R&DTestapic
Testapic innovates constantly to ensure that its user-centric optimization methodology continues to perform effectively.
With this in mind, Testapic pays particularly close attention to the “state of the art” applicable to all the tools and methods that it uses, ensuring that our solution remains valid, reliable, and competitive.
Pre-processing of user test videos: standardizing audio quality
Whenever a user completes a video test, the quality of the sound recording depends directly on the hardware used for the test and the external conditions of the recording. Consequently, there is no way to standardize sound quality, which is consistent for all users, as soon as the recording completes.
Our efforts are currently focused on achieving three goals:
- la Standardizing the audio/video encoding format across all platforms;
- la Normalizing the playback volume to avoid having to readjust the volume button constantly while viewing the recording;
- la Detecting and moderating “parasitic” peaks in volume, which detract from the listening experience, cause spikes in volume levels and give incorrect results when performing calculations to normalize the playback volume
Standardizing the encoding format
Settling on a single encoding standard is relatively simple, and relies on existing Open Source software.
Videos that are re-encoded in this way ensure full compatibility with the majority of video players and makes the process of tracking bugs much quicker than when a video suffers degradation during the recording process.
Normalizing playback volume
The idea is to apply a normalization factor to the entire waveform to ensure that the playback volume is set to a reasonable and consistent level.
This method nevertheless has a limitation: if an irregular, highamplitude peak occurs in the waveform, it will act as a point of reference, leading to bias in the normalization process.
Detecting and moderating peaks in volume
To remedy the normalization issue, we have developed and algorithm that can detect and automatically moderate any “parasitic” spikes in volume.
These optimizations mean there is no longer any need to adjust the playback volume. Users can therefore focus on their analytical work exclusively.
Enriching user testing videos with speech-to-text
The first work on machine learning and deep learning algorithms was undertaken in the 1950s. These algorithms have grown significantly over the last decade and are now available to developers via software libraries that are relatively simple to use.
Our data/R&D team is currently hard at work to create a remote user testing experience that is enriched by the latest innovations from the field of artificial intelligence.
In the near future, our clients and UX experts will therefore be able to benefit from automatic closed captioning of test videos, synchronizing the text with the video footage. Meanwhile, the transcriptions will serve as the basis to feed natural language processing algorithms, developed in-house by our data scientists, all with a view to automatically enriching our test data.
As such, it will for example be possible to identify the major blockers in our clients’ interfaces and applications even before starting the “human” assessment phase, to filter out irrelevant content and, eventually, to automate some steps in the analytical phase.
Internet users’ expectations and behaviors
Testapic’s R&D department also conducts a series of research programs, both independently and in partnership with renowned French institutions including CNRS, BPI, and CFI.
In the context of studying internet use, Testapic has undertaken a process of iterative research to study internet users’ behaviors and expectations. This research relates to a series of topics that are both fundamental in nature and highly varied, such as customization, filtering functionality, and behavioral differences between men and women. The aim of this research is to overcome potential failures that are likely to cause users to abandon their browsing. These studies also aim to highlight best practices that encourage a positive user experience..
This research process enables Testapic to be close to its users and listen to what they have to say (see feedback from our users) while remaining close to our partners by providing them with relevant, specific recommendations that aim to improve their user interfaces and features (see feedback from our customers).
As part of our continuous improvement process, our R&D department also constantly monitors the latest developments in fields such as,UX, usability, and the mobile web…
Testapic’s R&D activities have earned it the Scientipôle Initiative Certification in 2012, a place at the finals in the 2013 Bourse de l’Innovation awards and certification as an authorized organization for the Research Tax Credit in 2014.
Behavioral changes via the mobile web
Against a backdrop in which the technologies used to access the web are changing, the current performance of smartphones and tablets has considerably changed the behaviors of web users and the conventions of web design. User-friendliness is at the heart of this “mobile revolution” in terms of the way that information is both accessed and customized.
Building on sustained reflection derived from lessons learned during its previous activities, Testapic has undertaken a major research project into the effect of mobile platforms on web use. The issues that it seeks to address are derived in part from questions that arise relating to different usage patterns on mobile and non-mobile platforms.
Does web usage differ depending on whether a mobile platform is used?
This is the major question from which a number of related questions flow, in order to address the specific nature of services offered via websites – i.e., are web services used differently as a result of the mobile factor – as well as questions relating to the nature of the users themselves – i.e., is the way that the mobile web is used derived from the profile of its users?
In order to shed light on the topic, Testapic carried out a study of nearly 400 web users searching for real estate online. The aim of this study was to capture their use of the web, and their behavior, when using websites of real estate developers and real estate listings sites in both desktop and mobile formats. The results showed that users browse the same site in radically different ways, depending on the platform that they use (desktop vs. mobile web).
While traditional web platforms are more heavily used to carry out research and to compare different real estate schemes, benefiting from the large format of desktop screens, mobile platforms are primarily used for practical and utilitarian purposes, such as searching the address or phone number of a real estate agency. The issue at stake in this research is how users’ interactions with different web interfaces can be characterized, depending firstly on the interface type, and secondly on the platform used (desktop or mobile). Over time, Testapic set itself the goal of developing a set of “user/interface interactions” to serve as a reference point in classifying user behavior online.
With this in mind, Testapic’s research aims to:
- Identify user behaviors related to the “mobile web” and to distinguish them from behaviors associated with desktop platforms;
- Differentiate users’ behaviors and habits depending on the type of website and the services and features offered via the particular interface;
- Classify different interactions based on desktop and mobile platforms and website type.
Ultimately, the development of this body of reference material should provide website publishers with a robust and context-specific set of guidelines and recommendations to support the development of user interfaces.
Our R&D Team
Three experts in their fields who complete research projects that are linked to our business.
