The truth about BI tools - spreadsheets, consolidation and the cloud
Written by Ken Miller, Product Director at Panintelligence
OK, so I have 3 truths about Business Intelligence (BI) to share with you. Many of them I'm sure you'll be able to identify with.
#1: BI isn’t solved, so long live Excel
The world uses Microsoft Excel. Many of the reports which run in BI tools like ours are used to extract data into Excel. There I’ve said it! So what’s wrong with this? People can achieve a great deal very quickly with the data: they can create formulas, correct data and format it into any shape they require.
So why isn’t BI solved?
We all (surely) have access to Excel or a similar product – and yet I don’t think in my 20+ years of working as a BI consultant I have ever been to a company that considers BI completely finished – or even very good!
Why is this still the case in 2017?
It’s not a problem with Excel, it’s the best software I’ve ever seen. It’s 30 years strong!
It’s how it’s used. People create what they understand and rarely take the time to understand what others requirements are. People create their own version of the data, each with subtly different calculations and metrics. Data gets corrected in spreadsheets, not in the systems themselves. This leads to an ever-expanding list of changes and patches that must be applied to correct the data. The end result is, that spreadsheet you created takes longer and longer to update.
Soon updating your spreadsheet becomes a full-time job!
And you feel like this guy...
Instead of having one version of the truth in a central BI dashboard, spreadsheets become the source of other spreadsheets, the path of data becomes convoluted and hard to trace and errors become difficult to resolve. The data is hard to trust when one variable is quoted from multiple spreadsheets with different values.
The owner of a complex spreadsheet leaves the business, no-one understands how it works, and it is refreshed in the hope that nothing will go wrong (year-end arrives and it fails!).
There are many, many more examples…
#2: BI consolidates but can't fill all the gaps
BI is hard, and it’s because it is used to fill gaps. These gaps can even require creating data not captured or stored by the source application. No software product will be able to operate in exactly the way every business operates. It can never match the processes entirely and provide the data in the exact format required to provide the business with all the information required.
In addition, a single BI product is not used to fill the gaps in a single software solution. It is often used to bring data together into a single view from multiple applications. It’s used as the integration solution. Actually, this is when BI is at its best; it consolidates information from multiple business systems in one place.
#3: BI doesn't play well with the Cloud
OK, this is a tough one to swallow, but it’s getting harder. The a reason for this is the Cloud. Collecting data from multiple sources used to be about connecting to multiple databases and files held on the local network. This did have challenges, but now data can be in multiple cloud-based applications which provide limited and bespoke access via an API.
And whilst API’s are fantastic at providing transactional access (fetch, alter and update a record), they are very poor at providing business intelligence.
I feel better already getting that off my chest!
So, what’s the solution?
One solution is to extract the data into a database which itself can be on-site or in the cloud. This feels like Plan B, because the whole point of cloud SaaS-based offerings is to simplify IT!
Another solution, which feels like Plan A, is to use a vendor's own standard reporting. This means juggling multiple data sources and settling for the vendor's view of your data. Plan A could also entail lots of manual work to copy data to Excel, which compromises the security of data.
Does this feel like an accident waiting to happen to anyone else?
Share a link to this article and comment.