Bullet Graphs or Smartlines – Which One is Best?

August 19, 2014 by Stacey Barr

I’ve written about two very powerful graphs to use on performance dashboards, or in performance reports that need to be concise: my version of Stephen Few’s bullet graph, and my version of Edward Tufte’s sparkline (my version is called a smartline). Is one better than the other?

It’s not a matter of deciding which one to use, but rather why you’d use both.

Smartlines tell you the time, type and direction of a real change in performance.

Smartcharts can do this because they use the two powerful features of XmR charts. The first powerful feature of the XmR chart is the central line, and this line is roughly the equivalent of the overall current level of performance.

The second powerful feature of the XmR chart is the natural process limits, which tell you how much routine or natural variability your measure has, and therefore which period-to-period fluctuations you can ignore because they’re just routine variation.

With smartlines, you can apply a set of statistically valid rules to detect when performance has changed. So these little charts pack a lot of interpretation power into a small space. At a glance you can see if performance is changing, in what direction it’s changing, and when that change was triggered.

Bullet graphs give you a comparable snapshot of how far current performance is from to-be performance.

The smartline won’t tell you how far current performance is from to-be performance, because it wants your attention on what’s happening through time, rather than on the magnitude of difference between current performance and your targets. So the scaling of the vertical axis in a smartline is not based at zero, it’s designed so the chart data fills the available chart space.

At a glance, your bullet graph will tell you how far current performance is from where it should be or where you want it to be.

So when you use both bullet graphs and smartlines together, you get a very comprehensive answer to one of the most important questions you should ask of every performance measure: What is performance actually doing?

You can, at a glance, know:

  • is performance changing, or staying the same?
  • if it’s changing, is it changing for the better or the worse?
  • if it’s changing, when did the change begin?
  • is current performance improving at the rate it should to reach our targets?

The performance measure Purchasing $ on Approved Contracts measures the percentage of a company’s total expenses that are made via contracts with approved suppliers. Bringing the smartline and bullet graph together we can describe a fuller story about what performance is actually doing:

The smartline tells us that recently there was a dramatic upward shift in the measure, which moved it well beyond its two interim targets. And then the bullet graph adds more to the story by showing the actual size of the gap between performance now, and the stretch target of 80%.

DISCUSSION:

For the first five people who respond in the comments on the blog, you can send me the data and target for one of your performance measures, and I’ll create a smartline and bullet graph for you, and write a blog post about the story your measure wants to share.

Facebooktwittergoogle_pluslinkedinmailby feather

Speak Your Mind

Your email address will not be published. Required fields are marked *

  1. Chris Witte says:

    I am a Business Intelligence architect and I found your blog a few months ago while researching visualization and performance best practices. I am quickly becoming a Stacey Barr groupie and was planning to purchase your recent book this week. I love bullet charts for their high “bang / $” ratio and I’m getting converted to XmR concepts through this blog. As an end-to-end BI guy, sparklines / smartlines and Bullet charts really pack the important stuff in an intuitive way into a tiny sliver of screen real estate. I see these two underappreciated gems rising in market value sharply with mobile device adoption and embedded Business Intelligence. I recently accepted my new job and we’re in need of some serious performance measurement assistance in this organization. I would be honored to get some personalized visualizations from the rock-star Stacy Barr. If I make the cut for your services giveaway contest, please let me know and I’ll deliver you some relevant data when I get back to the office tomorrow morning.
    Thanks!
    First time caller

    • Stacey Barr says:

      Welcome, Chris! Hey I love your phrase ‘a tiny sliver of screen real estate’ – it’s so true. I wish more BI guys and gals understood this.

      You make the cut – please email the data for one of your performance measures to info@staceybarr.com and mention in the subject line ‘smartline bullet graph giveaway’. Looking forward to seeing it!

  2. Ray Grant says:

    Stacey, the addition of the bullet chart is a nice to have but not a necessity when the Smartline is giving you the information at a glance.
    We track our Suppliers quality performance using the smartline with the identified target. So not only are we looking for changes in performance over time we are also seeing our suppliers ability to achive an acceptable level of performance. I’m seeing the addition of a bullet chart as a duplication of effort and may even prompt a question ‘why’ and we don’t want questions. Charts should be clear without the need for questions. Or have I missed a point? By the way – I enjoy Measure Up and use its little tips to push points across the business. – Ray.

    • Stacey Barr says:

      Ray thanks for your point. I do agree that when the smartline can show you how far you are from target, you don’t need the bullet graph. But when the smartline is being displayed without a zero baseline – which you do so that the smartline can fill the chart area – the distance between actual performance and the target is amplified. The bullet graph has a zero baseline, so the distance between actual and target is accurate. That’s the benefit in my mind.

      Oh, another benefit is when it’s not clear at first glance which direction is good – whether actual performance should move up or down – and so it’s not always obvious if performance is better than target or worse. The shaded areas of the bullet graph make that very clear.

      So I do kind of like them both together, for a fuller picture.

  3. Kirsty says:

    The bullet chart certainly helps present performance gaps at any given point. I have been struggling with understanding how we could use the XmR chart to show progress over time for indicators which have very few measurement periods (normally baseline and endline) and have concluded that it really only fits to present data where monitoring is done on a regular basis, and XmR charts are thus not really suited for impact evaluations. Although it would be to know how you think they could be used, because the concept of natural variation is so important for managers to understand. Its all in the context. And also because using result based management approaches with
    mandated targets, means that managers focus on chasing the numbers rather than understanding better how to improve the implementation process which is what would drive better performance towards meeting these unrealistic targets. Without a change in attitude of donors to really focus on learning, M&E in development will still be seen as a punitive exercise, which comes down to numbers. Sigh.

    • Stacey Barr says:

      I hear you, Kirsty! It’s frustrating for sure. But when you’re measuring something irregularly (you call it impact evaluations), you’re right that XmR charts don’t help. They help when you’re monitoring something continual, that you want to manage through time, and not just after the fact. If you can devise some lead measures for your impact indicators, those lead measures would be good to measure more frequently and then use XmR charts for those.

      Sometimes though, it’s also worth seeing if your impact indicators really do lend themselves to being measured more frequently, so you can actually manage their performance and not just celebrate or sigh at your endline.

  4. Jeff Wilkins says:

    Hi Stacy – Next to quality control measures, our most important measures is on time delivery to our customers. We build custom magnetics (transformers and such) and there can be a lot of variation over time as we ramp up new projects, stabilize production and then ramp down. Here is our performance measure for this year compared to last year. Last year was a difficult year as far as on time delivery is concerned. Reaching our target of 80% on time has proven to be a huge challenge. Best regards, Jeff

    DYCO Delivery Performance to Promise Date – 12 Months Rolling ended July 2014 Average days late: 4.2 days
    2013 2014 Grand
    Months Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Total
    late 47 36 16 15 15 55 27 35 47 43 50 9 395
    on time 116 112 123 97 69 126 85 85 105 121 99 104 1242
    Grand Total 163 148 139 112 84 181 112 120 152 164 149 113 1637
    Rate 71.2% 75.7% 88.5% 86.6% 82.1% 69.6% 75.9% 70.8% 69.1% 73.8% 66.4% 92.0% 75.9%
    Avg Days Late -9 -6 -2 0 -2 -7 -4 -7 -5 -5 -2 -1 -4.2

    DYCO Delivery Performance to Promise Date – 12 Months Rolling ended Dec 2013 Average days late: 8 days
    2013 2013
    Months Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total
    late 60 62 55 98 77 56 25 47 36 16 15 15 559
    on time 92 76 85 82 77 98 97 116 112 123 97 69 1124
    Grand Total 152 138 140 180 154 154 122 163 148 139 112 84 1686
    Rate 60.5% 55.1% 60.7% 45.6% 50.0% 63.6% 79.5% 71.2% 75.7% 88.5% 86.6% 82.1% 66.7%

  5. Hi Stacey

    I have found your blog articles about performance measurement insightful. In my organisation performance measurement is a relatively young function <3 years old… it is a healthcare not-for-profit and we are still getting to grips with ensuring that our KPIs are the right ones and are driving performance improvement in all areas. We are in the process of evolving and refining our measures and balanced scorecard – and use statistical charts but not the bullet graphs or smartlines…. as yet…

    Sabin

  6. Mike Davidge says:

    Hi Stacey

    I have not come across bullet graphs before and they are an interesting concept. When trying to relate current performance to a target, I use a full XmR chart and add in the target as a line. Now I can use the mean and process limits to determine not just how far we are from target on average but more crucially how likely we are in future to hit the target. In the literature, this is referred to as ‘Capability’ and there is a fancy calculation you can do to give you the percentage likelihood of hitting a target. However I prefer to simply direct people to look at where the target line is. If it’s inside the process limits then at some point you will get performance that doesn’t meet the target. So you will need to work on that process to improve it – either by changing the mean, reducing the variation or both.

    • Stacey Barr says:

      I prefer to keep it simple too, Mike. Thanks for sharing what you do. Your approach is very approach for most performance report layouts. The bullet graph and smartline are specifically for dashboards, where you want to display a lot of measures in a single view.

    • Adding a target line to an SPC chart is not considered good practice. I would suggest that it not be done – use the bullet chart instead. Nice post Stacey.

      • Stacey Barr says:

        Hey Bill, thanks for your comment.

        I don’t like adding target ‘lines’ at all, but I do like using the ‘dots’ sitting above the dates by which we’d like to achieve an improved level of performance. I’m curious what you might think of that, based on your rationale for not liking target ‘lines’?

  7. Tommaso says:

    Very interesting use of ‘smart charts’. I have sent you via email some comments.

Connect with Stacey


Haven’t found what you’re looking for? Want more information? Fill out the form below and I’ll get in touch with you as soon as possible.



*We respect your email privacy.
PO Box 422
Samford, Qld, 4520
Australia
Stacey Barr Pty Ltd
ACN: 129953635
Director: Stacey Barr
Simple Share Buttons