We have some important news to announce today - WHAT Conference 2018 is no longer taking place. We had a great time last year and would like to thank everybody who showed their support and joined us at WHAT 2017.

For those of you who had already bought tickets, we appreciate your enthusiasm for #WHAT18. Of course, you will receive a full refund for your tickets.

Instead of being sad about saying goodbye, we wanted to share the colourful vision we had for WHAT 2018. We hope you like its lively spirit!

If you have any questions, feel free to reach out to us.

May the data be with you,

Your WHAT Conference team.

beatery

On October 25th, respondi launched beatery, a new interactive dashboard which connects declarative and behavioral data to get a more exact, speedier and simple insight into people and markets.

respondi CEO Dr. Otto Helwig said of their latest product that "...with beatery, respondi provides an intuitive and interactive ticket to the data-driven analysis of people and markets.”

By using the latest best-in-class technologies and analysis methods, beatery offers a customized dashboard to help clients accompany customers in their everyday lives reliably and insightfully. beatery tracks the use of digital devices of representatively selected people, combines this information interactively and visually with survey data and supports the analysis of this new dimension of granularity.

Reppublika

Reppublika is a new tool for media reach and real time campaign analysis based on passive metering, launched by MindTake earlier this year.

Reppublika's data is steered by tracking user behavior. The dashboard combines a range of different data sources, including MindTake's existing deep profiling data, the passive metering behavioral data from Wakoopa as well as other sources, cookie data, statistical knowledge from market research specialists, and more. Reppublika can rank every online service, cross-device, according to usage and reach. It can furthermore analyse the flows of traffic between websites, and drill down to the demographic profile of each site’s users on some 40 attributes such as gender, age, education, family status, even car or pet ownership.

We interviewed MindTake founder Klaus Oberecker about his new product earlier this month on our blog. You can find the full interview here.

As proponents of passive metering, we are happy to see the uptake of passive metering as an integrated part of the offering of panel providers. We look forward to seeing this adoption and innovation in behavioral data continue to grow with our clients and the overall market research industry.

ENGAGE – 101 tips to improve the research participant user experience!

Over the last 12 months, GRBN has worked on the participant engagement initiative, conducting research-on-research, analyzing metrics and listening to expert opinion. As we all depend on people’s willingness to participate in research, the initiative aims to bring clients, agencies, technology providers and data collectors together to improve participants' research experiences.

The new handbook is full of practical advice to help researchers deliver better experiences to research participants and greater value to clients. ENGAGE contains 10 golden tips to increase participant engagement as well as specific tips to online communities, online surveys, passive metering, and qualitative research.

You can access the full handbook and read our passive metering tips here.

About GRBN

The Global Research Business Network connects 45 research associations and over 3500 research businesses on six continents. More than US$25 billion in annual research revenues (turnover) are generated by these businesses. GRBN’s mission is to promote and advance the business of research by developing and supporting strong autonomous national research associations and undertaking global industry initiatives.

Klaus Oberecker founded MindTake Research in 2001 while still a student at the University of Vienna. What started as a two-man web agency evolved into a group of companies that cover the digital spectrum, offering expertise in panel and online market research in 21 countries, as well as design and development services for external clients such as museums, banks and media portals.

Interview with Klaus Oberecker, Founder at Reppublika, MindTake Research & Talk Online Panel

In early 2016 MindTake decided to get a license for Wakoopa’s passive metering technology. Why do you think passive metering is important, and what were the main reasons for you to decide to implement passive metering?

MindTake Research, our market research agency, is now 15 years old, and while this makes us a bit of a veteran, we’ve always been passionate about digital innovation – curiosity and excitement about technology is in our DNA. When we learned about passive metering tools and Wakoopa, we decided to get a license on a whim, without really having a concrete plan: we were just convinced this is the way of the future, and we thought it was cool to be able to get to all this data. Basically in the beginning we just followed our hunch, and built on a plan from there.

How have you incorporated passive metering in your research designs and data collection methods? How do you combine behavioral data with other data sources?

Initially we struggled a bit with finding the best way to monetise this new data source. Of course we pitched behavioral data to existing and new customers: for MindTake as a market research agency the obvious services using this kind of data were customer journeys, search term analysis, competition monitoring, trend watch, things like that. Our experience in the panel business with Talk Online Panel helped a lot with implementing and processing the passive data automatically. In this process we also enrich the behavioral data with both our own panel profile data, as well as data provided by our clients.

How has your experience with offering passive metering to existing and new clients been so far?

Market research is an established discipline with a tried and tested set of methods, so naturally, as with any new technology, you must first convince your clients of the benefits. But since we have a solid reputation as a "digital native" market research agency, there is a lot of trust on the client side and most are willing to give our suggestions a chance. It’s of course tricky with new terms like “passive metering” - no client knew or understood it at the beginning! But it all comes down to the fact that everyone wants to know more about their customers, target groups, and what they are doing online, and passive metering data is simply the best way to find out.

What do you see as the main challenges and/or key success factors when dealing with passive metering in your specific markets?

Speaking as a panel provider, the challenges for selling behavioral data are quite big, since no ordinary client can handle “raw, behavioral data”. Without pre-processing and “cleaning” it, this kind of data is of little use to ordinary research agencies or direct clients. Speaking as a research agency, the situation was different: offering the bespoke services to market research clients has been successful because we come from digital and know how to handle data. So from the perspective of MindTake that was successful, but came with one catch: All projects were different and required a new audience that we had to recruit and convince to install the tracking software, every time. This is hard work, and soon we noticed that letting the users go after each project was more expensive than simply keeping them. At that point we decided to recruit a bigger, representative and permanent panel and just keep collecting data.
That meant much higher maintenance and recruitment costs, so a permanent monetisation strategy was needed to at least cover the base costs. We saw this as a challenge and integrated the data into Reppublika, a SaaS (Software-as-a-Service) product that we had already started to develop previously.

Reppublika is a new tool for media reach and real-time campaign analysis based on passive metering. How does Reppublika apply behavioral data and what are the use cases for advertisers?

In a way Reppublika started as our approach to the challenges mentioned before, but it developed into much more than that. The module "Reppublika Ratings+" uses behavioral data based on Wakoopa, on other sources and on our deep user profiling data, and brings it all together to rank and rate all used services for a national market, be it apps or web sites. Advertisers can filter and search through it using a variety of criteria based on the usage data and a big range of demographic attributes. They can find target groups and decide where to place advertisements, and brands can also use it to monitor their competitors – everyone likes that function! If you, for example, wanted to know how the group of “20-29-year-old Austrian women who own a cat” spend their time online, you could find out. Not to keep you in suspense, in August 2017 they were mostly on Netflix.
In addition to Ratings+ we also have other modules such as "Reppublika Campaign Control", which is a control centre for monitoring and optimising online campaigns. It measures impressions and target group accuracy, and utilises automated brand-lift surveys to establish the effectiveness of your ad based on predefined Key Performance Indicators such as Recall, Recognition, Appeal and Purchase Probability. Campaigns can be tweaked and optimised in real time.
To arrive at this result we had to use our existing deep profiling data, the passive metering behavioral data from Wakoopa as well as other sources, cookie data, statistical knowledge from our market research specialists, our in-house team of developers, and the business relationships we have with advertisers, media agencies and brands all over Europe. It is a huge undertaking, but industry reception has been overwhelmingly positive.

What would be your advice to the industry to survive in the future of research?

Always keep an eye on emerging technologies and don’t be afraid to experiment and adapt them to your needs – no one ever survived in business by going back to the ‘good old days’.

In this part, we will focus on the experience users have with passive metering and surveys. We asked our respondents a variety of questions about different aspects of both data collection methods in order to get an idea of how their experience can be improved and how we can prevent a further decline in response rates.

Research results

We started the survey with a very broad question in order to get a sense of people's general motivations for participating in market research.

Why do you participate in market research?

Note: Multiple answers were possible

What we found in the answers is that incentives are still the overwhelming reason for people to participate. However, many people also indicate that they think their opinions and behaviors can have an impact on society, by giving them a voice.

The panelist were also asked to rate their experience with both surveys and passive metering, giving them a grade from 1 to 5.

How would you rate your experience with passive metering/surveys?

Although the results were close, more than twice as many people rated passive metering higher than surveys.

The participants were then asked follow up questions about their preferences to specify why they prefer one method over the other.

Why do you prefer surveys/passive metering?

For people who prefer surveys, the answers were mostly related to being able to share their opinion directly, as well as having the chance to choose the things they want to share. In the 'other' open answer category, they also indicated that they felt there were more offers of surveys than opportunities to be passively tracked, which is why they prefer surveys.

In terms of passive metering, people chose it as their preferred method because it is low effort/high reward. They indicated they feel they get better rewards with passive metering for the amount of effort and time they spend on it, and in the open answer they mentioned it was preferable because they could do it while working/continuing with their daily life. They also indicated that they can’t always remember everything they do, and so when they are asked to recount specifics in surveys, they aren’t able to answer correctly. With passive metering, they don’t have to remember their own behavior, because we can see their activity instead of having to ask for it.

The next questions aimed to understand what the participants think can be improved for both data collection methods.

How can surveys be improved?

Mostly, our panelists indicated that they wanted surveys to have better rewards and be more efficient. They also mentioned their frustration at not qualifying for enough surveys and therefore not being offered any, and being screened out during a survey.

How can passive metering be improved?

Mostly, participants felt that passive metering could have better and more rewards. Generally, they also wanted more control and clarity, such as making the tracker easier to install. In the open 'other' answer, many people suggested adding a pause or stop button to the tracker, and making it possible to turn it off for a short while if the participant wants to. However, this is already possible, suggesting there could be some improvement in how passive metering is communicated to the panelists.

Since privacy appears to be an important topic with participants who use passive metering, we decided to also ask how protected they feel in the general online environment as well as being passively measured for research purposes.

Do you feel that your privacy is protected in the general online environment?

Their answer was mostly ‘unsure’, and 23% of the people definitely did not feel safe.

Do you feel that your privacy is protected with passive metering?

When we asked the same question about passive metering, almost 2.5 times more panelists said they felt passive metering does protect their privacy compared to the ones who indicated they don't feel their privacy is protected.

When we asked the 28% who didn't feel protected why this was the case, there were some concerns generally about sharing information and private activities. Some people even said that although they didn’t feel their privacy was being protected, they also didn’t mind.

For those who felt like their privacy was protected, the most frequently chosen reason was that they felt the data was anonymous. Many participants (25%) also said they forgot the tracker was even installed, and that they understood about what the data was being used for, so they weren’t concerned about breaches in their privacy.

Interestingly, some participants indicated that they felt protected because they could turn the tracker on and off whenever they wanted. This is in contrast to those who indicated, while being asked what they would want to change about passive metering, that they would like the option to turn the tracker off.

Conclusion

From the early planning stage of selecting the data collection methods to the actual experience that people have during the execution, there are a lot of interesting take-aways from this study that can have an impact on the research results.

It's crucial to carefully choose the appropriate method for collecting data to get a better understanding of today's consumers and to reveal the insights you are looking for. Since people can’t correctly report their online behavior, passive metering will deliver more accurate data than survey data if the purpose of the study is to find out WHAT people are doing online.
A combination of both methods can also help to assess WHY people behave the way they do. The fact that people have a better experience with passive metering than with surveys, but still want to have the option to express their full opinion, can also be addressed by combining both methods.

Moreover, it's important to respect and appreciate the efforts of research participants by improving their experience. The main reason for people to participate in market research is still rewards, which might lead to the conclusion that the more attractive the incentives are, the happier the participants are with their participation. However, there are also other factors that affect the experience consumers have when participating in market research e.g. the way the data is collected in terms of usability and privacy or the effort it takes to receive an incentive.

To avoid misconceptions regarding the data protection, we see that the communication about the privacy of passive metering needs to be improved to ensure research participants are aware what data is collected for which purpose.

The participants' demand for more efficient surveys as well as their frustration when being screened out, can be addressed by using passive metering for targeting a specific survey sample. This way of sampling can improve the relevance of the survey, so the experience they have with it and prevents them from dropping out. Recent research has also indicated better targeting increases participation rates, reduces dropout rates, improves survey experience and yields the same or better survey data quality.

To conclude, we think there are two primary factors to always keep in mind when conducting a research in order to succeed in the future: Data quality and user experience.

The seamless integration of mobile devices into people’s lives makes the understanding of consumer behavior more complicated than ever before. In the process of trying to understand these behaviors, data collection has mostly continued with traditional methods – expecting respondents at times to bear the cognitive load with long, self completion surveys, resulting in “bad data” and a decline in response rates.

In a world where online sample becomes a cheap commodity, we tend to forget that sample is actually people. The people whose opinions we claim to value. But do we really know and respect the effects of our practices on the people we rely on for information? Sample is to research as soil is to growing food. We need to understand how we can best maintain and enrich our soil. In that, there is something for everyone.

That's why we think it's important to look into the way market research is conducted. In this blog post, we will present the benefits and challenges of passive metering and surveying. We will explore if consumers' perception of their online activities reflects their true browsing behavior by analyzing the websites and apps they use and investigating for which purposes one data collection method is better suited than the other. Additionally, the second part of the study will look into the research experience of participants for both data collection methods.

Research Design

For this study, we used data from 485 respondents in Australia, which includes their tracked behavioral data as well as their completed post-surveys. We passively tracked their online behavior during the entire month of April 2017. We had 228 desktop participants and 103 additional smartphone participants, and the rest (154) were cross-device.

Research Results

We started the post-survey by asking the respondents a simple question: How much time do you spend on your desktop device per day for your private usage? We gave a range of times as answers, and the respondents simply had to pick one.

Time spent on desktop

We found that only 26% of our panel was able to correctly assess how much time they spent on their desktop every day. The other 74% were either over- or underestimating their usage, with over half of the people overestimating.

Time spent on mobile

We also asked our respondents the same question in regards to their mobile usage.

In this case, people were able to estimate their time spent slightly more accurately – 29% - but even so, most of the respondents were overestimating their mobile usage.

Time spent per category: desktop

Apart from asking about their general online usage, we also asked the respondents to indicate which categories they are interested in. For each of those categories, they were then asked to estimate how much time they spend on it using their desktop device.

Overall, respondents were only correct 54% of the time, meaning almost half of their answers were wrong. People also overwhelmingly overestimated how much time they spent per category.

What is interesting to note is that out of the 547 correct responses, 544 were '15 minutes or less'. This means that people are very accurate when judging their overall usage if split by 'very little time' or 'a lot of time', but when asked to provide more details, their estimates are incorrect. Only three of the correct answers were above the '15 minutes or less' category (two times 16 to 30 minutes, one time 31 to 45 minutes.) Since so many respondents chose '15 minutes or less' per category, there was very little underestimation (1%).

Time spent per category: mobile

For mobile, we again see a slightly better percentage of correct answers – 63% – however, in this case, all correct guesses were in the '15 or less minutes' category. This means that once again, people can accurately say that they spend very little time on a category, but can not estimate how much time they spend if it is above 15 minutes.

Reported vs. measured: favorite website/app

We also compared the sites which people indicated as their most used per category to what we measured their favourite to be (we defined favourite as website/app with the longest duration).

We found that, comparing the top 5 indicated vs. measured per category, only 31% of the answers were correct.

It was very obvious from our research that people use the same sites/apps for a variety of different reasons. Many websites and apps are so diverse in content that in some cases that they are difficult to classify into one category. For example, Facebook showed up as a favourite in all 15 of our categories, including animals, business, and property.

The most common answer participants gave us was either 'I don't know' or incomprehensible words and numbers, indicating that people were filling in the survey not to express their opinion, but just for the sake of filling it in. Immediately, this made 10% of our survey data unusable.

Conclusion

From this first part of our study, we can conclude that people can't accurately report their online behavior because it is too complex. We also found that while people can accurately assess if they spend a very short amount of time online, they can't estimate longer periods of activity. People also tend to overestimate the time they spend online.

To avoid wrong conclusions from a single source of data, it is essential to carefully select and merge the right data sources to create an accurate and complete picture of consumers.

In the next part of this study, we will examine the research experience of participants for both data collection methods.

This blog post will give an introduction into behavioral data and how you can start working with it.

We will do this by showing you what metrics you can think about when looking at the different data files of Wakoopa, as well as creating small use case examples which you can use as inspiration.

Why is behavioral data so different from traditional survey data?

When talking about online data collection for market research purposes, we are usually interested in two different types of data.

• Opinions (stated), which refers to subjective data like emotions, intentions, moods or preferences; that is, all kinds of information that is inside our brains.
• Behavior (observed), which is a kind of record of our physical actions, like the products we purchase in webshops.

Both of the data sources are needed to fully understand consumers, but both are of a different nature and need a completely different approach.

People change their minds quite often. People might change their preference for a political party, a brand, or a product. But facts are facts; if last week I purchased a smartphone, nothing will change this fact, even if I am not fully satisfied with it.

Differences in approach to analysis

When we look at the way the two sources of data need to be analyzed to make sure we get valuable output, there are big differences between them.

The most visible difference between the traditional survey data and behavioral data lies in the initial output.

Surveys are aimed at getting a specific question answered, and this is done by sending a survey with targeted questions to a possibly pre-defined group of participants. The outcome of this survey are answers that are often pretty linear. On question 1 you can answer A-E, and that goes on for a number of questions, with some dependencies between them. However, in the end, the format of the output is predictable.

The part of defining the questions and drawing up the questionnaire is the critical phase here. When the goal of the research is taken into consideration during the definition of the questions, the questions can and will be formulated in such a way they will lead to answers and data that are useful for the project and are easy to process.

When we look at behavioral data, we are looking at an outcome which is the online data of an N amount of people for an X amount of days. Nowhere in this process it is known if this data will be useful, what data is actually useful and what data should be looked at.

This means that it can be a bit overwhelming when you start your analysis by merely looking at huge sets of behavioral data.

If you are looking at one day of data of a regular internet user, you are looking at hundreds of pageviews on desktop, and the same on mobile if the participant installed cross-device.

To give you an example of how quickly data sets can become large, when we look at our own data, a panel of 200 participants creates about 1 million rows of desktop data per month. Microsoft Excel cannot handle files with more than 1,048,576 rows. (source)

So you can imagine that this type of data can run into large numbers very quickly, and that calls for a different approach.

How to approach behavioral data from a market research perspective

Whereas the initial filtering on data and participants is traditionally done when creating the survey questions, this filtering is now one of the most important parts of the analytics on a data level.

Defining the points of interest, segments, and categories is important to start off from before looking at the data itself. Defining what you want to see in the end, what information is of interest to you, and what will add value is the same type of creativity and logic as defining survey questions.

So, are there specific groups in my sample I am interested in? Do I already know certain demographics that are particularly interesting? Do I want to create segments of people based on certain behaviors? Am I working with pre-defined communities or groups? Am I interested in a specific industry or domain? etc.

Answering these questions before even opening the data files will give you easy opportunities to filter or subset the data, which will allow you to work quicker and will give you the possibility to work with smaller files as well.

The next step is to further define your research questions. If you have decided that mobile data is interesting for you, you need to define if you want to see mobile data produced with apps, or on websites. If it is on apps, is it for all apps, a specific category, or a few apps?

Defining these initial research questions as detailed as possible can and will save you a lot of time when executing the analysis.

A simple example of a use case you might want to answer with behavioral data can be:

“Which website categories are more popular with my target audience than for the overall population?”

I have made all words that are variables in this question bold.
Looking at an example this basic, you can already define a lot of different metrics and more precise questions that will help you to filter and subset the data with, making it easier to handle.

Website: We know we are looking at websites, this means that app data is not of importance here and can be removed from the data.

Categories: The main object of interest are website categories, so we aggregate our data set based on the categories. No need to run analysis over exports showing all visited domains or even full length URLs.

More popular: By identifying what is more popular you create the main metrics of this question. Is a site popular when 1) it is visited most often, 2) it is visited by the most unique visitors, 3) most time is spent in total, 4) most time spent on average etc.? If all metrics are interesting, create multiple research questions.

Target audience: Here we look to identify who our target audience is. If you say you are interested in ‘young male population’, then define a young male. Is this based on age ranges 18-25 or from 16-30 etc? If your audience is ‘people who have visited website XX’ you can also do this based on behavioral data. If you have identified the audience, you can filter the data on the data of only these individuals.

Running the actual analysis

Continuing with the same example as stated previously, “Which website categories are more popular with my target audience than for the overall population?”, a lot of research questions could be formulated:

• Which website categories have the most visits from panelists between 20-30 years old?
• Which website categories have the most visits from all panelists?
• Which website categories have the most unique visitors from panelists between 20-30 years old?
• Which website categories have the most unique visitors from all panelists?
• Which website categories have the most time spent by panelists between 20-30 years old?
• Which website categories have the most time spent by all panelists?
• Which website categories have the longest visits on average by panelists between 20-30 years old?
• Which website categories have the longest visits on average by all panelists?

So even though we are looking at a research question that is seemingly very simple, we already have 8 different, more precise, questions based on it. If you start with a more complex research objective, you can create more research questions. But you will see that when you break down these questions into as much separate ones as you can, all the separate questions will be easy answerable questions.

Example of answering one of the questions

When the detailed questions are formulated, answering it is easy.
If we take: “Which website categories have the most visits from panelists between 20-30 years old”, answering needs the following steps (regardless of the analysis tools used):

• Subset data on target audience.
• Count the total number of visits to a website category and order them from high to low.

If you then want to zoom in on one category, you can remove all data that does not belong to this category, and do the same analysis, but change the category for the domain.

Conclusion

As with every task, preparation is key when we look at behavioral data analysis. Answering clearly defined questions is not difficult. The biggest trick lies in the definition and knowing what is of interest in the data, even before looking at it.

This way of thinking is something that is important throughout the entire research design, from sales to execution. When a prospect approaches you with a question such as ‘I want to know what people do online’ there should be an immediate response to further define this question so if you can identify which research methods are necessary, which people should be invited, and help the research team to come up with research questions that can be defined as detailed as necessary.

If you need guidance to set up a research or if you would like to discuss how you could approach a project, just contact Wakoopa. We are happy to help!

Top shopping domains

We created a top 5 for shopping sites in both countries, based on total visits.

Rakuten is the most visited shopping site in Japan by far, not just reaching number 1 in shopping domain visits but also number 2 across all domains in terms of reach. Amazon is used far more frequently in Japan than in Taiwan, which is due to the fact that Amazon doesn’t have a local website for Taiwan and has not expanded its services fully to Taiwan in the same way it has for Japan. Ruten, a consumer-to-consumer marketplace, ranked highest in Taiwan.

Top shopping apps

In the top 5 ranking for shopping apps, while Rakuten and Amazon remain the number 1 and 2 in Japan, Taiwan’s ranking is radically different from the top 5 domains, introducing Shopee as the number 1 app. Yahoo also makes an appearance here, reflecting the fact that it is the site with this highest reach overall for both countries.

Most active shopping time of the day

We also found that when it comes to shopping, Japan busiest hour is 10pm in terms of unique visits. We saw something similar when we conducted ourglobal case study, where the highest online activity on desktop for Japan was 9pm. Generally, our Japanese participants were most active late in the evening. In contrast, Taiwan’s busiest hour for shopping happened around 9am.

Average session duration

Within our panel, we found the average length of a session in Japan is more than twice as long as in Taiwan, and almost 3 times as long for a purchase session. This is something we also found when we examined individual paths to purchase. Our example for Japan consists of two long sessions, whereas for Taiwan the journey happens in multiple short sessions split over a week.

Individual path to purchase: Japan

Using our passively tracked data, we were able to reconstruct two paths to smartphone purchase journeys from our panels.

Our Japanese participant started her journey with seeing an advertisement from Mineo, saying that if she purchases a device from Mineo and attracts new clients she will get a ¥2000 Amazon voucher for each person she recruits. After seeing the ad, she looked up 'cheap sim comparison’ on Rakuten websearch, made the decision to switch from her DoCoMo device to Mineo, and decided on ordering a Fujitsu Arrows m03 phone. She might also have recruited a family member/friend, to take advantage of the deal.

From here, she went straight to Rakuten and searched 'smartphone case’. She looked at a range of different cases, while occasionally switching back to look at the device she just purchased on the Mineo website (perhaps to see what it looks like and compare how it would work with the cases she is seeing). She also checked phone cases on Amazon, but decided to return to Rakuten to continue her browsing. The next day, she returned to Rakuten and narrowed her search to the ‘Arrows’ range of phone cases.

After some more browsing she narrowed her search even further, specifically searching for ‘arrows m03 case’ on Rakuten and Amazon. Finally, she decided on two cases on Rakuten and added them to her basket. She stayed on Rakuten to also search for a screen protector, which she added two to her basket after reading reviews. The total time it took her to buy the phone cases and screen protectors was 2 hours and 19 minutes, and involved 345 pageviews.

In her post-survey, she confirmed that she bought a Fujitsu Arrows m03 phone, and that her family was her main source of information about the purchase as well as the trigger for purchase.

Individual path to purchase: Taiwan

This journey starts with our Taiwanese participant searching for the Huawei Honor 3C on Google. From there, he visited a price comparison website to see some of the prices for the phone. He also googled Taiwan Mobile, and started looking around the site at mobile devices, but didn't find what he's looking for and makes no purchase.

The next day, he went straight to Xiaomi (a phone manufacturer) on his mobile device to start looking for phones. Here, he looked at a range of different models, and subsequently looked up the Xiaomi Redmi Note 4x on shopping.friday.tw. A day later, he returned to the manufacturer website on his mobile device to look at different phones once again, but still made no purchase.

On day four, he used Google to find 'landtop’, a price comparison website.

Here, he looked at the 50 newest phones and their prices. However, the 'Xiaomi' brand of phones are not included in the list, so he looked up the Redmi Note 4x again on Google. From there, he returned to the manufacturer page and looked at the product pages for both the Redmi Note 4 and 4x.

Finally, he chose the Note 4x, and started looking at prices of the phone on shopping.friday.tw. Here, he found the phone he wanted, logged into the site, and made the purchase. The purchase session of the phone itself took 71 minutes.

The next day, he checked manufacturer site again, where he signed up and created an account, perhaps so he could buy accessories.

Two days after his purchase of the phone, he started looking at phone cases and screen protectors on GoHappy, TaoBao and Tmall, but did not make an purchase yet. However, he spent the following day by looking at various phone cases and screen protectors on Shopee, and after browsing for a while, looked on Xiaomi, also for accessories. Finally, he returned to Shopee and purchased both a phone case and a screen protector on the site.

The total purchase journeys happened over the span of a week and involved sessions on both desktop (shaded blue) and mobile (shaded orange) activity.

In the post-survey, he indicated that he did indeed purchase the Xiaomi Redmi Note 4x, that this purchase was a gift, and that he bought it online because he got a better price and because he prefers online shopping in general.

Conclusion

The main takeaways from this study are that there are major differences in the online behavior across Japan and Taiwan, as well as some similarities. One of the most striking differences is the sheer difference in amount of time spent online, which was almost twice as much on desktop for Japan and seven times as much on mobile. Japan was also mobile-first, whereas Taiwanese panelists used their desktop device more. Additionally, we found that the lengths of sessions for Japan, both ‘normal’ sessions and purchase sessions, were longer than for Taiwan. However, this does not necessarily mean they spent more time browsing before making a purchase, as we can see with the individual paths to purchase - it could simply be that in Taiwan, panelists browsed more often but in shorter sessions.

There was also a marked difference in the use of social media platforms, specifically Facebook. As in much of the rest of the world, Facebook had a high reach in Taiwan. In Japan, on the other hand, it reached only 6th in most reached domain, and 13th in most visited domain. The app performed even worse, only reaching 51 unique visitors out of 256.

Rakuten was clearly the number 1 site for shopping in Japan, both on desktop and on mobile. It was also used occasionally in Taiwan, but their shopping was more spread out over multiple domains, such as Ruten, TaoBao and Shopee.

With this study, we not only revealed the overall purchase behavior and customer journeys in these markets, but also highlighted some of the biggest differences (and similarities!) between them. We hope you enjoy this first part!

Research objective

While we have conducted studies in Asia before (Japan was one of the countries we profiled in our global case study), we had yet to demonstrate how behavioral data can be effectively used to paint a picture of two very different Asian markets. Although they are only around 1,000km apart geographically, Japan and Taiwan have so many cultural differences, that we decided to explore whether these differences extend to people’s online behavior and the way they shop online. We also used smartphone purchasing as an example for an individual paths to purchase in both countries.

Research design

We used a mixed research approach, leveraging both survey-based research and passive metering to understand the participants’ online behavior in Japan and Taiwan. Using a pre-survey, we targeted consumers in Japan and Taiwan who were considering purchasing a smartphone in the coming month. We tracked 256 Japanese and 225 Taiwanese panelists from the GMO Research online panel. In total, we tracked 910 days on desktop, and 802 days on mobile across all participants.

Research results

Device usage

The first and most striking difference we found when we examined the data of both countries was the discrepancy in device usage.

Taiwanese panelists spent roughly 1 ¼ hours on their desktop device per day on average, and only 20 minutes on their mobile device. In contrast, Japanese panelists spent 2 hours on their desktop, and 2 ½ hours on their mobile device per day on average.

This means that Japanese panelists spent almost twice as much time on their desktop, and almost seven times more time on their mobile devices than our Taiwanese panelists.

Moreover, Japanese panelists spent roughly 15 more minutes on their mobile than desktop devices, while Taiwanese spent 3,5 times more time on their desktop devices.

Top 5 domains

Apart from device usage, we also established a top 5 ranking of the most commonly used domains and apps across the panels. The ranking is made using reach, meaning the amount of unique visitors within the panel.

The top site for both countries is Yahoo. Yahoo has a strong presence in Japan^[1] and Taiwan^[2], being used more than any other search engine. However, it is used not only as a search engine, but also as a news site, a shopping platform, as well as for a variety of other functions. A notable exception from the Japan list is Facebook, which reached number 2 for Taiwan but came in at 6th for Japan.

If we rank the top 5 most used sites in Japan by total number of visits, it remains the same except YouTube and Amazon switch places. Facebook falls even further, becoming 13th.

For the Taiwanese top 5, if ranked by total number of visits, Facebook becomes number 1, followed by Yahoo, YouTube and Google. Ruten (a shopping site) joins the top 5 instead of PIXNET.

Top 5 apps

For both countries, similar apps appear in the top 5 – YouTube is number 1, followed by a combination of Line, Google Maps and Google Search. Just like with the domains, it is notable here that Facebook is not present in the Japanese top five (only 51 unique visitors), yet ranks fourth for Taiwan.

Top domain per category

We also identified the top domain per category for both countries, and highlighted the most relevant and interesting ones here. We ranked the domains and apps according to reach percentage^[3].

YouTube dominates the audio, video & entertainment category in both countries, having around 70% reach within the panels.

The highest reach in either panel is achieved by Yahoo, which is in line with the fact that it is the number 1 domain overall for both Japan and Taiwan.

While Facebook has the highest reach percentage in the social category for Japan, it is still relatively low when compared to Taiwan, where it is around 90%. This shows that there is more of an open market in social media platforms for Japan – in terms of total visits, both Twitter and Ameblo beat Facebook.

On the other hand, Rakuten has a much higher reach percentage in Japan than Ruten has in Taiwan, since there is less diversity in the shopping domains used in Japan than there was in Taiwan.

Top app per category

As with domains, YouTube is number 1 for audio, video and entertainment for apps as well. Here, we can also see the popularity of Line quite clearly, which is the region's most used messaging app with more than 560 million users across Japan, Taiwan and Thailand.^[4]

Yahoo makes an appearance again as the top news app for Japan, while News Republic is number 1 for Taiwan.

In terms of social apps, Twitter ranks higher in reach percentage than Facebook for Japan, but Facebook is still number 1 for Taiwan.

Coming up

In the next part, we will shed light onto the top shopping domains and apps for both countries, as well as examine some individual paths to purchase for two smartphone purchases.

^[1]Search interest for Google, Yahoo! Japan by time, location and popularity on Google Trends. Retrieved June, 2017, from https://trends.google.com/trends/explore?date=today%12-m&geo=JP&q=Google%2C%2Fm%2F03kjmh

^[2]Search interest for Google, Yahoo! by time, location and popularity on Google Trends. Retrieved June, 2017, from https://trends.google.com/trends/explore?date=today%12-m&geo=TW&q=%2Fm%2F045c7b%2C%2Fm%2F019rl6

^[3] Reach percentage is defined as unique visitors in relation to the total N of the panel.

^[4] McCracken, H. (2015, November 02). How Japan's Line App Became A Culture-Changing, Revenue-Generating Phenomenon. Retrieved June, 2017, from https://www.fastcompany.com/3041578/how-japans-line-app-became-a-culture-changing-revenue-generat

Most used app per country

By calculating the reach^[1] of the apps used by our participants, we can put together a list of the apps used by the highest amount of people in various categories per country.

Social apps dominate this category, with every country except Japan having the highest reach on either Facebook or WhatsApp. Line, which is big in Japan, is also a social and messaging app.

Most used gaming app per country

Using the same approach as above, but adding categorization, we can also find the most gaming app used by the most amount of participants per country.

Again, there is not a huge difference between the countries, with Candy Crush Saga clearly coming out on top in every country except France and Japan. It is worth noting that the Japanese game, Fishdom, has similar gaming features to Candy Crush Saga. France is the only country with a racing game as their top used gaming app.

Most popular YouTube search per country

Since our technology also tracks search queries (which are extractable from the URL), we can create the following ranking of the most frequently searched teams.

For Mexico, Descpacito is a song by artist Luis Fonsi, which was very popular in Latin America around the time of our study. In Brazil, Felipe Neto is a YouTube ‘star’ with over 10 million subscribers. In the US, ‘April the giraffe live’ refers to a livestream of a giraffe in the late stages of pregnancy and subsequent birth, which was broadcast live on YouTube throughout February and March.

The top search terms of the UK, Germany and Australia all relate to the first two singles from British singer Ed Sheeran’s latest album, ’Shape of You’ and ‘Castle on the Hill’. Those songs were released in January 2017 and broke records in a number of countries, including the UK, Australia and Germany. For France, Cyprien is a popular French YouTube ‘star’, also with over 10 million subscribers. The Japanese search term refers to a song by the Japanese girl group Keyakizaka46 called Futari Saison. It reached number 1 on the Japanese Top 100 Billboard Charts, and broke sales records for the first time in 20 years.

Most watched Netflix series per country

Using the series ID we identified from the Netflix URLs that our users visited, we are also able to find out which Netflix series were most popular for which countries.
The only country we didn't select a most watched Netflix series for was Japan, because it was used too infrequently to identify a series that was most watched.

We also segmented the data to look at the most popular Netflix show per gender.

Most visited news landing page

With our categorization, we can easily pick out the most visited domain per website category. For example, we identified the top visited news landing page per country.

The article for Mexico and Brazil was the same, but with different translations. In Australia, the most visited news page showed a weather map, which is understandable, since there was an heat wave in the country during our study.

Average number of pageviews per shopping session

The above graph shows the average number of pageviews per session when a shopping site is visited. We can conclude that in France, consumers spend a lot of time per session looking into products, perhaps using various websites for comparison. Australia, on the other hand, only visits on average 18 different pages in a shopping session.

Conclusion

We think the best way to visualize exactly what we mean when we talk about the power of behavioral data is by using this iceberg.

The visible part of the iceberg is the URL. This is the foundation of what we do – we passively collect behavioral data in the form of URLs, then adding identifiers in the form of timestamps and user IDs.

Even just adding simple components such as demographic data, or combining behavioral data with in-the-moment and post-surveys, can reveal invaluable insights into customer segments and customer journeys across the entire globe. Measuring and analyzing who (age, gender, etc.) browses what (websites, apps, search terms), where (country, category, device), and when (time, duration, frequency) helps us to improve the understanding of consumers’ online behavior. This understanding can be used to improve customer experiences (through better targeting and communication) as well as to refine products and services.

^[1] Reach refers to the total number of different people or households exposed, at least once, to a domain during our observational period.

With that in the back of our minds, we started planning our biggest, most ambitious case study yet – tracking 500 people each in 8 different countries, only using Wakoopa's passive metering technology. We wanted to create a truly global study that would cover a range of different markets to demonstrate that our technology can be applied anywhere in the world.

Together with some of our global panel partners (GapFish, Respondi, GMO Research, SoapBoxSample, TEG Rewards, and Netquest) we created 'Around the world with behavioral data.' Enjoy!

This study was first presented at WHAT Conference by Pavel Vilensky.

The anatomy of behavioral data

Before we can dive into the study, we think it's important to explain what exactly we mean when we say we use our software to collect passively metered 'behavioral data'. What we really collect is URLs, which contain all the information we need to analyze where users went, when they went there, and how they got there. While sometimes a URL can seem like one long string of letters and words, when we break it down, it becomes our guide to everything we need to know about the users' behavior.

Each participant is given a unique ID – this is what we can use to identify patterns of behavior from one user without having to use any personal information. Next comes the main part of the URL – the sub domain, top domain, and extension. After this, we can also see the 'path', namely which part of the website the user visited. If there was a referrer, we would also see where they came from in the path. If the URL contains a search query, this is also where we would find it. Finally, we collect the timestamp of each visit, so we can keep track of information such as how much time was spent on each visit, or what was the most active part of the day for our users. For app visits, this remains largely the same, except that the URL is replaced by an app ID.

Enriching the data

Using our collected behavioral data, we can then apply other information we already have in order to turn this raw data into useful consumer insight.

By using the information we have gathered from pre-tracking surveys (such as their age, gender, and location) we can start to segment our findings into different categories. Using pre-defined metrics, we can mix and match these various information points to extract the knowledge we want to have.

Research objective

Our research objective was to uncover the differences and/or similarities in cross-device online behavior across countries. For this objective, we chose to study (in order from left to right, top row first) Japan, France, Germany, USA, Mexico, Brazil, Australia and the UK. (Do you recognize all the countries' symbols? Keep them in mind, because they will appear again later on in this post!)

We wanted to see what we could understand about their culture, device preferences, search behaviors, media consumption and general online behavior by using passive metering.

Research design

We tracked the users over 21 days, from February 20 to March 21, 2017. In each country we had 500 desktop participants, of which 250 were also cross-device. We kept a fairly even split between male and female respondents, and tried to represent as many generations as we could in our study. In total, the time we tracked of all the participants together adds up to to 17 years on desktop, and an additional 11 years on mobile.

This volume of data was almost overwhelming, and it was a monster task to pick out the most exciting, interesting, and useful findings. What we will show here are some of the highlights of our discoveries, to demonstrate what we can observe from user behavior just with passive metering – without having to ask any questions at all.

Research results

Let's start with some easy insights to get the ball rolling – the top sites based on pageviews, aggregated across all participants. Using our own categorization, we can quickly find the most visited websites per category as well.

Top websites

No surprise in the most used site overall - Google is used by participants across the board in every country. Facebook makes a strong case for most popular social media platform, showing up in the top five most visited websites for every age category except <16 year olds. YouTube is the most used website for video streaming and online content consumption. We also noticed that amongst younger people (<16), it showed up very high on their top five list of websites. The older the age category, the lower YouTube showed up in their ranking, disappearing entirely in the top five of anyone over 45.
When we examined shopping and travel, we started to notice our first cultural differences. Amazon had very high pageviews in European countries (France, Germany and the UK) and, naturally, in the US. However, in Mexico and Brazil, the website that was used by most people for shopping was mercadolibre.com.mx/mercadolivre.com.br, which is technically the same website, just in Spanish for Mexico, and Portuguese for Brazil.

We can see Amazon has not made a huge impact on the Australian market yet, where eBay is still more popular. This is due to the fact that Amazon in Australia is very limited, selling only Kindles and various eBooks. It does not have the same range of items and services as it does in other English-speaking countries, which accounts for the discrepancy.

As we will see more often in this study, Japan is an exception in terms of international websites and apps, primarily using their native platforms rather than multi-national ones. Here, they used the Japanese eCommerce giant Rakuten most often for online shopping.

For travel, the top two contenders were Tripadvisor and Booking.com, coming out on top in six out of eight of our countries. Shermantravels.com offers exclusive deals on travel, and Jalan is again a native Japanese site, used only in their market.

Top five domains by gender

This was one of the insights which surprised us – on the surface, the top 5 for men and women were almost exactly the same, except for one domain. The order of the top five did also differ slightly - men visited YouTube more often and had it ranked at their number three, while women visited live.com more than YouTube. However, the difference in the amount of pageviews is also noteworthy. While both genders have Facebook as their number one, women visited Facebook roughly 50% more than their number two, Google.com, while men only visited Facebook on average around 15% more than Google.com.

Unique apps and domains per country

By measuring the average amount of unique apps and domains each country visited, we created the following graph. These numbers do not reflect the domains and apps available to each country, but shows their actual recorded usage.

Germans are clearly aware of the range of apps available to them, even though their web domain usage is not very diversified. In the US, people are aware that they have more choice in their online browsing, and subsequently they are less loyal to specific websites.

Desktop vs. mobile sessions per country

We collated the average session^[1] data we collected for desktop and for mobile activity and compared them per country.

The main insight we can take away here is that mobile dominates all markets by a large margin, the biggest difference between the two device types being in Japan. The fact that Germany had the most mobile sessions on average is concurrent with their status as the country with the most diversified unique app usage.

App vs. web on mobile per country

Using our passive metering technology, we can also track the two types of mobile behavior – apps and browsers.

Japan continues its trend of being an outlier by having a relatively low app to total mobile time percentage, hovering around 71% while the rest reaches between 86%-94%.

Time spent on mobile vs. desktop per day per country

By adding the data from cross-device users on mobile and desktop (on average) together we can see whether weekends (shaded area), different days of the week, or events such as holidays have an impact on which devices they use. For example, we can see here that Australia's mobile usage increased during the last recorded weekend, which was a long weekend due to various national holidays.

Most popular hour to be online per country

Another interesting insight we derived from the data is this time-based metric – when people are most active on their desktop device.

Australians are the only ones with a desktop usage peak at around noon, and once again Japan is the exception, clocking in at a late 9pm. The Latin American countries (Brazil and Mexico) have their most active period in the afternoon. The rest of the countries are most active from 5 to 7pm, which is usually the post-school or work period.

Coming up

In the next part, we look deeper into specific apps and online content consumption of the different countries.

^[1] We define a session as any activity from one participant that occurs without a pause longer than 30 minutes - for example, someone might be searching something on their laptop, close it, and grab their phone to continue their online browsing 25 minutes later. This is counted as one session.

You can download the available presentations that were given at the conference below.

Finding the right touchpoint - shifting the focus from reach to the quality of engagement

Ansie Lombaard, Global Innovation Director | Behavioral Data & Chris Davies, Global Innovation Manager| Behavioral Data, Kantar

Consumers today experience brands in more ways than ever before, and every experience can potentially change attitudes and behavior, positively or negatively. These experiences are linked to touchpoints, moments where consumers engage with and are exposed to a brand – whether in the form of a TV advert, a billboard, a shop display, or retailer site.

For years, ‘finding the right touchpoint’ has focused on reach, and shifted emphasis from traditional to digital. However, as technologies and consumer behaviors evolve, it is becoming increasingly clear that touchpoints perform differently – even if they have similar reach.

While reach is important, our case study shows that also looking at memorability and impact helps marketers to re-focus the search ‘for the right touchpoint’ on effectiveness and quality of engagement. We will demonstrate that the unique combination of survey and behavioral data helps to provide deeper perspective on what happens at the touchpoint.

In this context, behavioral data not only helps to identify the moment of brand engagement, but also offers insight into the broader behavioral context of such engagement. When combined with survey data, it supports improved media planning, touchpoint selection, and messaging strategies. In short, it makes ‘finding the right touchpoint’ easier.
Download presentation

Appfluencers: using passive data to identify the next big thing

Mark Jefford, Director of Data Applications, YouGov

2016 saw the launch and mass adoption of a number of brand new mobile apps, including the likes of Pokémon Go, Prisma and MSQRD. Apps come and go, but the people who drive this mass adoption share remarkably similar traits; they're affluent, influential and they take startup products and make them famous.

Using historical passive data dating back to 2013, we have identified The Appfluencers, the group of earliest adopters of new mobile app technologies. Tracking the online behavior of The Appfluencers over the opening weeks of this year, we have determined the apps that are most likely to reach a consumer tipping point over the coming months. Which apps will become household names in 2017? Which ones will be bought by the likes of Facebook and Apple?
Download presentation

Machine learning for behavioral analytics

Steve Dodson, PhD, Tech Lead, Machine Learning, Elastic

As volumes of data increase, manually searching and visualizing consumer or user behaviors becomes more and more difficult. An alternative approach is to use machine learning to automatically build behavioral models of these behaviors. These models enable users to gain deep insights into behavioral characteristics that are beyond the capabilities of classical search techniques.

Typical use cases include automatically understanding users that are behaving unusually and understanding the typical behavior of the population.

This talk presents real examples of machine learning techniques applied to real-time behavioral data, and describes the methodology behind these methods along with an overview of the machine learning space.
Download presentation

Around the world with behavioral data

Pavel Vilensky, International Business Development Manager, Wakoopa

Wakoopa leverages the power of the global availability of behavioral data and showcases how consumers around the world are using their devices. With a cross-cultural, cross-country and cross-device view we zoom into the behavior of consumers and demonstrate how they shop, search and socialize online. Besides this we uncover how content and media are consumed on the different devices by users. Finally, we compare different segments cross country to see if they behave in the same way.
Download presentation

Segmenting online customer journeys: the U&A for digital strategy

François Erner, Chief Innovation Officer, Respondi

Segmentation is a wonderful tool to devise targeted business strategies: clusters are defined with distinct needs and attitudes to address customers in the most efficient way. Traditional segmentations are based on survey or CRM data, but online behavior is a third natural data resource to be used in this context. All the more so when business decisions concern digital strategy.

Our research focuses on online customer journeys: What do customers do before they buy online? And, since we aim for a fine-grained understanding by means of clustering, what are the typical kinds of online behavior?

We have studied and classified online customer journeys related to beauty products – from the early steps to the actual purchase. We have gathered the browsing habits of 400 women from our panel who were equipped with passive metering technology and tracked for 2 months.

Several types of online behaviors were identified. In particular, they connect information sources (youtube channel, beauty blogs, brand websites… ) with visited online shops. These segments portray, classify, and weigh the online behavior of women in this universe, providing online retailers and cosmetic brands with actionable tools to manage their digital strategy.
Download presentation

How marketing research will become free

Maarten Stramrood, Director Advanced Analytics, VodafoneZiggo

The availability of data exploded the last few years. In this jungle there is so much information that the traditional marketing research agency could get lost. And now more than ever this will create huge opportunities for everybody. But be aware! Competition in your space will grow and unexpected players could destroy your business in weeks. This talk explains the opportunities, but also showcases some current threats.
Download presentation

Innovation track

Edwin Rietberg, Key Account Director, DAN DNA

DAN DNA introduces Motley for Persona-based marketing. It helps to turn data from social media into insights through personas for digital marketing. Motley was founded by two brothers David and Benjamin Borch, who are still in charge of rolling the tool out to other European markets. Motley’s insights about consumers is important to brands across Europe. Motley is able to turn volumes of Facebook data into actionable business insights. The tool has been used on clients such as Coca-Cola and Arla. With the Motley we offer our clients unique insights into their target groups. This is based on the people who actually interact with the brand on either owned or social media. This gives clients a clear, competitive edge in the digital era.
Download presentation

Daniel Tjondronegoro, Co-Founder, Beatgrid Media

Beatgrid offers a new single source cross media meter that measures a person’s actual exposure to TV, digital video, radio and OOH through the use of a smartphone microphone and geolocation. Unlike other technologies, Beatgrid’s proprietary Ambient Content Recognition technology can be completely performed within the app without an internet connection, while data and battery usage is negligible. This means it’s the mobile media meter that is able to scale with relative low panel costs. Beatgrid takes it down to the individual media consumer and ties it back to their user’s store traffic, purchases and brand opinion.
Download presentation

When reality hits: understand predictive behaviors today to prepare for the AR and VR experiences of tomorrow

Anjali Lai, Analyst, Data Insights, Forrester

Globally, consumers and businesses are feeling the early tremors of Augmented Reality and Virtual Reality – and early adopters have begun to respond. But will AR and VR technology cause a seismic shift in our lifestyles? And if so, which markets will be at the epicenter of change?

In our latest study, we blend our newest mobile behavioral tracking methodology with a sophisticated approach to passively tracked social media behavior. Through this, we can reveal how people across segments react to AR and VR concepts differently, how they are drawn to the technology for various reasons, and how they indicate latent opportunities for AR and VR to transform their activities and experiences.

In this session, we unveil our findings on what current consumer patterns forecast about the future trend of AR and VR, what this means for business and consumers, and why diverse methodologies are critical to predicting human behavior.
Download presentation

Results of the datathon

The winners from WHAT Datathon presented how they used their data science skills and ideas to turn raw behavioral data into valuable insights. You can read more about the results of the datathon here.

Measuring the impact of emotions on the reception of online media through a real-time-behavioral sampling approach

Malte Freksa, Director Business Development Mobile & Data, GapFish

To what extent do users differ in their mood and what impact can be observed on the acceptance of advertising? In this study we investigate differences in emotions: panelists who installed passive metering software were recruited via push-notification for an online survey regarding their in the moment emotional status when using a predefined website.

In addition, the internet usage data of the participants was collected via a measuring software which made it possible to analyze questionnaire and behavioral data at the same time. Results indicate the important role of emotions during the internet usage for both reported and observed data.
Download presentation

Measuring passive behavioral data: futureproof necessity

Peter van Eck, Sr. Research Specialist, GfK

In our always-on culture the context of the consumer is transforming fast. Understanding the complete journey of their digital shopping behavior is becoming the main challenge.

We help with answering relevant client questions as: which channels & devices are used in the path to purchase? When and how intense are touch points used? What differences are there between online and offline behavior? And last but not least what is the role of smartphones? All based on recent use cases from the Dutch market.
Download presentation

More information

You can also download the all presentations at once here.

If you want to learn more about the speakers, you can find the speaker descriptions here.

We hope you enjoy reading the presentations! Please feel free to reach out to us in case you have any questions.

We had initially planned the workshop for 18 participants. We set that limit to make sure everybody would benefit from the hands-on training atmosphere. As we found out later, we definitely could have had two sessions - the people on the waitlist could have filled another workshop!

When the first participants started arriving, workshop leaders Carlos, Zoltán and Bert made sure everybody was set up to follow the intensive R course we had in store for them.

Once all the machines were set up, our host Simon opened the workshop and welcomed everybody.

The first part of the course was 'Introduction to R', where Carlos explained the basic concepts of R, the logic behind it, and the objects. This background information prepared the participants for the second and third part of the course, where the exploration of real-life behavioral data began.

"The R workshop, connecting with other data nerds... [was] really thought provoking and also fun.”

After a short break the section on 'Exploratory data analysis with behavioral data' started, where the participants learned how to deal with more than a million observations. They also discovered how to read in the data, access it, run basic statistics and deal with 'time'.

After a delicious Dutch-style lunch of a variety of sandwiches, the course entered the third and final stage – 'Getting insights from the data' – where participants were given free reign to apply their previously learned knowledge to the behavioral dataset (of course, under the watchful eye of our teachers!).

"It was good to be around likeminded people and have a good time while learning."

At the end of the course, in recognition of their successful completion of the intensive analysis workshop, the participants received a special diploma.

"Very inspired by the workshop."

After the intense five-and-a-half hour course, the participants were invited to enjoy a drink at TQ's bar and discuss everything they had learned with each other.

"The workshop was useful, insightful, and relevant."

Afterwards, those who still had the energy joined other conference attendees at bar Louis in Amsterdam for the official pre-WHAT Conference drinks!

Overall, the workshop was a day filled with interactive behavioral data learning, networking, and fun - 100% of respondents would come again, and would recommend the workshop to their friends. 100% also rated the workshop content, location and organization 4 stars or more! We want to thank all the participants for being great students, and we hope to see you all again soon!

For more images from WHAT Workshop head to the facebook page.

Jokes aside, we decided to create WHAT for a purpose. WHAT is aiming to impact market research and encourage the industry to embrace the opportunity of data by working towards a user-centric, data-driven future. The fact is: data is the new research.

Welcome to WHAT!

The beautiful morning of April 26th was the start for this year’s WHAT Conference hosted at TQ Amsterdam. With this gorgeous view over the city from the balcony of the venue, we appreciated the weather being our friend almost during the whole day!

The first important point on our agenda: Welcoming our attendees warmly with valuable goodies and a nice breakfast to prepare them for a day packed with a lot of inspiration, knowledge and networking.

At WHAT, our world doesn’t just revolve around behavioral data. We also care about the world we live in, and want to make sure it is preserved for the next generations. In order to do our little bit to help, we made the items in the goodie bag out of as many recycled and reusable materials as we could. Our favorite: The special WHATer bottle to help reduce plastic waste from disposable water containers:

The air was thick with love and excitement for data. Everybody was curious to see the upcoming talks!

The opening words were given by our very own Simon, Managing Director of Wakoopa, who led the conference during the conference day.

"Invest in the skills, the people and the organization to harness data and give you the methodologies and insights that allow you to compete in the future of research."

"If you build it, they will come!"

As you can see, they came:

99% of the 152 registered market research and data professionals from all over the world showed up at WHAT to share strategies, best practices and valuable insights about behavioral data.

The longest travel distance to WHAT Conference: 9,305 kilometers

Mostly thanks to our numerous attendees from Japan!

17 speakers | 9 different nationalities

What would have been a better opening talk than inspiring the audience with creativity! Emmanuel Flores, Innovation Director at JWT Amsterdam, was kicking of with an award-winning case study, demonstrating how data can fuel creativity, or, in other words, how ideas can be supported by technology: THE NEXT REMBRANDT.

From creativity, to data, to insights, to vizualization

The opening talk was followed by loads of diverse talks to shed light on all different angles of behavioral data. Here are some snapshots from the sessions during the day:

"Finding the right touchpoint - shifting the focus from reach to the quality of engagement" by Ansie Lombaard & Chris Davies, Kantar. Thank you both for taking the long journey all the way from South Africa!

"Machine learning for behavioral analytics" by Steve Dodson, Elastic. Steve dove deeply into machine learning techniques and the logistics behind such methodologies.

"Around the world with behavioral data" by Pavel Vilensky, Wakoopa. Analyzing 28 years of behavioral data from 8 different countries? It’s actually not that complex.

"How marketing research will become free" by Maarten Stramrood, VodafoneZiggo. A provocative talk about how market researchers can remain relevant and save themselves from extinction.

Constructive breaks

Eat!

Chat!

Play!

Get back to work!

"The future of data visualization" by Gert Franke, CLEVER°FRANKE. "Don't throw information to people, tell a story!"

"When reality hits: understand predictive behaviors today to prepare for the AR and VR experiences of tomorrow" by Anjali Lai, Forrester. Thanks for coming all the long way from the US!

Before the final part of presentations, it was definitely time for a sugar boost!

83% of our post-survey respondents rated the food at the conference with 4+ stars. Our guess: It was because they "loved the red velvet cake". More than 80% of the cake was eaten – not a bad result as we had plenty ;) You should try it!

With their bellies full of cake, the WHAT Datathon winner teams presented their solutions...

...and got awarded with fame and prizes!

The day was wrapped up by Marc Tollens and Jan van der Vegt from KLM with their talk about "Predicting your next destination @KLM". KLM’s site leverages behavioral algorithms that can predict future travel destinations with 98% accuracy based on the visitors’ last search!

Our next destination: #WHAT18

The event was not just talks! We had a pre-event datathon, a half-day behavioral data workshop, a pre-event networking session and last but not least: the after-event drinks after the closing of the conference.

Party on!

The evening was laid back and filled with lovely music, delicious refreshments, interesting conversations, ...

...and ended with a lot of empty glasses!

WHAT a day!

Here are a few examples of the many comments the world’s first behavioral data conference triggered:

"Great initiative, great atmosphere, interesting people."

"Diverse content, not sales-y."

"Fantastic job - amazing venue, very well organised, good communication throughout, very well done!"

"Learned a lot, and great conversations afterwards."

"It was very informative and the networking was good. I'm still new to the world of behavioural data and this has been an eye-opener."

"You simply have to attend if you want to know what's going on in behavioral reasearch."

"Not enough people are seeing the value of behavioral data and how it can drive strategic decision making."

"Great conference for every marketing analyst, really nice (informal) vibe, great experience!"

"Please do another one of these! :)"

Key take-aways from WHAT Conference 2017

Still wondering, if you should attend the next edition of WHAT?

94% of the people who answered our post-event survey rated the show 4+ stars!

94% will consider to attend WHAT Conference again and would recommend the event to their colleagues and/or friends.

We were aiming to create something fresh, inspiring, enjoyable and knowledgeable. WHAT has proven to connect data and research professionals. Besides providing a day of learning from all kinds of companies within the consumer insights industry, it provided an interactive environment for networking and new business opportunities.

Moreover, we were able to create the conference without any additional monetary investment, even without any sponsorships.

Our goal will always be to become better and grow. Aiming for an increase in attendees and speakers will also enable us to cluster the talks in different tracks (e.g. data, technology, insights), so every attendee can create an individual schedule which fits their interests best.

Thanks, everybody!

First, thanks everybody for attending. Without you, this event wouldn’t have been nearly as amazing it was! Thanks for all the comments and ideas as well. We will take them to heart!

Also a huge 'Thank you' to all our speakers for contributing to the success of the very first edition of WHAT!

Thanks to TQ for making your lovely event space available to us! 72% of the people who filled in our post-event survey rated the venue with 5 out of 5. You truly have a beautiful space!

A big shout out to Simon, who was not not only an awesome Host at WHAT, but who is also a wonderful Managing Director at Wakoopa. Thank you for your time, flexibility, optimism, but also for giving us so much freedom in creating this event! It has been such a great experience to build something from scratch – and it's impressive to see WHAT is possible with a team that can be counted on the fingers of one hand, if you have the right people. And you are one of them!

And of course a big thank you to all Wakoopians who supported us before, during and after the conference!

More information

• Most of the presentations from the conference are available for download as a pdf-version here.

• For more images from WHAT head to the conference facebook page.

MAY THE DATA BE WITH YOU!

Awake your inner child, go out and explore. Be curious. Have fun. Fail. Try again. Kick ass. Exceed your own expectations. Shake off boredom. Amaze and be amazed.

All the best,
Your WHAT Conference Team

The participants could compete in three tracks:

The GfK Insights track asked participants to reveal meaningful insights that generate business value from the current data (e.g. path-to-purchase, segmentation, categorization).

The DAN DNA Visualization track challenged the participants to impress the jury with an outstanding visualization of the information and insights that they haven't seen before (e.g. spheres, flow charts, activity visualization).

Finally, for the IceMobile Predictive track the participants had to develop machine learning algorithms to predict the gender of a user based on their navigation in terms of sessions and urls (e.g. 200k sessions/14 million urls to train, and 50k sessions/3.4 million urls to test).

We started the day with a small breakfast, as we gave the participants some time to network, meet each other and form teams.

We also handed out special WHAT Datathon goodie bags, with a little message of support on the front for the tasks ahead.

The event was hosted at the TQ main event space, which was generously sponsored by them and provided a great open space (with a fantastic view!) for our participants to work.

After breakfast, the participants were given their briefing and access to the data.

At 11am, the clock started on the 24 hours of competition, after which the teams would have to present their findings from the behavioral data set to the jury.

During the 24 hours, we of course provided the participants with plenty of food, snacks, drinks and entertainment to keep them going through the long, sleepless night!

Many of the teams stayed through the night to make sure they had the best possible chance of winning one of the tracks. The teams worked hard during the night, as this graph tracking the activity on the machines provided by AWS shows. All times are UTC, which means there was a spike of activity at around 5am local time.

When the competition ended at 11am on Sunday, it was time for the presentations. All eight teams presented their findings to our jury of behavioral data experts from DAN DNA, IceMobile, GfK and Wakoopa.

There were some great insights, visualizations and predictive problem solutions presented by our teams. Some notable mentions include an analysis of the behavior of 'adult content' watchers, and some creative hand-drawn graphs.

After all the presentations were done, the participants enjoyed a catered lunch while the jury deliberated and decided on their winners for each track.

First up were the winners of the DAN DNA Visualization track, Team Cortex (Jesse Paquette, Sergio Bellon Alcarazo, and Marjolein Smits). Incredibly, Team Cortex was formed at the Datathon, and hadn't met before the event. They used a combination of Tableau and Jesse's own software, tag.bio, to visualize the user data.

Team Cortex used Tableau to visualize information about the users, such as their education, preferred bank, and age group.
Additionally, they also used tag.bio to create characterizations of user clusters.

The winners of the IceMobile Predictive track were Team TeaClub (Jedda Boyle, Madli Uutma, Taavi Kivisik, and Sebastian Mehldau) who had the highest AUC or 'area under the curve' score. The teams were given a mostly complete data set, with roughly 20% of it missing the genders of the users. The challenge was to predict whether the users were male or female based on their online behavior. Team TeaClub, by simply identifying some key domains which showed up predominantly for one gender, managed to score the highest and win the track.

The final scores for the predictive track, with Team TeaClub edging out the second place team by just a few points. The leaderboard was projected in the room all night, and teams were given the opportunity throughout the 24 hours to submit new predictions and see their score in real time.

Finally, the GfK Insights track was won by Team CodersCo (Mariann Lesko, Gosia Wrzesinska, Jan-Mark Wams, and Artem Duplinskiy), who used their own dashboard to visualize and cluster online fashion shoppers behavior based on the data set.

Using their dashboard, CodersCo visualized shopping behavior in order to provide insights for businesses about how their consumers are behaving online.

Even after staying awake all night and working hard on their presentations, some of the participants still had the energy to stay for some celebratory drinks.
WHAT Datathon was a huge success. We received a lot of great feedback, and even 100% of the respondents from our post-survey would recommend the event to friends/colleagues. And of course, we as Wakoopa can take away a lot of inspiration from the event. We are looking forward to the next Datathon in the future.

We want to give a huge thank you to everybody who came down to participate in the first ever WHAT Datathon!

Of course, the Datathon would not have been possible without the help of our fantastic sponsors. Thank you to DAN DNA, IceMobile, GfK, Blendle, Elastic, Catawiki, StartupAmsterdam, Neuro Flash, Fruitful Office, and Tony's Chocolonely for all the support!

For more images from WHAT Datathon head to the facebook page.