Ideas Like This identifies similar ideas within an event. Large innovation events can easily generate hundreds of ideas. Careful reading of such ideas can be difficult because the topic jumps around for every new idea; it is not easy to keep focused, or to know when a particular topic has been thoroughly explored. Ideas Like This finds all ideas similar to the one youʼre looking at, and presents them in ranked order for easy reading without page-jumping or interruption. Ideas like this is suited for looking at single ideas, where Triage is suited for looking at ideas en masse.
Business Value
Large-scale idea generation is a now a well-established business practice thanks to Imaginatikʼs historic leadership and offerings from many vendors. But the essential next steps — serious yet efficient reading, evaluation, triage, and review of hundreds of ideas — is not yet well addressed. Ideas Like This is one tool that can help.
The first and best way to manage innovation event content is by framing the event carefully in the first place. This means clear and named sponsorship, honest and measurable goals, explicit statements of business need and criteria for evaluation, and (in most situations) a finite time when ideas and comments are welcome, followed by an explicit evaluation and review process. But even when all of that is done well, large events, especially those run swiftly, will generate hundreds of ideas with inevitable repetition and overlap that even the best descriptive menu choices cannot really categorize. Sorting by author or date isnʼt effective, and menu-based categorization may help but is always broader than a particular train of thought that emerges.
This is the need that Ideas Like This addresses: when youʼre reading an idea that you find particularly interesting, with one click (and no struggling to guess which might be good keywords for searching), Ideas Like This provides you with a list of all similar ideas in the event, in most- to-least similar order, and with the full text of these ideas in one place for easy reading.
Ideas Like This is accessible to all users, from casual browsers to time-pressed review team members.
How Ideas Like This Works
Searching and classification of text is a huge topic, from academics who want to decide if an anonymous poem is more likely by Shakespeare or Marlowe, to corporations like Google or Microsoft who profit by creating comprehensive and accurate search engines. But in our research weʼve found that most of these methods arenʼt particularly useful, because weʼre analyzing ideas of only 20 to 100 words within a context of only 100 to 1000 other ideas. These numbers are too small for the statistical techniques that would benefit Google, and content too limited and diverse for the parsing of grammar and meaning that would benefit the academic.
Instead, Ideas Like This takes its core technology from chemical information systems, which is a highly accurate and mature field. Ideas Like This uses a similarity score that is fast and effective. These similarity scores are used to put the similar ideas in order (most similar to your “central idea” at the top), and also to create an interactive graphic display of pushpins:
The central idea – the one youʼre interested in, for which you want to know if there are any related ideas – is shown with a starred pin at the left. Other ideas are shown as colored pushpins. Any pin thatʼs close to the star is very, very similar: in the example above, the idea represented by the red pin is so close that it probably represents a duplicate copy-and-paste entry. The orange pin is at a middle distance; it has some legitimate similarities. The others are really quite far away, and while there might be a few common elements, they are most likely distinct from the starred idea. At least two pushpins are always displayed; if these are far to the right, then you can be very confident that your idea of interest is unique in this event.
In our extensive testing we find that about half of the ideas identified by Ideas Like This will feel similar to the central idea. On the other hand, if there are any similar ideas, they will be found. Your experience with Google is probably similar, namely that several of the returned links on the first page will be useful to you, but if there are none, then you are confident that what you sought just doesnʼt exist.
Advanced Controls
What we in Imaginatik Research think is “similar” wonʼt always be what you want. Are two ideas similar if they use the same words, which are otherwise rare? Probably. Are they similar if theyʼre by the same person? To some extent. Are they similar if they were written at about the same time? No, but I might want that for another reason.
So there is an “advanced controls” link that opens up a control panel, shown here. It's not necessary for regular use, but it can be fun to explore.
At the top, the advanced controls give you three preset definitions of what “similar” might mean. The first is the default, which emphasizes similar language, authorship, and any shared categories (such as Challenge, Area, or other menu-driven choice in the ideas). The second checkbox will bias strongly for content but puts a negative weight on authorship, that is, it will find similar ideas by different people (if they exist). The third checkbox does the opposite, and finds different content by the same author as your central idea.
In the center of the control panel you can adjust the weights of the five factors manually, so if you want to see ideas entered at the same time, push the date slider to the right and all of the others to the left. It’s a general-purpose, configurable search engine that works in real time.