Online writers have something that previous writers did not have instant access to, powerful reading data from analytics. While in the case of “a show of hands,” I found the analytics to be both instructive and depressing (since you can see how far registered users had clicked through), in the case of “Living Will” I find the stats to be provocative and inspiring.
Let me explain the way these work. “Living Will” is an interactive story in the form of an electronic Last Will and Testament of ER Millhouse, a coltan magnate who is using the Will as a last opportunity to reach out to punish and reward his closest relations and associates. At each stage of the storygame, the reader can make choices, first by selecting which one of the heirs she will inhabit (his son Nigel, his daughter Salomee, his errand boy Kip, or his gardener Gerald/Gerard). As the reading of the Will progresses, the interactor can choose whether to remain contented with the portion she’s been given or instead to filch from the inheritance of one or more of the other heirs. The story ends with an opportunity to choose whether to save the patriarch or let him die, assuming that the heir has sufficient funds, lest medical and legal expenses which mount with every click of the document.
Before I published the story, I inserted some analytics triggers into the code, so I can see how many readers choose which events. For this, I’m using Google Analytics. Now, three years after publishing that story, I can examine how the 4,587 unique visitors behaved during 5,456 visits and 9,454 page views.
The results are provocative if nothing else. Some highlights of the undigested results:
- Most readers who could save Millhouse did save the bastard (1,660 to 836).
- Readers most often stole shares from Kip (2,627), followed by Nigel (1,673), Gerard (1,486), and Salomee (28).
- Most chose to read as Nigel (3,211), the default character, followed by Gerard (1,864), Salomee (1,731), and then Kip (1,305).
Now, this is very raw data, and I should note that some of these behaviors are incentivized in ways evident in the code which I can explain another time. Also, I should note that my highest activity on the site came during the IFComp when readers were presumably exploring the work for reasons other than merely pleasure, but to test out the nature of the system. That said, in future analysis of the data, I can parse it out according to how the readers got to the site, how long they spent on the site, or what part of the world they live in.
I also, have new data from the version in French (trans. by Ugo Ellefsen, Alex Gauthier, Myriam Gervais-O’Neill Ă‰milie Robertson), which feeds into this same data pool.
Which is just to say that now my authoring experience can go beyond merely looking at (obsessing over) how far readers made it through the story to see the choices they made as they moved through the story. While Hollywood might use this info to optimize their storytelling for maximum viewer experience and the construction of a sure-fire blockbuster mega-hit, I find the analytics provide me with a fascinating reading experience, feeling a bit like Millhouse himself, smiling wickedly to see the way the heirs play and are played by his final act of will.