Welcome to a short presentation about how Jellybooks collects and uses data about readers to develop better book discovery and marketing approaches. You can follow the speaker, Andrew Rhomberg, on Twitter @arhomberg and his company @Jellybooks. This presentation is available for download at http:// jbks.co/ZNifGM To have titles that you have published included in the Jellyfactory, register at http://vip.jellybooks.com .
How does one avoid Amazon, the 800 pound gorilla, that dominates the ebook and digital publishing industry? Well, Jellybooks was founded on the principle of avoiding Amazon and creating alternative approaches for ebook discovery, ebook marketing and ebooks selling.
Where is Amazon weak? Most people go to Amazon to buy books that they already discovered elsewhere. They discover these through recommendations from friends, reviews, online tips, after browsing a book shop or similar. Amazon is the destination for cheap book, but it is not a good online substitute for browsing a physical book store or library just for fun. Amazon encodes all books, even samples with its proprietary DRM. This makes it tricky to share such samples with friend. Amazon does not share its data, even publisher’s sales data is delayed and not very granular, yet some of the most interesting opportunities arise from mining such data. Reaching an audience on Amazon is mostly guesswork. There have been schemes to game the system, such as making ebook free to climb the popularity charts, but Amazon keeps changing its algorithms without warning and its rules are opaque.
At Jellybooks we focus on discovery and have developed our business model around it. Our goal is to delight users with new book discoveries. In the process we aim to create systems that allow publishers to promote books through targeted promotions without alienating the consumers. Such promotions should be perceived as relevant and appealing by the reader and the intelligent use of data makes this possible. In tackling the problem, we have assumed that there is not one, but many different forms of discovery and we focus on 5 in particular: serendipitous, social, distributed, data-driven and incentivised discovery.
Project Cranberry by Jellybooks was launched in March 2012 - almost exactly 1 year ago - with a focus on serendipitous discovery. Project Cranberry is an ongoing attempt by Jellybooks to recreate a fun and entertaining discovery experience online. Its goal is to entertain the avid reader for 2-5 minutes during a break and help them discover some new book and provide samples (first 10%) for later reading.
To make it fun and easy we focused entirely on book covers without any text or other elements to distract the user. The navigation elements were kept to an absolute minimum with 1+4 icons in the top navigation bar. The website is responsive, which means the layout and number of covers automatically adjusts to the user’s screen size on smartphone, tablet, laptop or PC.
If a cover grabs the user’s attention, they can click on it to get 4 options: Sample the book for later (offline) reading on an ereader, smartphone or tablet. Share the sample with other online via Pinterest, Twitter, Facebook or email (more than ½ the reader use email). Sign up to be notified when there is a special ½ price deal (bingo – we just identified a price-sensitive customer) Or buy the book now and we link your though to you local Amazon, iTunes or online book shop (ebook or physical).
In Summer of 2012, we followed up with Project Blueberry to support social discovery. With this release we created a set of social media tools and optimized them for reader engagement. They are designed to maximize the number of users who download or share a sample after seeing it featured on a social network. Our solution is based on fine-tuning HTML mark-up specific to each network so that we can display cover, title, author and synopsis alongside the link.
One such example are Twitter Cards, where cover, author and synoposis are displayed in expanded tweets or when a user replies to a tweet. This effect is based on the use of Twitter Cards. Similar tools exist for other social networks, but the rules differ form one network to the next and our sharing/download links have a set of optimized mark-up for each social network, so that auhtors, piblicist and others don’t have to worry about this.
Project Cherry is currently in beta and will launch at the London Bok Fair. It allows author, agents, publishers, reviewer or bloggers to use the Jellybooks sample system to place buttons and widgets on their own homepage, blog or website, so readers can discover books in context and download an ebook sample for later reading. All buttons and widgets make use of the same social mark-up system as normal jbks.co sample links.
The user does not need to visit the jellybooks.com website, but can download an ebook sample directly from the author’s, publisher’s or blogger’s website for later reading. All samples are DRM-free, can be freely shared and are supported by “Send to Kindle” (.mobi files), “Send to Readmill” (epUB files) and the Jellybooks “access and download form anywhere” cloud infrastructure.
Now to the most interesting part of today’s presentation, which is Project Elderberry. Project Elderberry is a cross-platform effort to collect, store, distribute and analyse data of how readers recommend and share books online. It is being supported by the UK’ Technology Strategy Board and is currently under development.
We still tend to think tend of books in the traditional form such as hardback, paperback, mass-market paperback or now ebook, but in so doing we still think of books as “packages”.
However, increasingly a book is just a URL for the product page on Amazon, be it a link to an information page about the book, a link to a sample file we can download (i.e. jbks.co link) or a link to re-download an ebook we have previously purchased (for example on iTunes). In other words books are increasingly represented by an online URL and not the “container” or digital file.
URLs can reveal many things, such as what people share (as in the book links they include in their emails, tweets, Facebook posts, etc., when, where, on which device, etc. (most of which can be found in the meta-data of posts on social networks).
However, it can be quite messy and developers with the rights skills in visualization, machine learning and data mining are not necessarily well versed in the specialised language and practices of digital publishing.
Therefore, Jellybooks creates a clean and structured data set. We retrieve, index, cache, organize and clean up the data, so that others can make easy use of it.
We provide this cleaned-up data through an open API to 3 rd parties, so they don’t have to worry about data integrity and can focus on creating algorithms for visualizing and analysing the data.
At Jellybooks, we also do our own data-mining and we use the collected data to understand readers and what they read. One of our goals is to develop better discovery and recommendation algorithms for books.
At Jellybooks, we also use data to understand who influences whom and how this is subject, category or genre specific. Our goal is to find the right audience for each book (finding books for reader, and readers for books are related, but not identical challenges!)
Our goal is to understand online influence networks for books so that these can be targeted for promoting books without disrupting the natural flow of communication. Our aim is to target promotions, give-aways or special deals without them being perceived as intrusive or as spam. Publishing is extremely diverse, so targeting is crucial to an effective online discovery and marketing strategy.
The Jellybooks API is free, as we wish to encourage innovation and make the Jellybooks ecosystem, including the Jellyfactory, more attractive to publishers, developers, authors and readers. We make money, by charging a small commission on special deals (part of our incentivised discovery system, that we cannot cover in more detail here due to time constraints) and promotions where we help publishers reach a new audience. We also take a cut on each 3 rd party app that uses our API and is sold through our app market. This can be a small inexpensive visualization apps, or more sophisticated and pricey analytics tools. We also receive a small affiliate commission from buy links we refer to 3 rd party retailers, but this is a very small and insignificant part of our business model.
We are currently a focused team of 8, which makes us very agile, but we are continuously growing. Thanks for listening and visit us at http://jellybooks.com and http:help.jellybooks.com to discover more about Jellybooks. To sign up for the Jellyfactory register at http://vip.jellybooks.com or if you have over100 titles, email us at email@example.com. You might also want to check us out at http://pinterest..com/jellybooks or http://jellybooks.tumblr.com.
@arhomberg Magical Times for Book Publishing@Jellybooksjbks.co/ZNifGM
How NOT to be Amazon…Avoid the Gorilla! Amazon Jellybooks • A place to buy - you • A place to discover new already now what you reads and share them want most of the time with others • DRM on everything and • DRM-free samples to only for Kindle apps share + read on any app • Indiscriminate discounting • Earn yourself a better for everyone discount • Data is proprietary • Open data policy • Reaching audience is • Analytics for audience guesswork for author discovery
Berryliscious Path to Discovery Berryliscious Path to Discovery1. Serendipitous Discovery Project Cranberry by Jellybooks2. Social Discovery Project Blueberry by Jellybooks 3. Distributed Discovery Project Cherry by Jellybooks4. Data-driven Discovery Project Elderberry by Jellybooks5. Incentivised Discovery Project Pineapple by Jellybooks
the.book.as.URL• Open and accessible even at Amazon• Can be tracked as in how shared, tweeted, pinned, liked, rated, reviewed• Can track the detail of when, where, who, on which device…• Can track engagement generated as in download of sample, purchase, share… Lots of reader-centric meta-data
…but it is messy• Many different 3rd party APIs• Do coders speak (let alone like) ONIX?• Different ISBNs for each edition• Non-standard format and meta-data• Some social media data data not cached… scares non-bookish folks: coders like to do funstuff, not “cleaning” data
Project Elderberry by JellybooksWe•retrieve•store•clean•structure•index•organizeURL data
From Data to Insight – 3rd PartiesProject Elderberryprovides bookish datain a form hackers like,so they can focus on:•data visualization•Data-mining•Price optimization•Book recommendations + many more applications
From Data to Insight - JellybooksAt Jellybooks we use data to•Understand what people read (genres, categories, interests)•who they influence•who influences them•where and how they discover …so we can find the audience for (every) book.
Use Data to Find an Audience Infer reader interest from: •What samples they download •What samples they share •What samples they re-share •What they quote It’s about what people do, not what people say! (ignore ratings, reviews, etc.)
Use Data to Reach an AudienceDeduce from link sharing:•Who is influencing a reader ?•Whom does a reader influence ?•How is it category/genre/subject specific•Where do people discover?•How do they transition from discovery tosampling, purchasing, sharing, quoting,reviewing, praising, flaming…
Elderberry Business ModelFree• to submit books to the Jellyfactory• to access API (for in-house or non-commercial use)£€$• services to identify audience for a book (promotions, analytics, etc.)• licensing web apps through the Jellybooks app market
Team Jellybooks@andyroberts_uk @madebysplendi @jamiebrooker @johanbrand d @ismasan @sidane7 @gwynmorfey @arhomberg presentation available at: jbks.co/ZNifGM