Parallel Exploration for Product Catalogs
What Is Parallel Exploration?
An illustrated, narrated introduction to Parallel Exploration can be viewed in a separate tab or window (not recommended for mobile devices).
As an alternative, here is a purely textual introduction:
Almost all on-line commercial product catalogs are based on the paradigm of faceted search/browsing: You enter keywords and apply filters to zoom in on sets of products that are of interest to you.
This paradigm has a major limitation: You can view only one set of results at a time; if you don’t like the current set, you have to leave the result set to try a different query. Consequently, you are likely to wander from one search result set to the next without an effective way to cope with the hundreds (or thousands) of potentially interesting products that are being offered. One frequent result is that customers hesitate to buy any product, because they believe that there must be a more suitable product somewhere in the product catalog.
Parallel Exploration extends faceted search/browsing to overcome this limitation: All of the result sets that you have viewed are stored in an exploration tree. So you can easily go back to previous result sets and view multiple result sets at once. The construct of an exploration tree is the basis for four main ways in which Parallel Exploration helps customers make up their minds when shopping.
- The exploration tree enables you to explore in a number of different directions to see where the best products can be found, without losing track of what you’ve found so far. So you can quickly try out specific queries, such as “Find related products from the same company as this product” without losing track of what you’ve found so far. Whenever you see a promising product, you can put it on the comparison table.
- At any time, you can view the products in the comparison table, viewing them side by side with different levels of detail and adding your own annotations.
- Until you’re satisfied with one or more products, you can go back to the exploration tree – for example, to the search result set in which you found an especially promising product – to continue gathering more suitable products.
- At any time, you can create a bookmark that saves the current state of your exploration, so that you can come back to it later – or share it with others to get their input. Among other advantages, the bookmark functionality enables groups of people who are jointly deciding about a purchase to exchange ideas and results to converge on a decision.
Together, these features of Parallel Exploration enable customers to take control of the process of choosing a product from a large set, exploiting and increasing their own knowledge - as opposed to having to choose passively from sets of recommended or promoted items that may not come close to meeting their current needs.
How Has Parallel Exploration Been Applied So Far?
Parallel Exploration was developed over several years at the German Research Center for Artificial Intelligence (DFKI) in the group of Chusable founder Anthony Jameson. In the domain of city events and places, a much earlier version played a key role in the innovation activity 3cixty, funded by EIT Digital, which won the 2015 Semantic Web Challenge. At the end of 2016, DFKI transferred the intellectual property rights to Parallel Exploration to Chusable so that it could be commercially exploited.
How Is Parallel Exploration Being Deployed by Chusable?
A number of important on-line vendors make it possible for affiliates to design websites that access their product catalogs. As a first step of exploitation, Chusable has been developing and testing a website based on Parallel Exploration as an affiliate of one of these vendors.
Detailed information and demonstrations are available on request.
Chusapedia: Analysing and Designing Choice-Supporting Interventions
Teilnehmer am MOOC “Smart Service Welt”: Beachten Sie bitte auch die für Sie bestimmten Hinweise auf der Hauptseite von Chusapedia.
Who Can Benefit From Using Chusapedia?
Chusapedia is used by specialists who design choice-supporting interventions, such as:
- a website for helping customers choose products on-line
- a small part of a mobile application that helps users choose a method to perform a particular task
- a multi-month, multi-person behavior change program to encourage people to eat healthier food
More generally, a choice-supporting intervention is any set of measures taken to help people make choices in a particular context.
Why Do We Need Something Like Chusapedia?
Analysing an existing intervention – or designing a new or improved one – requires the application of a wide range of knowledge about how people make choices and how they can be helped to make better choices.
This knowledge is normally available only in the heads of experienced practitioners and in a wide range of publications based on research in psychology, computer science, and other fields. The relevant experts tend to use different conceptual frameworks and to adopt complementary perspectives.
The web application Chusapedia makes available a coherent synthesis of this wide range of knowledge and enables analysts to apply the knowledge effectively.
What Can I Do With Chusapedia?
Specifically, Chusapedia helps the analyst to answer the following questions in a systematic way while making use of relevant general and specific knowledge:
- Who are the choosers whose choices are being supported?
- What choices of each chooser need to be supported?
- What are likely steps that the choosers will take when making each choice – with or without an intervention?
- What known strategies and tactics can be applied to support these choice steps?
- What features of the intervention help to realize these tactics?
Answering these questions with Chusapedia helps to understand the gaps in any existing intervention and to generate well-founded ideas for a better intervention.
What Knowledge Is Exploited by Chusapedia?
For the analysis of any given intervention, Chusapedia makes use of knowledge encoded in an underlying model of how people make choices and how choices can be supported. The default model is based on the ASPECT and ARCADE models from the book Choice Architecture for Human-Computer Interaction. It is possible to add extensions that deal in more detail with:
- choices in particular domains (e.g., health and well-being)
- particular types of choice problem (e.g., configuration problems)
- ways of using particular technologies (e.g., recommender systems) to support choice
It is also possible to introduce models based on different conceptual frameworks (e.g., from the persuasive technology or behavior change fields). So people accustomed to using these conceptual frameworks can benefit from the functionality of Chusapedia.
Models can also be formulated that have little or nothing to do with choice support – for example, a model of risk factors and possible mitigating strategies in a given domain. Models like these make it possible to apply the general user interface functionality of Chusapedia to a wide variety of tasks.
The development and use of models other than the default model currently requires the collaboration of the Chusable team; but support for the independent use and development of alternative models is being added step by step during the second quarter of 2018.
How Can I Gain Access to Chusapedia?
Chusapedia will remain freely available for use in both practice and research at http://chusapedia.chusable.com .
Stored demos that quickly introduce the main concepts and functionality can be accessed via the main menu.
Even as Chusapedia continues to be developed, Chusable offers the service of helping clients apply Chusapedia efficiently to improve their own choice-supporting interventions. Part of this work can include the creation of new underlying models that are especially relevant to the client.
Where Has Chusapedia Been Presented Publicly?
Successive versions of Chusapedia have been presented and made available in a number of scientific or business-oriented talks, including the following:
- Two workshops at the 2017 International Conference on Recommender Systems (Como, Italy, August 2017).
- A keynote talk at the conference i-KNOW 2017 (Graz, Austria, October 2017), from which the annotated slides are available.
- The international workshop New Trends in User Expertise and Interactive Systems (Paris, France, October 2017)
- The German-language MOOC Smart Service Welt, Daten- und Plattformbasierte Geschäftsmodelle, which began on April 23rd, 2018, and for which (free) registration is still possible. (The final lecture of Week 2 is directly relevant to Chusapedia.)
- An English-language version of the same MOOC, titled Smart Service Welt – Data and Platform-Based Business Models, which will begin on June 19th, 2018 and which is already available for (free) registration.
Presentations to smaller groups, focusing on content especially relevant to the group in question, have also been made and can be arranged on request.