Emerald Group Publishing Limited
Copyright © 2000, MCB UP Limited
Counting on numbers can add value
Keywords Statistical forecasting, Presentation graphics, Reports, Management reporting
One of the results of the information age is our utter acceptance of the visualization of information. In the span of a century, we have moved almost completely away from the Victorian era's way of communicating thought, which was marked by the weighty use of adjectival and adverbial embellishment in everything from love letters to narrative business writing. The endemic emergence of the Internet to lead us into the twenty-first century reinforces just how far we have come. Now, "pictorial metaphors" across our computers' tool bars symbolize actions that we take, and we have come to rely upon symbols (or numbers) to replace words.
In fact, today, in the "executive information age", we are more likely to try to report, describe, convince, and plan by using some numerical mode of communication than use a strictly narrative mode for imparting information. It is a simple fact of life that in environments of information overload, what we want to say can often be more concise, comprehensive and clear if we describe it in a graph, chart, timeline, or by supplying a final number or a formula for computation.
Evermore sophisticated computer programs make scenario building, forecasting, cause-effect explanations, survey results, annual reports, and strategic plans easy to generate, simple to modify, and very visually appealing.
While information managers are creating more of their communications exclusively or in great part with numbers as the basis, they are also faced with information being presented to them in numerical formats. In these instances, the managers must be able to evaluate the veracity and relevance of the content, and also be able to critically review the methodologies used to collect data and present it.
The bottom line is that numbers make good spokespeople whether we are using existing numbers to help us make decisions, or numbers that we have generated to help others make decisions that will have an impact upon us. The focus of this column will be to review some basic points in presenting numerically or visually based information to others.
When you decide to make your point by collecting data (numbers), have a goal or intended action in mind. Do not simply ask your clientele how often they turn to the Internet for information unless you are considering some specific action (canceling hard copy or a proprietary database) for which collected data would be relevant.
There are numerous ways to collect data within an organization. A survey or a sample is an acceptable way in most library and information-based settings to gather information for decision-making. A government-sponsored census may need to count everything and everyone, but very few of us need a universal count to draw valid conclusions. The key here is to make sure that the conclusions are valid. Therefore, the survey must not be biased and the sample must be representative. Also, remember that every organization, even very small ones, will have anomalies. How these are dealt with may realistically depend more on organizational culture than any accepted statistical standards.
Another method of data collection commonly used in the LIS environment is to maintain transaction statistics. Based upon a predetermined list of categories, every interaction is fitted into a category and counted. Categories may include a kind of service (directional, research, reference, bibliographic instruction); type of client served (student, public, accountant, engineer); or the day/time that service was rendered.
A third way of data collection is often labeled as "experience or event" collection. Here, the idea is to analyze people's reaction to something to which they have been exposed, to extract out the common elements, and to report the disparate ones. Then, transform the experience or event into "data" by conducting interviews or debriefings after the fact. Some situations lend themselves to real-time debriefings as the subject is experiencing whatever is tested.
Yet another manner of using numbers is by by-product. One looks at what has been collected for perhaps another reason, and then applies that number to a new or unintended purpose. In libraries where librarians and information professionals bill their time, the billable/non-billable hours per person or unit are often used to justify additional staff or to make the case that the unit needs some equipment to free staff from non-billable activities.
Finally, we can use the experience of others to draw our own conclusions. This method simply involves reading published or disseminated results from which one can show similarities or differences that justify adopting or dismissing these conclusions internally.
Make sure that everything adds up
When showing a trend(s), be sure that you are following actions over a significant and all-inclusive time period. You do not want to make policy based on exceptional circumstances. Most entities experience "seasons" in the commercial sense, but many have never identified them.
If your "business" is subjected to large variations of peaks and valleys during the year, then it is always a good idea to graph-plot your plans along the lines of these fluctuations. For example, do not schedule payments for a major new system installation in the same month or quarter in which your business pays its taxes. Regardless of how small a sub-unit of a larger whole you may be, it is always a good idea to make yourself aware of recurring cashflow issues, special circumstances anticipated in a given time period, and opportunities to limit the financial impact of your needs within the greater business. Remember to incorporate these milestones on any visual that you may generate.
When presenting numbers, there are a few common pitfalls that require you to be on your guard. First, there is the temptation to switch cause and effect. It might be true that If A, then B, but is it also true If B, then A? Second, is the correlation being drawn strictly by chance? Third, is the conclusion true, but not for the reason stated? As was alluded to above, also take care to avoid using a sample that is not representative or a survey that is biased.
Presenting the numbers
Visual appeal plays a large role in convincing us to look at numbers, and to even accept their veracity. This may be for better or worse! When presenting numbers in any format it is, therefore, critical to use whatever advantage you have to bring attention to your numbers. It may be helpful to keep the following points in mind:
If you are showing dramatic trends, a line graph is an excellent communicator.
When large numbers are being addressed, the best tool is a bar graph.
To make comparisons between timeframes (month-to-month, year-to-year), use a table.
To show percentages, use pie charts or bar graphs.
To overlay an illustration or display several scenarios, bubbles work the best.
Use colors whenever possible, and depending on the point that you want to make you can use contrasting or complementary color schemes. (Red should be avoided unless you want to convey crisis. Black is the best for baselines or constants, but is not upbeat enough for good news or for setting high expectations.)
If possible, use colors in lieu of different geometric designs (a line graph of dotted and broken lines, with an addition of cross-hatched lines to make yet a third point is much more difficult to follow, and will not enable you to take advantage of the psychology of color).
Whenever possible, use graphics or clip art to represent the subject of the presentation. (If you are addressing human resource needs, then use stick figures on your graph. If you are discussing technology, use computers. The fewer words that need to be processed by the reader the better, and he/she will understand what your icon represents.)
Keep font and font sizes as uniform as possible. Too much variation in these tends to confuse, and distract from the numbers.
Narrate as little as possible. (Referring back to the common fallacies, while it may be true that a picture is worth a thousand words, a thousand words are rarely worth one picture.)
Do not try to make too many points in one numerical representation. The maximum number should be three points.
If you are presenting a package of numbers, use a very brief title as the header for each one. This descriptor should explain what the graph, chart, or survey results are designed to show. (Do not have the descriptor state the conclusion. The reader/interpreter needs to do this by him/herself or you will run the risk of making your information looked rigged.)
Counting on numbers can add up to success. There is a social expectation that we will use numbers, visual symbols and visual representations wherever we can today, because the Internet and computer technology in general has socialized us in that direction. Visual representations tend to be accepted at face value, and as containing more reliable information. Technology has freed us from the burdens that generating and editing numerical displays formerly presented by first generation spreadsheets. Something as basic as word processing has given us an entire visual vocabulary of symbols that impart action, and are understood across national (language) boundaries. Most importantly, numbers and visual representations tend to communicate more efficiently than do words. But like words, we must choose our numbers carefully and understand their potential for multiple (and unintended) meaning.
Roberta I. Shaffer Dean, The University of Texas at Austin, School of Library and Information Science