Blackfriars is currently in the process of analyzing all the survey returned from our
new, larger marketing survey. But as we got into the analysis process, it became clear that Excel really wasn't the right tool for this job. After all, not only do we need to do tests on 300 answers to each question, but we also have to look at industry specific slices of the data, compare responses of B2B, B2C, and nonprofits, and also analyze the effects of company sizes on the responses. Frankly, the possibility of being afflicted with the strain of staring at Excel for days on end was too much for us.
We were considering buying either the SAS or SPSS tools when we stumbled across
The R Project for Statistical Computing. This open source statistics analysis package is actually a clone of the S language developed at Bell Labs. As a former Unix afficionado, I have the greatest respect for the work done at Bell Labs, so we decided to give it a try. The results have been wonderful.
Now we will say that R is not for the faint of heart. Unlike a lot of systems that are presentation systems that know how to read databases, this is really a programming language that knows how to do scientific and statistical plots. That said, it provides us with powerful automation for crunching data, and the speed that that automation gives us is worth the steep learning curve. For Blackfriars, it is the right tool for the job.
Clients won't see the R graphical output in our published research; we publish using the Adobe Creative Suite, and that isn't going to change because of the page and color control needed for professional publishing. But we do believe that clients will see more insight in our published research because we are now spending our time thinking about what the data means instead of fighting with Excel. Pictures communicate better than spreadsheets any day of the week; R is letting us see pictures in minutes instead of days. We can't say enough good things about the package.