1. In simple terms, what is MacGuffin and how did it come about?MacGuffin’s a literary jukebox: a website and app hosting fiction and poetry in both text and audio form. Writers have to upload an audio reading of their work along with their text; readers can read, listen, or toggle between the two. We hope to encourage self-publishing authors to follow the lead of amateur podcasters, who’ve shown in recent years that technological barriers to entry have fallen away. You don’t need a studio to get decent audio (in fact, you can get a near-broadcast standard recording with just a quiet room, a smartphone, and a bit of practice).MacGuffin uses a ‘broad-folksonomy’ hashtagging system for content curation and discovery. End-users can add tags to anyone else’s content, to describe it (e.g. #bears #woods #porridge #dangerousblondes #furniturerepair) or add it to a reading list (e.g. #jimsfavouritefairytales). Then whenever anyone searches for one of those tags, that content will be returned. Tagging a bunch of stories or poems with a unique tag creates a playlist you can share – kind of like how people used to make mix tapes for friends.
The idea arose a couple of years ago, when we built an app called LitNav, which featured short stories (many in translation) set in cities across the world. All of the stories were in text and audio form, and categorized in terms of length, location, genre, and so on. One of the most common responses to LitNav was from writers, saying, ‘I have a story that’d be perfect for this – how do I put it on there?’
2. MacGuffin was created in partnership with the NESTA Digital Research and Development Fund For the Arts, which seeks to support projects that make the arts more accessible through the use of technology. What was this partnership like? What did you learn from working with people from the partner organisations, such as fffunction?
fffunction brought a lot of UX and data architecture nouse to the project, while Manchester Metropolitan University took on the research and app development. It’s been great to work with people with such broad experience across digital development and research disciplines; it’s certainly increased Comma’s (and my own) expertise. Probably the biggest take-away for Comma is the value of agile development: co-design with users, early prototyping, testing (both in the lab, and ‘guerrilla testing’), and iteration. We’re pretty pleased with the user-interface aspects of the website (I think the ‘upload story’ function is miles more usable than comparable functions on self-publishing latforms built by huge corporations with infinitely greater resources), which is testament to fffunction’s careful approach to UX.
3. What would you say are the top three ways in which digital platforms have disrupted publishing? What other apps or platforms have most impressed you?
Most of the literary apps and platforms I like best originate from outside the traditional print publishing industry (whether built by non-profits, or communities of volunteers). They tend to be better at discarding models and structures that essentially just ape print publishing. I love the Poetry Foundation’s app, for example, which lets users find content according to theme. I think it’s been a big influence and inspiration to a lot of literature organizations building apps on a small-ish scale and budget. Librivox and the Internet Archive are great from a community/volunteer point of view. Likewise Archive of Our Own and PinBoard (the book-marking tool). Wattpad’s a useful model for user-generated content, and while it’s not really aimed at the same kind of readership/writership as MacGuffin, you have to pay attention to a platform that’s so successful in growing a community.
4. Each research project funded by NESTA has to answer a research question. Yours was addressing how ‘broad-folksonomy’ content curation can be applied to literature to learn about changing reader behaviour and taste. Have you drawn any conclusions yet? What have you found out from your research so far?
It’s pretty early days still, but we’ve answered the first part of the question by building a platform to facilitate this kind of broad folksonomy approach. Over the coming months, we aim to publish updated analytics data comparing read/listen events with location data (essentially, telling us the extent to which people prefer to consume audio content when they’re travelling/commuting). There are also lots of live analytics (for individual stories and tags), which anyone with a MacGuffin account can access via the website.
5. Why did you place so much prominence and importance upon searchability in MacGuffin?
Lots of literature platforms/websites feed readers curated content (or aggregate content from 3rd-party sources), but we wanted to see what would happen if we placed almost all curation in the hands of the writers and readers themselves. The idea is that, as more content is uploaded and tagged, it accrues into a huge, searchable database of literature. The system really comes into its own if you want to find literature according to theme. For example, a search for the tag ‘journey’ returns two of Shakespeare’s sonnets, two short stories in translation by contemporary writers, a classic story by Tolstoy, an essay about cheese by GK Chesterton, and two self-published stories by MacGuffin users. All of them are, in one way or another, about journeys. You can also search with multiple tags, so a crime-fiction fan on a 15-minute bus journey in Liverpool might search #crime #liverpool #15minutelisten, to find a story suitable for her commute. From the search results page, the ratings give you some indication (or forewarning) of what other readers thougt about the quality of the text and audio.
The tagging system has been both a technical challenge to implement, and a UX challenge: will readers take to it? Will they be sufficiently motivated to tag other people’s content? Are they put off by not having curated content visible when they first land on the site/app?
My first instinct was to take a purist approach to this: to insist users find content with tag searches alone, and hope they would acclimatize to it. But we’ve rowed back from that a bit, and now returning users can also browse some categories (Trending, Recently Published, Highest Rated, Most Completed, and Editors’ Picks). The rationale is to ‘gamify’ the experience for writers, motivating them to share MacGuffin among their networks in order to push themselves up the rankings (in Trending, Highest Rated and Most Completed). When there’s a flurry of new stories published, the Recently Published tab essentially acts like a news feed, while Editors’ Picks gives new users a taste of the breadth and quality of work on MacGuffin.
6. MacGuffin uses analytic tools so that writers can track who is interacting with their work, where they are located, whether they are reading it or listening to the audio file, what they rated the content, and whether they reached the end, or ‘dropped-out’. What effect do you think this knowledge has or will have on authors’ future creation of work?
IMO, the most interesting finding so far is that readers graze more than we anticipated. The MacGuffin drop-out analytics show that in many poems and stories, about 50% of readers/listeners drop out within the first 10% of the content (which, depending on the length of that content, might mean on the first page, before scrolling down, or exiting during the first few seconds of audio). This isn’t just the case with contemporary, self-published stuff (which might be attributable to the variable quality of self-published content). We uploaded a lot of classic, public-domain literature to MacGuffin (e.g. Chekhov and Joyce), and it’s a similar picture for them. I think this reflects the way people read: they browse a few sentences, decide it’s not for them, then move onto the next thing. It might not even be a digital phenomenon; I suspect it’s how many people browse in bookshops too. But it’s quite sobering to see it described so starkly on a graph. It’s a reminder of how important the opening sentence, opening paragraph, and opening page are.
Another thing we’ve noticed (though the sample size is not yet big enough to draw firm conclusions) is that readers quite often drop out when they get to a specifically dmarked break within a story, e.g. a paragraph break with an asterisk or section number, denoting a change of scene. It’s almost as though this gives the reader permission to pause and break away from the story, and seems to run contrary to the notion that short, sharp sections help reader retention.
Of course, drop-out analytics are only a part of the picture. The data from the classic content shows that, taken in isolation, drop-outs are no indication of ‘Literary’ quality. On the other hand, I do think that shepherding a reader from the start to the end of a story is one part of a writer’s job, and writers who are learning their craft can use MacGuffin to sharpen their work. You can examine the dropout stats of your story, edit that loose page or paragraph, then republish, and once you’ve accrued a few reads, compare the graphs for different versions.
7. MacGuffin has ‘professional’, award-winning writers alongside relative unknowns. How important to the project’s overall aim was this democratisation in terms of who can upload to the site?
Very important. The market has a poor track record of recognizing literary innovation (and if it does, it tends to be long after the fact). I think indies generally do better (though not always), as they’re working to smaller scale, and can accommodate more risk. Even so, they each have their own predilections, and there’s only so much they can publish. So self-publishing is a really important corrective, or safety valve.