Other Links

Fabio Bastian explains Agile BI

13 March 2009

Highly experienced CIO & CTO Fabio Bastian explains the possibilities of Agile BI in a Software as a Service context

Businesses offering software-as-a-service (SaaS) applications are acquiring hundreds (or thousands) of new customers per week and generate an incredible amount of data from the resulting transactions. It is common place to see software platforms that service thousands of customers each week, with each interaction leaving a trail of data that tells something about the health of your business. 

Hidden data trends are incredibly valuable. They show when your sales model is working, when your customers are engaged, and when costs are under control.  It is knowledge that, in most cases, is automatically collected by your application and is available to all that want to look. Yet many businesses, particularly small businesses, fail to take advantage of application data. They rely on end-of-the-month finance reports and in the process miss out on critical insights.

The solution is to make trends visible so that the right people can understand and act on the information. But how do you make sense of the data? There is a perception that business intelligence is complex, costly to implement and therefore reserved to large corporations. Many projects start excitedly with the vision of timely insights and great decisions to come. By the time requirements are finalised, and all needs are known (I emphasize ALL) the organisation is facing a huge and complex project that is likely to cost a significant sum and take years to complete. The business case no longer works and many choose to delay or shelve the project and continue to rely on monthly tabular reports.

Enter agile business intelligence.

Agile techniques work wonderfully in the business intelligence world. By its very nature, business intelligence lends itself to incremental improvements as new insights are uncovered from existing data and newly discovered trends. Each discovery leads to new reporting needs, which can be implemented in future improvement cycles, typically every 2 or 3 weeks. For example, you discovered a jump in sales revenue last week and now would like to find out the reasons behind it? Simply add a break down of sales revenue by product into a future improvement cycle. Another example, you know that one of you channel partners is bringing far more businesses than others, but which one? Simply add a break down by partner in a future improvement cycle. In a short time a powerful business intelligence platform emerges and is producing value from day one.

Essentially, start small, show value early on,  improve constantly and incrementally. It is critical to show results soon, so start with simple metrics that are useful immediately.

Sure, but how do I start?  Businesses differ in their intelligence requirements. Some require complex product life-cycle metrics, others are interested in number of orders in the period. I recommend to start with sales and then build up from there. Sales trends are a must-have and are most powerful when visualized graphically.

To start, aim to generate a graph of sales trend across different time periods, say last 12 months, last 12 weeks and last 14 days. By using multiple time periods, you can quickly visualize long and short term trends and adjust your plan accordingly. In a recent assignment, we built a simple one page report to show sales trend for a subscription product against these time dimensions. Nothing fancy, just three graphs with subscription trends by month, week and last few days. The result was a great tool to compare forecast and actuals. Sales trends were clear and visible to all  and as a result many changes were made to marketing and sales plans.

Before you say it is not applicable to your business, its too difficult, your business is too complex, etc here is what you need you need to start:
1. data extract containing "sales date", "product id", and “sales amount” 
2. simple tool to plot the trend

And here is what it would look like:




Simple but incredibly powerful. This information alone can transform the understanding of your sales cycle and highlight deviations in the plan. 

That's it, your first iteration is complete. Use future cycles to enrich reports and track other relevant groupings and data breakdowns. Say, you would like to understand sales by product category - that is, which categories are selling better. Add the category code to the extract and enhance the report show it. Add additional groups later to represent other demographics like customer age bracket, customer address, etc. Before you realise you are well on track to business intelligence success.

In future, consider adding operational metrics associated with costs. For example, total number of customer support tickets, average call answering time, total number of product returns, average time to fulfillment, etc.  Operational trends are invaluable to optimise business operations and processes.

At the end of the day, there is an immense amount of business insights in application generated data. It is up to each business to capitalize on this asset. With agile business intelligence one can start simple, build knowledge incrementally and deliver tangible benefits immediately. 

Written by: Fabio Bastian

Candidate & Client Quotes

• Advisory + M&A