Ecology of a marketing instance

This post is best viewed on a desktop computer or laptop.

When I think about words on a page, I think of ecosystems. To me, the parallels are clear: words are like species with defined roles (nouns, verbs, adjectives, and so on) in a textual and graphical ecosystem that ultimately provides a service – to convey a story. Do you follow me?

I've been wanting to explore this analogy, so when Ute collected marketing literature at last year's Labvolution Trade Fair in Hannover, I seized the opportunity (admittedly very slowly – Labvolution was in May). After pruning the collection of flyers, brochures and postcards to eliminate catalogs, assets in German and pamphlets that didn't feature either a product or a service, the final set included 67 handouts. We extracted the text from the headers, body, figure legends and calls-to-action to create a database of the content. Then, while Ute examined the traits of a good marketing asset, I took an inexperienced and fully unprepared dive into the language. It's taken me a number of programming classes, the surreptitious use of my daughter's colored markers, and a new pair of reading glasses to resurface with an outcome.

And here it is, summarized in 3 graphics.

I'll cut to the chase: The author of each handout was careful to avoid repetitive wording (see Graphic 3) and maintain a light balance between text and space (see Graphic 2). But as a whole, language across all documents was constrained where marketing can be most impactful – textual imagery.

Authors are all using the same adjectives and adverbs to describe their products and offering.


Graphic 1 – A landscape in form and function

Just like a natural landscape is a mosaic of abutting and interacting ecosystems, Labvolution was an instance of marketing with a pattern of messages defined by handouts. Some documents told stories about instruments, others about consumables, and yet others discussed both. Some addressed imaging needs, others offered tools for cell culture. And some boldly showcased products for several applications.

Some thoughts:

  • The lines connecting the documents mean nothing, zero, zilch. But they make the graphic look more interesting.
  • General lab needs, sample prep and detection ruled Labvolution in 2019. Then again, maybe the collection wasn't exactly random.
  • Given the emphasis, analytes like DNA, RNA and protein are likely to be among the most frequently used nouns in these texts.


Graphic 2 – The lightness of language

The density of populations in an ecosystem play a key role in determining its function and health. So too does the density of language in a document. While lengthy text is generally avoided, cramped text can be even more off-putting. Messages at Labvolution were conveyed lightly – with either few words or lots of pages. Interestingly, the outliers on the dense end did not correspond to documents ranked as messy. Layout can make up for word density.

Some thoughts:

  • I hold my breath when I trace lines. It's exhausting.
  • I expected the density distribution to be skewed by extreme outliers. It wasn't.
  • I wanted this graphic to also show the range of document formats (postcards, DIN A5, square). It didn't.


Graphic 3 – Diversity for sustained attention

Species diversity brings stability, resilience and beauty to ecosystems. In language, word diversity retains the reader's interest with richness of meaning and ever-changing scenery. Thus, word diversity also brings stability, resilience and beauty to a message. Each flower pair shows the average per document diversity (ie., the average of the diversities calculated for each document) and the overall corpus diversity (ie., the diversity of the content from all documents as one body of text) of just nouns, just verbs, just adjectives & adverbs, and of all words. Stopwords (e.g., but, that, or, so, etc.) were removed in the analysis of all words but kept in the analyses by word types.


Some thoughts:

  • It's really hard to write a long word in a small leaf.
  • "as" can be an adverb.
  • "is", "are", and "be" were the most common verbs. Duh. Next time lemmatize the verbs.


Top 10 adjectives & adverbs
(no stopwords)

high, available, sample, highly, lipid, different, easy, well, reliable, fast

Top 10 nouns

DNA, analysis, system, RNA, sample, cell, PCR, data, application, product

Top 10 verbs
(lemmatized; no stopwords)

use, include, provide, base, allow, offer, design, optimize, develop, test

What have we learned?

As expected from the focus on sample prep, detection, and general lab needs, "DNA", "analysis", "RNA", "PCR" and "sample" were among the top 10 nouns used in the language. Of the 3 word types examined, nouns were the most restricted in terms of diversity; and the 2 most heavily used words overall in the text corpus were nouns. This makes sense. Science is a stickler for terminology.

Once we remove the stopwords "is", "be" and "are", the most commonly used verbs were also characteristic of the dominant topics (test, design, optimize) and of the nature of the documents (include, provide, offer).

The adjectives and adverbs were surprising. Although this word type has the most freedom to vary and space was not a limiting factor, adjectives and adverbs were not exceptionally diverse. While diversity was high for each individual document – almost 8 out of 10 words were unique — diversity across the entire text corpus drops to about 3 unique adjectives or adverbs out of every 10. That is, authors are describing products and services with the same limited vocabulary.

This is an opportunity. Broaden the imagery of your marketing text to stand out in the crowd.