Business Intelligence as a Service – What makes the perfect fit
Cloud computing is transforming virtually every area of business software—and Business Intelligence, data analytics, dashboards and self-service reporting are no exceptions. Consequently, companies such as Birst, GoodData and Bime are delivering BI ‘front-ends’ on a software-as-a-service (SaaS) basis to great effect. But for many companies, a full Business Intelligence as a service (BI-as-a-Service) solution, rather than just a front-end, may make more sense.
BI-as-a-Service means getting the benefits of a full, end-to-end business intelligence solution “pattern”, but with the ease and simplicity of deployment associated with Cloud. And it’s important to understand that this is different from the likes of Birst and GoodData (i.e. providers or Cloud BI front-ends), who provide fantastic user interfaces, but where you still have to do the heavy lifting of your handling data, yourself.
In other words, it’s a bit like saying, “These Cloud BI tools make great front doors… all you need to do first is to find the plot of land, dig some foundations, lay the bricks, and build all of the house apart from the front door.”
Whilst many Cloud BI tools are simply “front-doors”, BI-as-a-Service solutions are the finished house (but still built really quickly).
What makes a great Business Intelligence solution
We think that there are three main functional ingredients to ensure that you get a great business intelligence or self-service reporting solution. These are:
- Your data is seamlessly extracted from multiple sources and pulled together into one place
- A well designed data warehouse, in order to organise, join together and host your data in a high-performance way
- A groovy front-end, that your users will like to use, and which will make your data look pretty
To validate this analysis, go and read the case studies section on almost any BI tool vendor’s web site (cloud or traditional) and many, if not all, will talk about how their technology is running over an existing Data Warehouse.
But in contrast, SaaS BI tools like Birst and GoodData excel at the last piece – the groovy front-end.
Alas, the real work (and hence cost and time), is in the first two pieces i.e. getting the data extracted and then organised into a ‘single version of the truth’ (or, as we would call it, a Data Warehouse).
A business intelligence as a service (or BI-as-a-Service) solution delivers all three areas: extracting data from your core systems, organising it into a high-performance data warehouse, and giving business users access to this warehouse through a natty front-end. It’s still fast to implement and easy to afford, like any other Cloud system, but crucially, it’s end-to-end.
So who does a BI-as-a-Service solution make sense for?
BI-as-a-Service solutions make sense in the following situations:
- You have business users who want to self-serve to access and create reports and dashboards
- You have bottlenecks around a small number of skilled employees who are creating and managing reports for others
- Management information is slow/hard to access/capitalise upon
- You have the data, but accessing it is sufficiently slow or painful that sometimes you don’t
- You need something to happen fairly quickly (weeks or months, not years)
- You have finite IT resource, or your IT resource is already stretched working on higher value projects
- You have a business-led requirement where IT support will be hard to get (as they’re perhaps busy elsewhere)
- You have a corporate business intelligence (BI) solution already, but you need a tactical solution for departmental/divisional/regional level which is faster or more agile to implement
In these situations a BI-as-a-Service solution will often make more sense, cost less and be faster to implement than either a traditional BI solution, or a Cloud BI front-end coupled to a build-yourself data warehouse.
If however you already have your data beautifully organised, and an IT team managing change in the Data Warehouse, than both traditional and Cloud BI tools/front-ends may work fantastically at bringing that data to life.