my research

Right now my work focuses mainly on quantitative research, large-scale analytics, and helping Intercom make informed business decisions. However, up until recently I devoted the vast majority of my research career (12+ years) exploring the intricacies of mobile user behavior, in particular focusing on mobile search, mobile app and mobile device usage patterns. I explored how people access and consume mobile content, the types of daily information needs mobile users encounter, the effect of contextual cues on mobile behaviors as well as prototyping and evaluating proactive mobile and social applications that aim to enrich their experiences. My goal was always to use the insights from my research to propose, to prototype or to evaluate new engaging mobile experiences. And at Yahoo I had the chance to work directly with mobile product teams on such endeavors.

My research interests include: the mobile web and mobile search space; mobile-app and mobile device usage; personalization and social applications in a mobile setting; mobile HCI; context-aware computing, location-based computing; mobile user experience; and mobile user behavior modeling.

I’ve experience in designing and carrying out live mobile field studies, most of which tend to combine mixed-method approaches in which qualitative insights are gathered using techniques like interviews, surveys, refined Experience Sampling Method and novel diary studies, and quantitative insights are gathered by logging actual mobile usage normally via custom built logging software. I’ve also experience in rapid mobile prototyping and development (mostly Java) — ranging from low-fidelity paper prototypes to proof-of-concept mobile prototypes – normally on iPhone or Android handsets. Finally I have experience in analyzing and mining large-scale mobile user behavioral patterns using technologies including Hadoop, PIG, Hive, SQL and R. My research has been pretty varied and involved elements of design, prototyping/development, data mining & data analysis and human-computer interaction (HCI).

Below is a list of the research themes/areas I worked on up until 2016 with a list of projects or research studies in each area.

#1 Mobile Messaging & Notifications

An In-Situ Study of Mobile Phone Notifications
With Martin Pielot (lead) and Rodrigo de Oliveira while @ Telefonica Research

A mixed-method study conducted in 2013 with Martin Pielot (project lead) and Rodrigo de Oliveira. The study involved 15 participants in which we collected real-world notifications through an Android logging application alongside subjective perceptions of those notifications through an online diary. We found that participants received an average of 63.5 notifications per day, mostly from messengers and email. And these notifications were viewed within minutes despite the phone being on silent mode or not. A paper describing the study design, insights and implications was presented at Mobile HCI 2014 and won the best paper award!

Comparing Mobile Instant Messaging Behaviors with Traditional SMS
With Rodrigo de Oliveria while @ Telefonica Research

In this study we combined in-depth interviews with 9 active WhatsApp users along with a large-scale survey with 131 participants to understand more about the motives and perceptions of using mobile instant messaging over more traditional SMS. Our insights point to the perceived benefits of WhatsApp — such as cost, sense of community and immediacy — while SMS was still considered a more reliable, privacy preserving technology for mobile communication. A paper describing these studies and our insights was presented at Mobile HCI 2013. This paper was nominated for a best paper award!

 

#2 Understanding Mobile Device Use

In-Home Study: Quantifying Mobile Device use while Watching Television
With Christian Holz, Frank Bentley and Mitesh Patel @ Yahoo Labs

A mixed-method study conducted in Winter 2014 involving 7 participants in 7 diverse households. We set out to understand more about mobile device use in and around the home as well as in conjunction with watching TV. In order to understand this use we equipped a primary participants mobile phone and tablet with Android logging software which tracked all the apps the participant used and when. We tracked indoor location using 4 BLE beacons around participants’ homes, primarily in the living area, kitchen, and bedroom. Finally, we placed a dedicated phone that ran sound-printing software by participants’ main television sets to detect and log which TV programs our participants watched. To understand more about the motivations of using devices in and around the home, and in particular while watching TV we also conducted interviews at the beginning and end of the 2-week study period. Participants were also asked to leave a voicemail message once per day summarizing how they used their smartphone and tablet that day. A paper focusing on understanding more about how and why people use their mobile devices at home and while watching TV has been accepted for publication at TvX 2015 :-)

Behaviors and Motivations behind Mobile Phone Personalization
@ Yahoo Labs. Qualitative study done with JP Carrascal (during his internship)

In early 2014 Yahoo acquired a startup called Aviate. Aviate is an intelligent homescreen launcher for Android phones. Throughout 2014 we conducted some research with the Aviate team to explore a few questions around mobile phone personalization. In particular looking at how and why people choose to arrange apps on their phone, how they find and install new apps, how they decide which apps to keep, etc. These research questions were explored in two ways.

Firstly we conducted a large-scale study using a large sample of Aviate users (~500K), which detailed what apps people had on their phone and where these apps were located/arranged on their phone (i.e. on a screen, in a folder), before the user chose to install and use the Aviate launcher. Our goal was to learn more about exactly how people, Android users specifically, arrange their apps. We then conducted a second, small scale qualitative study in which we interviewed 20 participants and asked them detailed questions are their motivations and perceptions of personalizing their smartphones, arranging and installing apps on their smartphones, etc.

The ultimate goal of all this work was to add to our fundamental understanding of how smartphone users engage with and personalize their mobile phones and to derive insights that could be used to inform the design of future homescreen launchers or similar mobile device experiences.

Understanding App Usage at Scale
@ Yahoo Labs

Also working with Aviate, I conducted a range of analyses throughout 2014, which explored how people use apps on their mobile phone. In particular looking at which apps people use, the categories of apps used, the frequency of use, temporal patterns of use, location patterns, etc. What’s great about these analyses is that we’ve worked with large samples of mobile users (i.e. 100-500K). A subset of these analyses and the associated results has been publicized on both Yahoo Labs and Aviate Tumblr pages. And a fantastic resulting infographic of the results was produced by the Aviate design teams – click on the thumbnail to the right or here to see the infographic. We found for example that despite having 100 apps installed on their phones, mobile users typically engage with just a handful of those apps on average each day.

How Teenagers use their Smartphones?
With Frank Bentley, Kent Lyons, Beverly Harrison and Matt Rafalow (during his internship) @ Yahoo Labs

One of the first projects I worked on when I joined Yahoo Labs was a study of how teenagers used their smartphones. This was a very, very cool project! We conducted a 2-week mixed-method study with 14 teens, 7 males, 7 females and an even split of Android and iOS users in Summer/Autumn 2013. We equipped the teens phones with custom-built software which logged all apps the teen used, along with timestamp info and duration of use details. We conducted in-depth interviews at the beginning and end of the study period. And finally our teens called a voicemail number twice daily to tell us more about their device use and events for that given day.

We found some really interesting insights about how and why teenagers engaged with certain apps. It was also an eye opening experience working with teenagers! To deal with the qualitative data we conducted one of my first fully-fledged affinity diagrams which took a few weeks. The thumbnail photo below shows proof of what we deemed “the war room” where we tried to make sense of the wealth of qualitative insights we collected. I worked primarily on the quantitative insights and found insights like: our teenagers engaged with 9 different communications apps on average and spend 3 hours per day interacting with their phones!!!

#3 Mobile Web & Mobile Search

An In-Situ Study of Mobile App & Mobile Search Interactions
With JP Carrascal (Intern) @ Yahoo Labs

We conducted a study in summer/autumn 2014 to understand the interaction between mobile search and mobile app usage. The motivation behind this work is that smartphone users spend much of their time transitioning between mobile search engines and mobile apps (and vice versa) when trying to find information. To shed some light on this interesting behavior we designed and conducted a 2-week, mixed-method study involving 18 Android users in the Bay Area. We collected real-world mobile search and mobile app usage data via custom software installed on participants phones. Rich subjective insights on why certain interactions between apps and mobile search occur were collected through in-depth interviews at the start and end of the study as well as via a daily online diary. We found that when smartphone users engage with mobile search, they engage more intensely with other mobile apps.

We found that certain categories of apps are used more intensely alongside mobile search. Furthermore we found differences in app usage before and after mobile search and show how mobile app interactions can both prompt mobile search and enable users to take action. Full details can be found in our CHI 2015 paper available here

Understanding Requirements of Place in Local Search
With Henriette Cramer @ Yahoo Labs

In Winter 2013, Henriette Cramer and I designed and deployed a large-scale survey via Mechanical Turk to shed light on what factors of location are important when choosing a target place of interest in a local search scenario, namely restaurant search. The survey was deployed in December 2013 and collected a range of demographic information as well as details regarding participants’ current and most recent locations. The core of the survey was a local restaurant search scenario in which participants were asked to provide freeform textual responses to the following question: “Imagine you are searching for a place to eat. It will be tonight, and you will be with a good friend. What requirements are important for you in terms of where the restaurant is?”. We received responses from 306 participants, 145 female (47%), 160 male (52%) and used thematic analysis to find key themes/insights. We identified 4 key factors that influence choice of location in a shared restaurant search scenario including proximity and getting there, temporal / seasonal effects as well as social / shared location preferences and constraints. We recently had a CHI 2015 WIP paper accepted which can be found here.

Exploring the Social Side to Mobile Search
With Sofia Reis (Intern) and Antony Cousin (Intern) @ Telefonica Research

While at Telefonica Research I was involved in a range of studies which aimed to explore the social side to mobile space. The motivation behind this research is work in mobile computing has highlighted that conversations and social interactions have a significant impact on mobile Web and mobile search behaviors. However, to date, this social component has not been fully explored and little is known about why and how mobile users search for information in social settings. A deeper understanding of this social dynamic will help improve future mobile search experiences. So we focused our efforts on understanding (1) how groups of users engage and collaborate in mobile searches whilst in co-located group settings and (2) what types of novels interfaces, applications and interactions we could provide to enrich their experiences.

We published a study aimed at understanding why and how people use mobile search in social settings at MobileHCI 2012 in San Francisco. This paper reported the results of 2 studies: a survey involving almost 200 users and a two-week diary and follow-up interview study of 20 users. Our results point to the motivations, circumstances and experiences of using mobile search in social settings to satisfy group information needs.

More recently we focused our efforts on a proof-of-concept mobile application designed to enhance such social experiences by providing an easy means of interacting with and sharing mobile Web content among co-located groups. We also conducted an exploratory field study of the prototype and found a number of design implications that could enhance next-generation social mobile services. The application is called WaggleBee (screenshots can be seen by clicking on the thumbnail). A WiP paper was presented at CHI 2013 and a work-in-progress paper was published at the Searching 4 Fun Workshop as part of ECIR 2012.

I have also contributed an essay related to this topic to a book published in Jan 2013 called Designing the Search Experience by Tony Russell-Rose and Tyler Tate. The essay contribution is entitled Exploring the Social Side to Mobile Search.

Exploring Proactive Mobile Search Interfaces
With Mauro Cherubini, Joachim Neumann and Nuria Oliver @ Telefonica Research

Social Search Browser (SSB) is a novel, proof-of-concept research prototype designed to enhance the search and information discovery experience of mobile users. The application proactively displays the queries and interactions of other users in a given physical location. The application incorporates social networking capabilities with key mobile contexts such as location and time to encourage discovery of new, interesting content while on-the-move. The prototype also taps into the social dimension to search and information access by allowing friends and other users to answer your queries while you are on-the-move.

We carried out two live field studies of the prototype, both of which yielded very interesting results. The results of these field studies have enabled us to outline a number of important implications in the design of future mobile information access applications of this nature. And we’ve successfully published two papers describing the prototypes and the field studies, one at IUI 2010, a second at WWW 2010. See the publications page for more details.

Patterns of Mobile Web Use at Scale
With various collaborators @ UCD and @ Telefonica Research


Over the years I’ve conducted a number of studies related to understanding mobile search and mobile Web patterns. Most of these projects involve large-scale, offline log analysis of mobile search and browsing patterns with a view to understanding evolving patterns of behavior. My research in this space started as part of my PhD, in which I had the unique opportunity of work with a dataset from a mobile operator in Europe. The first dataset I used included a collection of more than 30 million mobile Internet requests generated by just over 600,000 unique European mobile subscribers over a 24-hour period in late 2005.

A subsequent dataset from the same operator focused primarily on mobile search and involved over 6 million individual search requests generated by over 260,000 individual mobile searchers over a 7-day period during 2006. While these numbers might seem to pale in comparison with operator datasets of today, at they time processing, analyzing and interpreting such data was no mean feat!

When I joined Telefonica I worked with their datasets, analyzing the search queries submitted through Telefonica / Movistar mobile portals to see how users in Spain engage with operator portals. All of these analyses looked at how people engaged in mobile search, focusing on things like query length, query distributions, query variation, query topics, etc. Results from these works have been published in a Research in the Large Workshop at UbiComp 2011, Mobile HCI 2008 and Transactions on the Web (TWEB) 2007.

In addition I’ve gave an invited industry talk at ECIR 2013 on the topic of mobile search in which I took a nostalgic look back at research in the mobile search space between 2005-2013. And many of the insights from these studies were included in this talk. The slides from this talk are available here and for those interested, the talk was actually recorded and can be watched online here.

Understanding the Intents and Motivations of Mobile Web & Mobile Search
With Nuria Oliver @ Telefonica Research

To compliment my studies of mobile search and mobile web use at scale, I’ve conducted a range of ethnographic field studies to investigate what motivates people to use mobile search, the impact of context on search behaviors and novel success / failure metrics for proactive mobile search experiences. I’ve also focused on temporal patterns, the influence of location and triggers for mobile search. Most of these studies involve diary studies, interviews and/or surveys and most of these studies where conducted while I was at Telefonica Research. These studies have been published at published at Mobile HCI 2011 and Mobile HCI 2012. See the publications page for more details.

Mobile Information Needs

A Large-Scale Study of Daily Information Needs
With Mauro Cherubini and Nuria Oliver @ Telefonica Research

Taking a step back from how people engage with their devices, use apps or access/consume online content via their mobile phones, we’ve taken an in-depth look at the underlying information needs of mobile users. The most recent of these studies has been one the most comprehensive studies of daily information needs to date, spanning a 3-month period and involving over 100 users in Spain. The study employed an intelligent experience sampling algorithm, SMS technology and an on-line Web diary to gather insights into the types of needs that occur from day to day, how those needs are addressed and how contextual and demographic factors impact on those needs. Participants were probed 3 times per day via SMS asking them about their information needs. Participants could respond via SMS and send textual snippets of their information needs. Participants then accessed an online diary at a later, more convenient stage to review all daily information needs snippets. Using an online diary , participants could provide a range of additional details about their needs like where they were located at the time the need arose, who they were with, if they satisfied their need, how they satisfied their need, etc. This study resulted in almost 12,000 SMS snippets and over 9,500 associated diary entries. Findings from this study were published in ACM Transactions on Computer-Human Interaction (TOCHI) in 2014 and we had the opportunity to present the work at CHI 2014.

Understanding Mobile Information Needs
With Barry Smyth @ UCD

In 2007, as part of my PhD, I conducted a diary study of mobile information needs. The goal of this study was to understand the intent behind mobile information needs and how those information needs change based on context. We compared user needs while at home, at work and while on-the-go. The study was in the form of a paper based diary study in which participants were asked to carry around a small notebook which we supplied and to use this notebook to keep a diary of all their information needs while they were at home, at work or mobile. We define mobile as being away from their own personal home or being away from their desk at work. We had 20 participants take park over the 4-week study period.

Our findings indicated that when users are mobile there is a clear location and temporal dependency in their information needs. When we examined user goals, we found that traditional Web taxonomies such as navigational and transactional needs were non-existent among the diary entries, thus requiring the addition of two new taxonomies that capture the unique constraints of mobility, specifically Geographical and Personal Information Management (PIM). Details of this study were published at IUI 2009.