Technical Details for Jellybooks DISCOVERY
Using virtual focus groups, Jellybooks invites test readers from its existing panel via email and via
Jellybooks will recruit the necessary number of test readers. We set a goal of recruiting
- Between 300 and 500 test readers for a single-title campaign
- Between 500 and 700 test readers per title for an A/B or side-by-side test
Publishers have the option to recruit test readers, as well, or exclusively through their own channels. In
the latter case a 20% discount applies. Jellybooks provides a unique campaign URL for each channel (email
list, Facebook site, Twitter or Instagram account etc.) that the publisher might wish to use for its own
recruitment efforts, so that conversion and audience engagement for each segment can be measured
Participants are paid no monetary compensation. Our experience of having conducted hundreds of test reading
campaigns over the past five years shows that receiving a free eBook and the knowledge that somebody is
paying attention to them is sufficient reward for participants.
Jellybooks RADAR campaigns can be executed by publishers with a global audience or can be focussed on selected markets
including UK and Ireland, Germany, Austria, Switzerland, USA, Australia, New Zealand and South Africa in any combination.
Territorial restrictions are available too and based on the IP address of the user’s device. Users reading outside
the designated territory see their access blocked.
Registered test readers receive eBooks that have been specially modified by Jellybooks to record their
reading and engagement data. Test readers can choose from a range of reading apps for smartphones, tablets,
laptops and PCs including Apple Books, as well as the Jellybooks Cloud reader. eInk based devices such as
Kindle, Nook, Tolino and Kobo are currently not supported, because these device support only the ePub2
Data is only transmitted when readers click on the purple “sync reading stream” button at the end of each
chapter. The system is essentially double opt-in and participants explicitly consent to their reading data
being collected. Participants also have the ability to view visualisations of their own reading data online
and access all the data that Jellybooks has collected and holds on them.
The eBooks distributed by Jellybooks are not encrypted but contain a range of “social DRM” features
including visible and invisible watermarks and other tracking and identification features. In addition,
participants know that they are being observed during the test reading campaigns which greatly inhibits
casual sharing and “piracy”.
Results and Analysis
Results are available in real time through Candy, the Jellybooks data portal where they can be viewed,
downloaded and printed.
A first analysis of trends is usually possible after two weeks when the first fifty to hundred readers have
completed their book. The total test duration is about six to ten weeks.
All data collected is available both in pseudonymous form (reading and survey data for individuals) and in
aggregated form (reading data for the entire focus group or sub-segments thereof). The platform is fully
The data Jellybooks collects is processed, formatted made available in real-time through an online data
portal called “Candy”. Candy is designed to be used by across the business with all colleagues. Access is
granted to any team member so designated by publisher. There is no “seat” or access charge.
There is an option or creating automated reports that are formatted to be printed or downloaded for
distribution at meetings. This feature is available through the Candy dashboard and KPI tabs.
The feedback that Jellybooks collects during a test reading campaign can sometimes provide key insights
into how readers respond to a book. These comments are private and not posted or redistributed publicly. As
a result they are very frank and informative. They provide insight into what resonates with readers or
annoys them and also allows the publisher to check if the reader making the comment has actually read the
This is a special format that can be used in combination with any of the other formats. Here we ask test
reader about the movies, TV series or Netflix series they watch (alternatively we can also ask about
magazine or newspaper reading and other entertainment formats). This a multiple-choice questionnaire
presented to test readers when they download or start reading an eBook.
We then evaluate how the reading KPIs (completion rate, satisfaction index, recommendation factor,
cover-match factor, velocity) vary depending on the user’s entertainment preferences. It is a very powerful
tool for better understanding what kind of audience is attracted to an eBook and how to position the
The test works best with popular movies and TV Shows. In fact many movies and TV shows are based on book
adaptations and make for perfect comps. Asking test readers about comparable book titles (comps) is often
less successful, because most books are read by a relatively narrow slice of test readers.
Note that in all reader analytics campaigns, we already ask if test readers have read another book by the
author previously (are they “fans” of the author or not). This “fan” question is already standard in all
tests and is not part of the cross-media format.
Readers do judge a book by its cover, and Jellybooks has the data to prove it. Traditional interview
techniques often fail when testing covers, because readers, having little context or commitment, judge the
cover primarily on its aesthetics. In contrast, Jellybooks applies a tried and tested scientific concept:
the double-blind A/B test. This allows Jellybooks to test covers without users being aware that they are
being tested on the cover.
The Jellybooks cover-match factor (CMF) measures how well readers think a cover matches the content of a
book. We also undertake A|B testing, where we split the focus group randomly into two sub-groups and give
each sub-group a different version of the cover without readers being cognizant of this. Readers are only
aware that we are collecting their reading data. They do not know about the specific variations in
“packaging” (cover, title or description), which is why the test is referred to as double-blind: readers are
blind to which group they belong to and the tester is blind to which readers get what version (random
assignment by algorithm). This ensures that any cognitive bias is minimized during the study. As part of an
A/B test, Jellybooks can also establish the pull of the cover, i.e. are readers more likely to choose the
book based on its cover. This requires the test to have a choice between at least two titles: covers matter
a lot when readers have to choose between books and select which book to read.
Between 500 and 800 test readers are recruited when executing an A/B test to ensure that statistically
valid results can be obtained. A/B tests are best suited for testing covers that are based on very different
concepts or propositions.
Follow-up Reader Surveys
For an additional fee Jellybooks can prepare post-campaign surveys that ask specific questions that arise
as a result of the campaign, which might include questions such as why users didn’t like the cover, what
they think of an alternative cover, alternate genre classifications and more. There is a per-survey fee for
the preparation and implementation of any custom-survey. There is no limit or fee regarding how many users a
survey is sent.