PowerPoint files – presentations or documents? A search challenge

I was browsing through some back issues of the journal Organizational Studies and came across a very interesting paper by Professor Dr. Dennis Schoeneborn (University of Zurich) on the use of PowerPoint in organisations, especially consulting companies, for both presentations and as a format for documents. Very often project reports need to contain many charts and diagrams and under the time pressures of a consulting engagement it is a great deal easier to work with PowerPoint than with Word to make visually elegant reports. Schoeneborn found that in the consultancy (not named) that he used as a case study a collection of 565 primary project documents contained 495 PowerPoint files. His paper explores the issues about whether the need for consultants to use PowerPoint as an accepted ‘standard’ outweighed any organisational guidance on its use.

This paper started me thinking about the implications for search. When presented with a list of results with a file format indication (either in the hit record or in a filter) PowerPoint files might be overlooked as a source of in-depth information and analysis because the assumption is that they are just corporate presentations with the standard five bulleted items per slide. In addition the ranking algorithm may be biased against PowerPoint files for the same reason.

The next issue is that even when used as documents these files may have relatively few words per page as the usual approach is to use graphics and charts supplemented by a text commentary. Not only may there be a paucity of keyword terms but depending on how the graphics content  was inserted in the document the crawler and document filter may not be able to extract terms from this content.  Excel tables and charts are often inserted as images so that they can be sized more elegantly. Based on my own experience version control of these PowerPoint documents is often weak, with small changes made as an outcome of a client meeting without any of the version metadata that is (hopefully) applied to documents in a Word file format. PowerPoint presentations also tend to be given ‘clever’ titles and so any bias towards the text content of the title could be negated.

The paper is based on just a single case study but it aligns sufficiently well with my own experience to cause me to look more carefully in the future at the use of PowerPoint in an organisation. If it is used extensively for project reports then some test queries on a collection of PowerPoint files might usefully be undertaken. If nothing else guidance on version management would probably be a good move.

Martin White