Show Me the Numbers

As part of my reading for my Personal MBA, I am documenting the key lessons I learnt from Show Me the Numbers. It is not intended to be a summary or review, rather a reflection of how the book has influenced my numbers presentational style.

In the world of Finance, we are told “the numbers speak for themselves.”

Numbers don’t speak for themselves.

The right numbers displayed in the right way speak for themselves. That is the goal.

I find that it is something all to common in Professional Qualifications. Key fundamental and complementary skills are taken as assumed knowledge or irrelevant, and completely ignored. So many courses and exams to become a Chartered Accountant, but there is nothing about presenting the message of the numbers produced. Those numbers are the deliverable, the output, the whole point of it all.

This is similar in other industries, for example perhaps 90% of Dentists, Osteopaths, Plumbers, Electricians, end up operating their own businesses, yet there is typically not one module on how to do this effectively. A more balanced, multi-disciplinary real world approach is needed.

Show Me the Numbers

This was not the easiest read, that is not to say it was tough, just that I am still on a basic level and this went beyond that. I would have preferred a more finance focussed approach. Our needs are simple and relatively consistent. That said, it was a great read or perhaps more appropriately review, as it is the “pictures” that hold much of the value.

In my experience we don’t spend time considering how we share the numbers produced, at least not as much as we should. I wanted to be better, I know how crucial getting the message across is. This book is one I would never have read if it weren’t in my Personal MBA prospectus. As I am targeting a more holistic approach to close the gaps in my education, including many of the key neglected fundamentals. It was not a read that excited me. Which meant it was more important. A key area, a fundamental, important, neglected.

Before this book, I had decided that I prefer numbers to graphics. I’m probably not in the majority, but as I read it, I realised it wasn’t a preference, but rather because the examples I’ve seen are generally not very good.

We are typically bad at producing meaningful graphs. They are relatively new and Excel has changed the need for careful deliberation without the historic manually intensive process needed to create them. As we are not taught how to present numbers, we add a pie chart, we add 3D, it looks impressive, but it doesn’t tell anyone anything, other than you know how to create a 3D pie chart. These are distracting from the message.

A few people may have an innate ability to effectively use graphs, the rest of us are left floundering. We have not seen enough good examples. Even slowly glancing through the book, reading the commentary and summaries will improve your exposure to possibilities. It will make you better.

What is the Message

Numbers should be used to make a decision or report on the impact of a decision. We should share what is interesting. The numbers themselves should be interesting, not the way they are displayed. We should not be tempted to “jazz them up”. We do not want to distract from the message. If they are not interesting, then they are the wrong numbers.

We need to realise there is an eloquence in simplicity.

We use numbers to display KPIs, balanced scorecards, and performance dashboards. They should all tell a clear story.

Tables or Graphs?

Tables for individual numbers

Tables make it easy to look up individual values. They also combine different units of measure with greater ease. They summarise data well and are better for smaller data sets (e.g. <20). Where the users are looking at one data point relative to another at a time.

For the table design there is a need to balance lines, fill colour and white space. The goal is to make it easy to read and digest. It is best to keep the orientation and alignment simple and consistent. It is worth remembering that precision does not always add to the message.

Graphs for the space between

Graphs can be used to tell complex stories. Graphs are a visual display of quantitative information and they allow an interpretive view of relationships. Graphs show the overall shape to the data, like trends and patterns, basically the relationship across the data. They are better for larger data sets.

Explained a different way, graphs are for the space between data points, tables for the data points themselves.

How and when to use Graphs

This largely depends on graph types. The types of graphs commonly found are points, lines, bars, boxes and shapes (areas/intensity). They all take practice to read and the recipient should be considered when deciding. For example shapes and boxes should be left to those with greater experience and a specific need.

Pie charts were excluded from the above and this is as Stephen Few is very clear on his opinion of pie charts – “Speaking of “difficult to read”, allow me to make it clear…that I don’t use pie charts, and I strongly recommend that you abandon them as well. My reason is simple: pie charts communicate information poorly.”

Stephen goes on to explain that the donut chart is even less effective than a pie chart and 3D donuts are particularly useless. Similarly funnels rarely work, as the intervals are not consistent enough. Waterfalls have specific uses, very specific and in my view radar and staked graphs should be reserved for the experts. I don’t understand them, so it would take a stroke of luck for me to effectively use them to convey the message.

As a quick reference basically directly from a summary in the book –

NominalDots when you can’t use bars because it doesn’t start at zeroAvoidHorizontal or verticalAvoid
Time seriesDots only when values were not collected at consistent intervalsEmphasis on overall patternEmphasis on individual values.When showing distributions as they change in time.
RankingDots when you can’t use bars because it doesn’t start at zeroAvoidHorizontal or verticalWhen ranking multiple distributions
Part to wholeAvoidHow parts of whole have changed in timeHorizontal or verticalAvoid
 DeviationWhen does not begin at zeroWhen combined with time seriesHorizontal or vertical. Always vertical when with times seriesAvoid
DistributionEmphasis on individual valuesLimit to a few linesEmphasis on individual intervals (avoid for multiple)Avoid (box for multiple)
CorrelationScatter plotAvoidHorizontal or vertical.Avoid
GeospatialVary point sizesMark routesAvoidAvoid

It would appear, if you are uncertain, go with a bar graph.

Don’t distract from the message

Graphs can be used to completely distort the message, depending on the scale used. It is key to make sure the message is accurate and not designed to hide the reality (although perhaps useful from time to time, is likely only delaying the inevitable action needed).

Visual perception is key to an clear message whether it is a graph or a table:

  • Avoid fill patterns and line style. Rather use colour and position.
  • Misalignment is very distracting.
  • Limit contrast to the message. Ensure it is not distracting. Overly dramatic colours can also distract.
  • Proximity, similarity, enclosure, continuity and connection help show the relationship or groupings.
  • Organize by group, priority or sequence

The Takeaway

First understand the story, then the message and then take time to consider how to convey that message. Always asking how will the information be used. Sometimes we hack numbers and how we display them until they look interesting, perhaps that means we are using the wrong numbers.

Learning requires practice, but to get started avoid 3D graphs, pie charts and anything that requires too much interpretation by the viewer. Remember the message and if unsure go with a table or a bar graph.

Facebook Comments

Leave a Reply