This is just a quick follow-up to my post on using word clouds or wordles for competitive intelligence. ScrewTinny – the competitive marketing intelligence service that I’m developing, now creates wordles based on vendors’ marketing copy, but there was a problem when looking at wordles containing the marketing copy for more than one product.
I’ve now updated some of the wordle-creation logic in ScrewTinny so that word counts from multiple vendor are now normalised. What does this mean? Well, previously I used an absolute count of the number of times each word occurs. This is fine when the wordle represents only one vendor’s marketing copy. But when combining the marketing copy from more than one vendor, the resulting wordle could be skewed towards the vendor with the most copy.
The problem with an absolute count of word frequency is that if ScrewTinny had found more content for vendor A than for vendor B, the results for vendor A would be more prominent than those for vendor B. What I now do is workout the ratio of the frequency of each word to the total number of words ScrewTinny has ‘read’ for that vendor.
Using this new logic, I’ve re-created the wordle showing the most frequently used words across all 13 ESB vendors that I looked at in the market ESB product segment.
ESB Segment (13 vendors) Marketing Copy Wordle
I’ve also reproduced the gallery of each of the 13 ESB vendors from my previous.
*Caveat lector. As I’ve stated before. ScrewTinny is currently being developed and tested and as such erroneous tokens and symbols are not yet filtered out of the wordle output. This will improve over time.
What is ScrewTinny? ScrewTinny, is the competitive intelligence service I am building to automate my REPAMA Methodology. ScrewTinny will look to visualise vendors’ marketing strategy by reading and analysing marketing copy.