The 11 Secrets to Ensuring Your BI Implementation Delivers
Embarking on a self-service reporting or BI implementation for your company probably makes sense.
After all, well implemented Business Intelligence solutions can deliver huge returns. Delivering the insight you need to find new sales, rationalise inventory, streamline performance, boost cash. All the while saving lots of time creating manual reports.
But how can you ensure you’re successful? Like many IT projects, a BI implementation can be tricky. Over 1/3rd of all IT project overrun by at least 100%. Approx 3/4 of Business Intelligence projects fail. Scary stuff.
Use these great trade secrets to ensure your BI implementation delivers, on-time and on-budget. Whether it’s sales analysis, self-service reporting or data visualization that’s driving your requirements, these techniques will help you deliver BI that benefits your business, fast.
- Build a business case – before you can get started on your BI implementation, you need to justify it to yourself, your colleagues and your budget holders. Learn how business intelligence software pays its way in this article then build a business case for your organisation
- Identify your requirements – develop a list of requirements for your business intelligence implementation. Split this into two parts: what data you need to analyse (e.g. revenue, margin, debtor days, KPIs); and what features you need from the front-end BI/reporting tool (Tip: the first part of the is more important!)
- Select the right technology – Evaluate the types of BI solution available (e.g. traditional on-premise; cloud business intelligence; hand-cranked solutions) and shortlist those most suitable for your needs based on your requirements, budget, timescales and attitude to risk
- Define your Facts and Dimensions – Define the Facts (business events), Dimensions (business entities) and Measures (numbers and values) you need to satisfy your reporting and analysis requirements.
- Use bite size chunks – based on the your BI requirements as defined above, organise your project into bite-size chunks e.g. Sales Analysis, then Inventory Analysis. This way you can deliver function-by-function, or department-by-department, gaining value from your BI implementation in each area then moving on.
- Prioritise features – business intelligence companies are great at telling you the latest, greatest feature is essential. The reality is that most BI users use basic self-service reporting more than any other whizzy feature. Decide what’s important to you and select accordingly. Don’t be sold on the bells and whistles.
- Find and evaluate your data – perform an audit of where the data that supports your reporting requirements is held. Will the data format, structure or shape need changing to make it useful for reporting? Will calculations or manipulations need to be applied to it to get useful information? Is it in more than one place?
- Use a Data Warehouse – if the answer to any of the questions in ‘Find and evaluate your data’ was “yes”, then you need a Data Warehouse and an ETL layer. A Data Warehouse allows you to get all your data into one place (one version of the truth) and apply business rules to it (e.g. currency calculations, KPIs, rebate/cost/discount rules, data cleansing). A DWH also makes it easier for the user to report, as the data can be arranged into a format optimised for reporting (so called, “de-normalisation”). ETL (Extract, Transform and Load) takes the data from the source systems, transforms is and loads it into the Data Warehouse
- and a meta-data layer – a meta data layer abstracts the business user from nasty, technical field and table names. So rather than adding ‘CNM001’ from table ‘CSMST05’ and joining it to another 5 tables using complicated SQL, the user can simply drag-and-drop ‘Customer Name’ and ‘Net Value’ onto their report. A meta-data layer also allows you to pre-define relationships in the data and enforce security rules.
- Tackle “GIGO” – garbage in, garbage out. This is one of the number one concerns people have in BI implementations. As part of your data assessment, you will already know where you data is and in what sort of shape. Start ti improve it now. You can read more on data qualitity and GIGO in this blog post
- Develop key-note reports – make sure you deliver one or two key-note reports early. Reports that look great and deliver value. This will gain momentum for your BI implementation
Using these tips and techniques, your BI implementation will rock – helping you deliver the stronger, smarter business you want to be.
Download our free e-book, ‘The Complete Guide To Evaluating and Implementing Business Intelligence and Self-Service Reporting’ to learn in detail about each of these topics and more.