentive to stay. |
| requirements during the project! Gone are the days | | | | Increased usage of Real Time data. |
| of the massive data warehousing project, which was | | | | End to end Data Lineage automatically captured by |
| obsolete before it was completed. | | | | the tools. Better metadata management of the |
| A typical DWH project should provide usable results | | | | systems will mean that users can easily see where |
| within 3 - 6 Months. | | | | the data came from and what transformations it has |
| Advice & Best Practice | | | | undergone, improving the trust in the data & |
| Initial Phase | | | | reports. Systems will also be self documenting |
| Successful BI projects will never finish. It should | | | | providing users with more help information and |
| perpetually evolve to meet the changing needs of | | | | simplifying ongoing maintenance. |
| the business. So first 'wins' need to come quickly and | | | | Integrated, real time Data Quality Management as a |
| tools and techniques need to be flexible, quick to | | | | means to measure accuracy of operational process |
| develop and quick to deploy. | | | | performance. This would provide cross system |
| Experience is Essential | | | | validation, and verify business process performance |
| Often we have been brought in to correct failed | | | | by monitoring data accuracy, leading to better and |
| projects and it is frightening how many basic | | | | more dynamic process modelling, business process |
| mistakes are made through inexperience. A data | | | | re-engineering and hence efficiency gains. |
| warehouse is fundamentally different to your | | | | Packaged Analytical Applications like finance systems |
| operational systems and getting the initial design and | | | | in the 80's and packaged ERP (Enterprise |
| infrastructure correct is crucial to satisfying business | | | | Requirement Planning) in the 90's. Packaged BI may |
| demands. | | | | become the standard for this decade. Why build your |
| Keep Internal Control | | | | own data warehouse and suite of reports and |
| We believe that BI is too close to the business and | | | | dashboards from scratch when your business is |
| changes too fast to outsource. Expertise is required | | | | similar to many others? Buy packaged elements and |
| in the initial stages, to ensure that a solid | | | | use rapid deployment templates and tools to |
| infrastructure is in place (and use of the best tools | | | | configure them to meet your precise needs. This |
| and methods.) If sufficient experience is not available | | | | rapid deployment capability then supports you as |
| internally external resource can be useful in the initial | | | | your business evolves. |
| stages but this MUST include skills transfer to internal | | | | |
| resources. The DWH can then grow and evolve (with | | | | BI for the masses: As information becomes more |
| internal resourcing) to meet the changing needs of | | | | critical to manage operational efficiencies, more |
| the business. | | | | people need access to that information. Now the BI |
| Ensure Management and User Buy In | | | | tools can technically and cost effectively provide |
| It may sound obvious but internal knowledge and | | | | more people with access to information, BI for the |
| support is essential for the success of a DWH, yet | | | | masses is now reality and can provide significant |
| 'Reporting' is often given a low priority and can easily | | | | improvement to a business. The increased presence |
| be neglected unless it is supported at a senior | | | | of Microsoft in the BI space will also increase usage |
| business level. It is common to find that there is a | | | | of BI and make it more attractive. BusinessObjects' |
| limited knowledge of user requirements. It is also true | | | | acquisition of Crystal and recent release of XI will also |
| that requirements will change over time both in | | | | extend BI to more people, in and outside the |
| response to changing business needs and to the | | | | organisation - now everyone can be given secure |
| findings/outcomes of the DWH implementation and | | | | access to information! |
| use of new tools. | | | | Conclusion |
| Strong Project Management | | | | The potential benefits from a BI/DWH |
| The complex and iterative nature of a data | | | | implementation are huge but far too many companies |
| warehouse project requires strong project | | | | fail to realise these through: lack of experience, poor |
| management. The relatively un-quantifiable risk around | | | | design, poor selection and use of tools, poor |
| data quality needs managing along with changing user | | | | management of data quality, poor or no project |
| requirements. Plan for change and allow extra budget | | | | management, limited understanding of the importance |
| for the unexpected. Using rapid application | | | | of metadata, no realisation that if it is successful it |
| development techniques (RAD) mitigates some of | | | | will inevitably evolve and grow, limited awareness of |
| the risks by exposing them early in the project with | | | | the importance of training..... with all these areas to |
| the use of proto-types. | | | | consider using a specialist consultancy such as IT |
| Educating the End Users | | | | Performs makes considerable sense. |
| Do not under estimate the importance of training | | | | |