Throughout the pyramid principle research class, I was in touch on LinkedIn with my former McKinsey communications specialist colleagues – we have a group with sometimes vivid discussions. I just wrote a summary of the findings of the class for that group, and I thought I might as well put it here, for the sake of other English language readers interested in the effectiveness of the pyramid principle – which is what we researched in a Master class at Groningen University.
The students carried out 7 small research projects: 5 experiments aimed at readers, one round of interviews with writers (learners of the pyramid principle) and one literature study. I’ll give you a summary of the results.
In true pyramid fashion, I synthesized the findings. The main conclusion is that there is room for nuancing of the pyramid principe, or perhaps even a need. Here is the support:
1. In some situations, it is good to be cautious with the pyramid principle. As you all know, it is a direct and forceful way to communicate, and many users are hesitant to use it in its full force when communication a message with negative consequences. This seems to be justified. In an experiment in which we compared two versions of an advisory letter about changing the opening hours of the university restaurant, one with good news (longer hours) and one with ‘bad’ news (shorter), we found that the ‘bad news’ version was evaluated by readers in terms like ‘forced’, ‘compulsory’, and ‘imposed’. Moreover, many readers of this version were unable to point out the main message.
2. There is much variation among readers in their appreciation of the pyramid principle. What some readers like about a pyramidal text (direct, personal, short, to-the-point), others don’t appreciate and see as too direct and too self-confident. Interestingly, some readers see the directness as less professional, probably because they are used to the more indirect and fluffy style of most professional writing.
For me, this means that in practice I am going to stress even more the importance of knowing who your reader is, including his or her report and style preferences.
3. Main message upfront is not confusing: hardly any reader starts reading a report from the end. However, the table of contents in a standard pyramidal report is confusing, because readers miss the familiar words like ‘conclusion’ and ‘recommendations’. Most readers do not recognize that it is all there right in front of them, by means of message titles. In other words: readers do not read the captions of a report as if it is a newspaper, and they need more grip in order to find their way around a non-traditional report.
I actually think this is an important weakness of the principle. A structure should be self-explanatory; message titles in reports do not seem to work that way.
4. It is unclear whether applying the whole pyramid principle is worth the effort. Writers consider it quite difficult to apply the principle, even though they recognize the value. What readers value, however, are the general characteristics of a well-written text, not necessarily the pyramid principle in its full consequences. Readers appreciate a clear structure and concise style – but more roads lead to that. It also seems to be that the main gain for writers is the preparation: they are forced to think before they start writing. That, too, is not pyramid-exclusive.
On the one hand, this lack of distinction stems from the limitations of our work: our experiments could not distinguish between ‘good writing in general’ and ‘pyramid’. On the other hand, I think this finding shows that there are good reasons for teaching writers in less demanding environments (than McKinsey) a ‘light’ version of the principle.
5. There is room to further develop the pyramid principle, on the basis of more modern insights in to human communication and cognition, that state the importance of creativity, metaphor, association, narrative… which makes the pyramid principle look at least one-sided, with its emphasis on rational logic.
The small scale of all the experiments makes it important to be cautious with generalizing. Yet I still think the findings are quite interesting – don’t you?
