10 Reasons Why BI Projects Worry IT Managers

Rolling out a quality BI product through the enterprise is not a task that can be just assigned to a single silo group, it requires management and work groups to commit to a long term project, it requires cross group co-operation, data cleaning, data sharing and most importantly education. 

Resistance groups may pop up during the project when people feel the challenges out way the benefits to them or better reporting will threaten their own jobs; while others may believe that company data should not be openly available. 

To address these blockers and help get BI implementations off the ground more quickly we have compiled a list of the top ten most common cultural issues with BI project rollouts. We then asked the CEO of Yellowfin Business Intelligence, Glen Rabie, to detail some of his own frontline experiences with BI projects and provide some solutions to alleviate the anxiety.  

 
1. Complexity

BI projects bring with them a new set of problems, company data is generally is disparate systems that must now be linked, data is often duplicated, data archiving and cleaning become more important, there are political issues to solve, different hardware and opsys to consider, which appliance to use and then the massive problem of educating people. What is it about BI projects that make them so complex?

 

To be honest I do not actually think BI is that hard. The true complexity lies in a number of items – one, actually defining the reports and data that people want. Business analysis and really engaging the end users is critical. Two, data cleansing – many applications allow users to enter junk – and the BI tool by exposing data makes this data visible. Now you have to go back and clean all that up! Lastly, you have the big bang approach. To many companies feel they have only one shot at a BI project. In my view this is the wrong approach. Start small and develop an incremental framework for delivering BI over time. Let you users see what they are getting in stages and allow them to learn what they actually need. And yes – give people time to improve business processes to clean up that data.

 
 
2. Change

This single issue can stall an IT project at any point, people fear change or perceiving a threat to themselves, whether it results in more work for them personally, greater scrutiny or worst case make their jobs irrelevant/obsolete. Business Intelligence tools can result in a great deal of change – what are the cultural impacts of this change?   

  

Like any project there are always a bunch of people that just like doing it the old way. Their power is derived from being the gate keeper to data – take that away with a BI tool and you do change their working lives. The problem with the BI project is that the gate keepers are critical to success. If they block or attempt to de-rail the project then there is little chance of delivering a great outcome. 

There has been a tonne of articles written on the blocker so I will not delve to deeply. However, another more subtle failure occurs when key users do not get involved in a project due to other higher priority tasks. For BI to be successful a series of sign-off steps need to occur to ensure data accuracy and that what will be delivered will be of business value. If the key internal users do not get involved and are just waiting for the delivery of the BI project – the project will fail. And as usual the project team is held accountable. Drive the change management and really make sure that all stakeholders are involved and sign-off on every step within the project.

 
 
3. Level of Investment

Like any IT project, BI projects require investment in time, money and human effort. The investment can seem too steep for the payoff for some.   Why do BI projects suffer from a perceived lack of ROI?

 

Well this is the recurring theme of BI. Business users want it, and yes it does cost money and yet putting your finger on the value of information and how that translates into accountable business benefits is incredibly difficult. The way we like to think about it is simple – imagine working without internet access – in the early days no conceivable benefit but now who could imagine a knowledge workers life without it. Sure you can continue to use excel and cobbled together processes and save money by not going out and doing that BI project – but in the long term will your business be better off with out that investment? Probably not!

 
 
4. Just another IT Fad

Like many of the must have technologies that have propagated through international corporations, from CASE tools in the 80’s to WEB2.0 today. IT managers fear Business Intelligence for the masses will have limited usefulness and after the initial interest period will gather dust as people get on with their established ways of working. What is you view on this? 

 

I think the opposite will occur. If we look at the history of business computing the persistent trend is to making users self sufficient. Lets take Excel as an example. In the early days there were a handful of excel super heroes in an organisation – now it is ubiquitous – everyone pretty much knows how to navigate a spreadsheet. What is lacking from the spreadsheet, however, is the centralised access to multiple data sources – and that is BI. People now are starting to want direct access to their data without IT’s intervention. And that is a trend that will not stop.

 

5. What’s wrong with Excel ?

Business and IT groups have become used to using the tools they have at hand. Traditionally Excel is the application that people are quite comfortable for most of their data manipulation and reporting. Do IT managers believe it is sufficient for most staff?

  
Look it depends – I do not want to speak for all IT managers but I suppose a large set would feel this way. There are a myriad of problems about using excel – most of them not visible to management – and so in that sense not a problem. 

Some of my favourites – How did users get their data that they are manipulating in the first place? How much time do they spend putting it together and re-running each month? It’s a great way to blow away a few days every month. How accurate are their assumptions and calculations – and on it goes. Sure excel is a great tool but it does not deliver consistent, auditable and replicatible BI – end of story!

 
 
6. Measuring ROI

CIO’s and business groups want to implement BI across the organization to drive new efficiencies, but once that decision is made. The next phase is to establish how to prove those efficiencies are occurring and that people are embracing the technologies benefits.

  
As I mentioned earlier – measuring BI is incredibly hard to do. What’s the value of email? Why bother with PCs. Simply put – BI tools are enablers – they will become critical business infrastructure. 
 
 

7. Executive expectations

Often a key driver for a BI project is the executive team’s desire for a sexy business dashboard. Do many IT managers fear that the BI rollout will not meet the lofty goals executive have envisioned.
 
The needs of the executive team differ vastly from those of operational users. In my view it is best to deliver operational reporting first and then the executive dashboards – problem for the guy paying the bills I know! However, the reason for this is that if you deliver high level dashboards to the exec team if the data is trending the wrong way they will want to drill down for answers from their staff. If their staff do not have access to the same data at a lower level – how will they respond? Well basically they will query the data – the numbers can’t be right – that tool is wrong! So it is a fine line – my approach is to deliver less data initially but deliver to all segments of the business. So that the exec get sales data but so do the customer care people etc. That way everyone has access to the same data all be it from different views.
 
 

8. My Company is to Small for BI

Some organisations bypass Business Intelligence, in the belief that their smaller size or lower staff count will not benefit from business intelligence. They assume that BI is only relevant for government or fortune 500 companies. What is you take on this?

 
Well BI is for everyone. Running a large business or a small business is the same. If you do not have total visibility of all your key metrics then you may miss those critical opportunities or risks for your business. Sure you budgets to implementing BI are going to differ – but so then will the complexity of your business. For example a small business is not going to have 100 different data sources, which means getting something up and running will be a much lower cost – and the benefits to the business almost instantaneous.
 
 
9. Data Security and Management
Business groups may desire greater freedom of access to company data.  However, IT just see’s the problems. Suddenly with the data aggregation, comes the issues of data cleaning, data partitioning, extraction from disparate stores and keeping the sensitive data access controlled.
 

Any BI project will come across security issues. If you have a robust BI platform in place then security should not be a headache. My number one advice in this area is to make sure that whatever tool you use it must have row level security infrastructure. You have to be able to write a single report and share that with many users in the confidence that they will only see the data that is relevant to them.

 
 

10. Data Confidence

IT having rolled out the enterprise wide BI project, now discovers that it must manage the ongoing problem of data trust, people are making decisions that affect the business daily and those decisions are being driven/supported by the business intelligence tools. How will IT guarantee that the reports and dashboard elements can be trusted.
 
 

And now for the bombshell – data quality really gets to way much airplay. It’s good for buy-in but really you do not need 100% accurate data. That is not the role of analytical BI. Analytical or non-operational BI is about trends and overviews, not transactional accuracy. If your sales trends are heading South – but a couple of records are inaccurate by 5 or 10% – what should the exec worry about? Care about the trend of your business not the quality outliers!

However, we all know it is an issue, and I did touch on this earlier – there are a multitude of applications out in the business world that allow users to enter junk data. The options are: one, to educate users on data entry. Make accurate data entry a KPI so that data is cleaned over time. Use your BI tool to highlight poor quality data and fix it at the source. Two, for those with a bigger budget – change your applications so that more data validation takes place at the front end. Again this is about fixing it at the source – do not try and clean data continually at the BI layer. Cause it will just never cater for all issues.