BI & Visualisation

Intelligence, not Information – Presenting Your Message

Business Intelligence Confusion vs Information

Organisations base most of their tactical and strategic decisions on ‘the numbers’: it seems like a trivial thing, but the way you present your final numbers can have a huge impact on whether people digest what they say, and hence make the right decisions.

Numbers are often produced by accountants but also sales & marketing and supply chain planners amongst many other operational users, often in big spreadsheets with hundreds of rows and multiple tabs. They create detail and order, and ensure that every little bean is accounted for. This helps the Board to sleep well at night, knowing that their Finance & Operational teams have all the numbers under control.

Unfortunately it doesn’t help senior management or the Board in actually understanding what’s going on in their business. Senior managers have short attention spans, and just don’t have the patience to sift through all of these spreadsheets, looking for problems hidden deep within the numbers.

There are a number of solutions on the market and ultimately the right solution will be based on your individual use case requirement and potentially existing architecture. We have designed and built solutions with IBM Cognos BI, Tableau and Motio amongst others but ultimately it is all about the design and telling the story.

Which Numbers? Choosing your KPIs

So we need to cut down on the numbers. But which ones do we keep? And, more importantly, how do we convince our more detail-hungry stakeholders to live without the rest?

It’s all about ownership, buy-in and creating consensus:

Step 1
  • Group your stakeholders into subject area communities (eg Board, HR, Operations, Employees). You’re never going to get eveyone to agree on everything, but by creating reasonably similar-minded groups of people, you’ve got a much better chance of lots of smaller successes.
Step 2
  • Pick the most enthusiastic bunch – these guys are going to be your guinea pigs and your evangelists, spreading the message about the new reporting framework, and making the rest want it.
Step 3
  • Run a series of workshops with your pilot group to come up with 3-6 KPIs for them. You may want to start by brainstorming as many indicators as they can think of, and then prioritising them. During this process, you’ll find ‘master’ indicators, that relate to quite a few other ones. Try to get an idea of:
  • How they’re calculated (agreeing definitions and a ‘business glossary’ is important)
  • Where the data comes from (and how reliable it is)
  • How often they’re measured (eg daily, weekly, monthly)
  • What underlying detail (other indicators, or reports) can also be provided)
Step 4
  • Use statistical analysis to back up your hypothesis. Most of the time, the ‘gut-feel’ of business managers (the workshopping approach) is the best way to get your KPI list, but sometimes the statistics can surprise you – especially with complex businesses. If nothing else, it removes the subjectivity and achieves consensus faster.
Step 5
  • Prototype your new reports and dashboards!

Sourcing your Data

Before you open your new reporting tool and start designing dashboards, take a step back and think about where your data is coming from. The more ‘structured’ your data is, the easier it will be for you. A dashboard built on unstructured and inconsistent data, no matter how slick it looks, is like building a castle in the air – sooner or later it will come crashing down, and all your work will have been wasted.

Unstructured Data Examples Data Source Issues
  • Spreadsheets
  • Extract files
  • PDFs
  • Web feeds
  • Social Media & other ‘Big Data’
  • Constantly changing formats (eg new rows, name changes) will break your reports
  • Data quality – how reliable is the ultimate source of this data?
  • Consistency of terminology – how can I drill down from the salaries & wages GL account to the employee detail from Payroll, when cost centres have different names?

Of course, unstructured data isn’t all bad – for ‘real-time’ operational reporting you’ll have to use it raw and unadulterated.

Ideally, what you really want is a dimensionally-modelled data mart, with nice clean reliable data from multiple sources. See what your IT department already has, or have a look here for some tips to tackle it yourself.

Charts, Tables or Text?

Business Intelligence Charts

Different people process visual information in different ways. Some prefer charts and graphs – some even like pie charts – and some prefer tables. Others prefer to be told the story in narrative form. The trick is to know your audience, and be prepared to change the style (or present it in several different ways) based on their feedback.

There are a few basic rules though:

  • Keep it minimal – don’t clutter the screen
  • Don’t overdo the colours – use your corporate palette if you have one
  • Enable drill-down to more detailed reports – to keep the ‘detail guys’ happy

What Next?

At the end of the day, every organisation is different and you need to find your own approach that suits the personalities involved. It would also be a good idea to come and talk to the Business Intelligence team at Cornerstone – we’ve done this hundreds of times across almost every industry sector.

If you’ve got big ideas for your reporting framework, and want someone to bounce them off, contact Hamish Dwight today.

Case Study – Coates Hire Operations

“We wanted a simpler way to deliver information, one that wouldn’t tie our guys up in spreadsheets,” says James Welch, Chief Financial Officer at Coates Hire. “We realised that a summary-level dashboard could give our brand managers an instant overview of what the top priorities should be, and we also wanted a capability to drill down, right to the source.”

Coates Hire

Case Study – Coates Hire Operations