Business intelligence is about providing information to the powers that be, allowing them to make informed decisions in a timely manner. There are many different elements to business intelligence and there are even more approaches you can take. In this article we’ll dive into a few of those and we’ll look at how you can improve your business intelligence maturity so that you can make even better decisions.
Firstly, business Intelligence is a term that is easily confused. Some people use the terms business intelligence and business analytics interchangeably. Some regard BI as one element of a wider range of data driven methodologies. Some don’t even call it BI, opting for management information instead. For the sake of simplicity, I’m going to go with BI as an umbrella term for all of that.
As I’ve already mentioned BI is a broad area. Waterstons have people focused on BI, but that can mean anything from producing some operational reporting from your stock system, to developing a platform for ‘what if’ analysis of sales and margin figures. BI can cover many things, but most commonly we find that requirements fit into one of these four areas:
- Operational Reporting – typically direct reporting from your line of business system.
- Data Warehousing, dashboards and classic BI reporting systems – more advanced reporting, combining multiple data sources and giving you a structured and clear way to retrieve information.
- Business Analytics – statistical analysis of the information provided to you, from your data warehouse, line of business systems, or both, allowing you to look for insights not handed to you on a plate report.
- Planning and Predicting – taking the statistical analysis one step further and attempting to predict how your business might perform in the future.
But how do you do all of that good stuff? Where do you start? And how do you mature your BI offering so that you can predict how many ice creams you’re going to need to stock in your store in Tadcaster in July? An important question, I’m sure we can all agree.
Rubbish in, rubbish out
You need to start with the data from your line of business systems; the stores of data from which your BI systems will be built. You need to ensure the data in your systems is accurate and that you can get access to it in a timely fashion. If your data is inaccurate and outdated then your reports are going to be nonsense. It’s as simple as rubbish in – rubbish out and you’re going to need to tackle any data inaccuracies before you even think about making judgements on them. You can use reporting systems to highlight these issues, but you’re going to need to work to ensure the data is accurate.
With your information accurate and available, you can start to look at your operational reporting. This is where you can find out what is happening, or what has happened, usually on a daily or weekly basis. How many ice creams did we sell yesterday? How much money have we made this week? This kind of thing is usually pretty straight forward but is obviously extremely useful, if not critical, in the day-to-day running of your department or business.
With operational reporting under your belt, you now want to get more out of your data. You want to start producing views of your data that are a lot more meaningful and give key decision makers a snapshot of the business as a whole. You might want to look at a monthly or weekly comparison. How much did I make this month and how does that compare to last month? What is the average transaction value of sales in my stores? Who are the customers that spend the most and what products do I need to stock more of to meet demand? Typically, these questions are defined by your KPIs (Key Performance Indicators).
To answer these questions, you need to start thinking about stand-alone reporting systems. You don’t want to rely on the data coming out of a single system, you want to look at the bigger picture and to do this you need data from more than one system.
Enter the data warehouse…
Traditionally, a data warehouse was the answer here and in many scenarios it really still is. A data warehouse is a central repository for all of your data, structured in such a way that getting the information out that you need is a doddle. This isn’t a simple undertaking and can require extensive development of ETL (extract, transform and load) procedures, database schemas, OLAP (Online Analytical Processing) cubes and then all the various reports and dashboards that your users are after.
Data warehouses usually fit the bill and provide flexibility, scalability and reliability but they do require considerable development expertise and they’re not the only option.
…or one of the newer options
You could instead use one of the many tools that allow you to do the extensive ETL processes and schema designing within a few clicks and the odd drag and drop. You can then start surfacing data and presenting dashboards with speed and ease. You might not have the control or structure you would with a data warehouse (and you’ll likely still need some form of developer input to create all the relationships and calculations you need) but it is certainly a quicker way of getting that more intelligent reporting system.
Whichever route you take, your business intelligence is becoming more mature and you are able to answer questions immediately that would ordinarily take hours to answer with multiple spreadsheets, a few conversations and best estimates.
Next up are those questions that are much trickier to answer – because you don’t know what they are yet! You might have a hunch about something or you might have spotted a potential new sales trend, but you don’t have the details yet, so you want to explore the data and find out more. Again, being able to spot these trends and understand them quickly is key – if we wait for the monthly sales meeting it could be two or three months before any real information can be provided. How long does it then take you to resolve these issues? The impact of delays like this could be catastrophic – you could end up behind the curve, trying to catch up with your competitors or, even worse, chasing trends that aren’t actually there.
At this point you need to start taking an active role in your BI system. You need people across your business to understand the data and to be able to look into these areas of interest at the drop of a hat. You need the tools to enable you to do this. You’ve got your data warehouse and your source systems producing reports, you might have a self-service tool that you’re using to produce some shiny graphs for the board meetings, but are you really drilling down into the detail of the issues and establishing why things are happening or are you just passively viewing the reports once they’re produced?
You might need access to further data, new data sources to really resolve the issues. Maybe you need to know what the weather was like to answer why no ice creams were sold last Tuesday. Maybe you need to know your stock levels. Maybe it was because there were road works outside and no-one could actually get to the store. Being able to explore these theories, to add new data sources and to look at what information they provide you with, are key to this level of intelligence. Traditional data warehousing can assist with this process, but you typically need a self-service, data discovery type tool for the in-depth analysis you’ll require.
Predicting the future
Once you’ve got your business analytics in place and you can undertake the kind of analysis required, you can then start looking at using the information you discover to predict what will happen. You’ve established what your average daily sales are for ice creams. You know how much, on average, your sales increase on weekends and likewise on sunny, warm days. You can then use that information to predict what your likely sales are going to be in your store in Tadcaster.
Building on that, however, you can also try to plan your stock levels for the store to ensure you meet the expected need. Introducing a Corporate Performance Management tool into your environment will allow you to plan on a large scale and introduce advanced approaches such as What-If analysis. If we reduce the cost of a strawberry ice cream and push it to customers, how much will this increase sales numbers? How much will this affect the revenue? Will the increase in sales of strawberry ice cream affect the sales of blueberry and chocolate and the other regularly priced ice creams?
These kinds of predictions can be used to drive production or stock decisions, but can also then feed into marketing campaigns and customer engagement. We’re dropping the price ten percent, so start telling our customers and driving them to the store. This approach is at the very mature end of business intelligence and requires the most active engagement, but the increase in efficiencies throughout your business can be massive.
Whatever the maturity of your business intelligence, it’s important to continually improve as your company grows. You might well have a data warehouse and some great dashboards right now, but what happens when people start wanting to ask new questions? If people become engaged with your BI offering, no matter how big or small it is, they will want more and more: new measures, new dimensions, different visualisations, more interactivity and generally more everything. Likewise, your business demands may well require you to start carrying out some advanced analytics and if you don’t have the people, the know-how or the technology to keep answering new, more complex questions then your business could be left behind.