 Kia ora kato katoa, welcome everyone. Welcome to innovative methods in market research. Digital data collection via TikTok by Ana Maria Areos, the director or NCEO of Senseata UX Research, as well as senior researcher, Gabriel Camago. I am Marina Sanko, your co-host and the AES convener of Arturoa New Zealand. I will monitor the chat function throughout the seminar. For any questions you may have. Please talk. Sorry, if you can please keep yourselves muted. That would be wonderful. When there is an appropriate moment for questions, feel free to raise your hand. But however, before we start, I'd like to acknowledge the traditional custodians of the lands which we all come from. I'm speaking from Wellington, Arturoa, and I acknowledge our leaders, past, present, and emerging. And today we are very fortunate to have a visually enticing presentation by Ana Maria and Gabriel. Ana Maria has over a decade of experience in behavioral science and cultural studies. She pioneered Senseata UX Research during the COVID-19 pandemic and her market research company has successfully grown and collected over 1.5 million surveys in over 30 nations. They specialize in providing an enjoyable user experience with nonverbal interaction types, increasing response rates way more than traditional surveys do. So thank you, Ana Maria and Gabriel, for your time today. We really look forward to learning from you. Over to you. Thank you, Marie, for having us here. We're very glad to share with you our experience in these couple of years, what we've learned so far with this new paradigm. So I'm going to share. Are you watching my screen? Yes, yeah. OK, so we call this actually, this was, I hope there's no misunderstanding because this is the title of this presentation. We're talking here about the TikTok era, which is maybe some of you are familiar with TikTok, which is a social network that found people to expect entertainment experience of at last 30 seconds. So the attention span of human beings in the TikTok era is around 30 seconds. So the traditional paradigms of social research are very far from that approach. So we are making some important steps towards collecting data massively in this era. And we really believe that what we have to share can be very valuable for social researchers all around the world. So this is a fact that every organization needs quality data. But the speed at which the world is changing is way faster than the speed at which social research or research in general is changing. And collecting good quality data is not always easy. You probably are familiar with this. We didn't have time to know your backgrounds, but you're probably familiar with data collection and decreasing response rates all around the world, especially among the millennials or millennials populations. So we know as a fact that people find service boring. They demand a lot of cognitive efforts. Traditional service is used to have sometimes types of questions that can be had. They are very much based on words, categories, texts that is cognitively demanding. And as you might know, surveys are expensive. They cost a lot of money. So sometimes you see here to the left a picture of a digital survey. And this is a traditional survey in a digital way. But if you see this, it feels more similar to an exam than to an entertainment experience, which is what people expect when they're online in the TikTok era. So in this era, people will give you five minutes probably, but not too many as before. So if you see maybe market research or some customer experience research can be done with five or 10 questions. But when you're doing proper profound social research, you need to do a lot of questions. To even just in demographics, you need like 10 questions. Then you have several models. You need a lot of questions. So time is the key variable to having successful data collection in this era. So I was myself a social researcher for the last 10 years, and I used to suffer with this survey experience. It was frustrating to pay a lot of money, wait a lot of time because usually data collection took months to be completed. And sometimes information, the data collected was not satisfying. It was small samples, I mean, the biggest sample that you can pay, which is normally very small. So we were wondering if there's another way, and that's how Zsata was born at the fastest data collection platform for organizations and with a specific emphasis in culture and behavior research, which is the background that I come from. And I want you to experience by yourself our paradigm. This is your code. If you have your phone there, you can scan the code and do a sample, a very, very short survey. It won't take more than two minutes. If someone can't get the link, maybe we can share it here on the chat. Anna, are you ready to move on? You ready? OK, that's great. So I think that you were able to experience some of the features. And we will talk about that more in depth here. I like sharing because this is the way we started before the pandemic. We actually did data collection in person with this type of devices. And this is a project. It was an impact evaluation for children at school. It was a sustainable mobility project that we were evaluating. And the interesting thing here is that these type of interfaces are known by everyone. I mean, children from, I think, two year old to people 80 year old, they all know how to interact with Facebook, Instagram, TikTok, whatever, smartphones in general. So the big insight that we have is people are all day long clicking on a screen. Why if we leverage on that ability to collect data in an engaging, fun, and fast way? And it actually worked after the pandemic started in 2020. We launched. It was a survey about mental health during the pandemic. How do you feel? How does the lockdown affected your emotions and mental health? And it became viral. We didn't expect it to happen. But we got over 400,000 surveys of that single survey with no advertising, no intentional efforts for that to happen. And it showed us the opportunity, no, survey responses. And it was only one questionnaire. And we got 500,000 survey responses. Because people at the end of the survey, they were able to share it with people. And they did. So it was insightful. People can even find surveys of something valuable so that they share it with others. It was really mind blowing, because traditionally, people are reluctant to answer surveys since they do even have to pay them. So we started a new business model in that moment. And we do social research surveys which are not paid. I mean, people answer them for free. And we advertise on social media, or we use different advertising methods. And that's how we have collected these mass surveys so far. We have been working in South Africa, Nigeria, India, Latin America, and North America. We haven't worked in Australia or New Zealand so far. We hope so. We hope to. So I think that the magic of this is to be able to answer the most possible question in the least possible time. And we do this by reducing cognitive effort, limiting the number of options, reducing the number of words, and exploring number of question types. This is really challenging, but it's part of the, this is part of it. I think this question, maybe we can answer to it later when we address the security and the things. So the possibility of capturing non-verbal question types and information is part of the paradigm of social media and the opportunities that come from smartphones. Before in traditional surveys, it wasn't as easy because people were not familiar with emojis or these kind of symbols, but now everyone is. So now we can do this type of things. So regarding survey times, 90% of our users can respond up to 30 questions in three minutes. And that's because we focus on the automatic system. And you probably know this. Maybe you are familiar with Daniel Kahneman, the author of Think Fast, Think Slow, Novel Prize in Economics. And this whole idea of the two different systems, the file system and the slow system, we humans tend to put our 95% of our time, we are in the fast system. We like doing things that are simple, instinctive, effortless and conscious. This is totally different to what traditional surveys do. Traditional surveys are complex, require that people think, demand cognitive effort, logic analysis, and it normally takes more time. But not only the time, but people sometimes get bored. They quit the task, and that's why we have so high, I mean, no way, but social research are having very low response rates currently because people don't want to do that type of surveys anymore. And how do we do this? How do we keep users in the automatic system? How do we keep our surveys fast? This is part science and part art. Because we analyze the time per question of each of the questions. We have a very careful process of creating each question, start from the variable. Then sometimes when we even work with researchers, they have already maybe the questioners, but we help them to simplify the cognitive effort in the surveys. We simplify the number of words. We sometimes divide questions. For example, sometimes you have a multiple selection question. You have seven options, and people can choose all of them. But in our experience, it doesn't happen. It's very uncommon that someone chooses seven out of seven. Sometimes it's just laziness, or you don't know. But what we found, for example, is that it's faster in terms of the total time. If you have seven different questions, yes or no. Then if you ask people to choose among seven things, because you have to read the whole seven things first, then think if you have these or not. So what we found, for example, it reduces the time if you separate, divide the seven options in seven questions. This is just an example, but that's what we do. We look at the time and we try to keep every question below the threshold of 10 seconds per question. This is just an example here. So many researchers believe that doing online surveys, and this is more common every day now after the pandemic, everything is more virtual digital. And people believe that you just take your questionnaire, the one that you always use, and you just put it online and send it to people. But we believe that it should be actually totally different piloting, how you approach people who are already online, because the experience that they're having in every platform is totally different to what traditional surveys do. And I think this all belongs to the economy of attention and you're probably familiar with this concept. Today, everyone is competing for people's attention, not only products and services, but also NGOs, government news. Everyone is trying to capture people's attention. So we want to capture their attention under time to answer a survey. We need to talk and approach them in the language that they use. This is an example of what should not be done. This is an example of a survey that I got this email in March this year. It's like four paragraphs. I already spent like five minutes reading only the email to ask me to please answer a 20-minute survey. Even if they pay me, I think it's like, I don't want to read the email. I actually didn't read the email. I noticed that they were paying for answering the survey like one month too late. But anyway, so this is our paradigm, which is very different. We advertise something on one social media or on WhatsApp or the social network that people are already using. And then we get them through one experience like this that you already experienced yourself, which is very simple and engaging. For us, it's really exciting to see how people like, actually how can they interact with our platform for seven minutes, 50 clicks or 70 clicks. And it has allowed us to run very interesting research for NGOs, universities or private companies as well. Also, we created this concept of survey test or survey quiz is similar to the one you experienced. And sometimes not always, sometimes we just say like, hey, please help us answering these two-minute surveys. But many times we prefer to launch a question that is interesting or relevant for people to answer, like how healthy are you eating? Or do you know your financial profile or what's your risk of getting COVID? Something like this is a question that starts like a conversation. And this is what people find at the end like a symbolic incentive to spend their time, respond honestly and then even share it with other people because it was an interesting or fun experience and they also found valuable the profile that we gave them at the end. We built this profile specifically for every survey based on the answers of the participants. So if they, we normally have two absces and then we create at least nine profiles. So sometimes people really, really get value from our surveys. This is, well, I think in this part we can talk more about security data and Gabrielle is gonna do that. Yeah, okay, so I'm gonna start by answering some of these questions that I've seen on the chat. First, what are our data security policies? So we have the data is encrypted in our server and it's also transported using a secure layer. We usually keep the data for a period of time sometimes agreed with the customer for our own, having a backup, but that could also be agreed with the client in some sense if the client needs it to be erased one month or immediately after data collection is done that could also be done. And that's kind of like our policies. In terms of where our data is stored I'm checking with the development team but I think our server is in the US because we mainly work in Latin America that's where less lag we have. I'm not sure if we could develop something but I think it is possible to develop something where for that survey, it could go to a different cluster. I would have to check with the development and I'm still waiting for an answer on that. But yeah, we use US servers because that's what it's better for us since we mainly work in Latin America. Okay, there's another question that I could also answer right now, Mark asked for university type surveys. How do you shorten the ethics or in the US DIRB sort of forms, it depends on the university policies we have been able to agree on having one short paragraph and then a link that sends to a longer website that people can check, that short paragraph has to be decided with the university and that can also be done and set up according to the client's needs. It is something that in some cases reduces response rates but I think it's designed to do that because it is an agreement where people have to be aware of what's going on. So it sometimes increases the time that it takes to reach a minimum survey. Yeah, also sometimes the informed consent can be downloaded to the device where people are answering. And there's no question about time. How long it takes? Yeah, give me a second, a development just answered. We can configure to save the data in a different server but that has a cost in the terms that some features might not be available. So it could be done. I think the question was about how long it takes to develop. Yeah, I'm going back to the question on where the data is stored. So we can configure it to be stored in a different place but it has a cost in terms of what features might be available for the platform. In terms of how long it takes, it depends. But we've been able to launch things in like a couple of weeks when it's something short. And we have also been able to collect data pretty quickly, arrive at a minimum sample but that depends a lot on the questioner length on how clear the research is by the customer. So we also have projects that have taken two or three months between translating the questioner to our paradigm and then collecting data but we can actually work very quickly. Okay, so let's keep going. So real time monitoring, we provide this sort of dashboard that presents how the survey data collection is going in real time. So you can check this and see how many people are being collected, where are they coming from, how long is it taking, some of the questions. And this allow us to be maybe checking a quotas if that's something that's important for the customer, checking that we are actually getting the people that we want to get and it just gives some peace of mind and see how that's going. I think we can keep going Anna, please. Yeah, I think here it's interesting to see the numbers. This is a real survey that we did in April and you can see the dates. It goes from the 4th of April to the 28th almost we collected 29,000 surveys in that period. So it's very fast if you compare with traditional data collection times. We have also been able to, one big concern with this kind of data collection is data quality, of course. We need to be sure of compliance of non-manipulation and we've been able to develop some technology and algorithms that allow us to identify attempts of manipulation if there's a political topic. We've also been able to detect bots and with times we can also check if it's non-compliant or inattentive users. We've come to know how long a question should take for someone to actually read it and not just answer because they want to get to the end. So we have developed technology to ensure that we have high quality data because it is something that we really care about. We not only want to get fast data but we want it to be good too. Our platform not only includes the ability to collect regular surveys but we can include survey experiments which allows you to test casual hypothesis. So we could have experiments kind of like priming experiments where you give some information before and then randomize who see what information. We also can have AB tests of different messages. We can have choice experiments or discrete choice experiments or conjoined experiments that there's different language in different disciplines. We have that developed in our platform. So we can randomize the order of questions, the amount of response options if you want to do a list experiment. We can randomize a lot of things inside. Would you be sharing the comments in the chat for me? And this is really, I think it's a really useful feature in market research and in social research in general to not only have observational data but actual causal data. So we have here an example of a choice experiment. This is just an experiment that we did to be able to measure how much people is blaming different entities or institutions for not eating well. And we randomize the two options that people see and based on that, we can actually estimate who is to blame. We also, this would be a conjoined experiment with only one attribute but we can have a conjoined experiment with two or three attributes. There is some trade-off here, normally conjoined experiments. Normally conjoined experiments have like a lot of attributes. We try to keep it short because as you know, we like to people not having to think a lot and get bored and just quit our survey. So we try to keep just two or three attributes maximum and have very short descriptions. So there is a trade-off there but we do believe that it ensures higher quality data. Yeah, I think here the counterfactual of this type of question is that normally people, surveys ask people to order several things like how do you rank these five things? And that's a very difficult question for people cognitively to do. Like if I ask you rank these five types of food or rank your favorite, whatever, it's cognitively difficult. So this way we simplify the cognitive effort and at the end we have a much more interesting analysis of the information that is closer to behavior than just opinion. Yeah, so this allows us to rank something without having people actually having to rank that. And in this case, we were able to see that most people blame individuals for why they eat wrong. Yeah, Anna, do you want to answer those questions about the social media? I can't because I'm sharing a screen. Oh, okay. Sorry, but that's right. But you can keep going. Yeah, so here's another example of the two different messages and different designs that we tested. We can randomize who sees which message and which design and we can also have a control group that sees none of them and then ask a question about whatever is related to this. So in this case, it was a slogan used for pro-choice campaign where we have different kind of messages and then we asked people if they were supportive of abortion in different cases or in any case or never and we were able to see which message actually increased support for abortion policies. And this can be done for anything, anything that's graphic or text, we are able to randomize it as an A-B test. And just to conclude, we've come to realize that surveys have come to ask a lot from respondents, right? They expect a lot from the respondents and they don't give that much back in return. It's not very fun to do a survey. It takes your time and it takes effort. So we have come to develop a platform that makes it fun, that makes it fast and that makes it, actually in some cases it can actually give you something like saying, hey, you're not eating very well so how about you try to follow new habits? Or you have a financial profile that puts you at risk on something. So we give something back and that has ensured us to be able to collect a massive amount of data faster than traditional surveys. Making it fun, making it fat, making it smart for people is how we ensure this. Virality, we also have the ability for people to, you finish the survey and if you like it, you just hit a link and it will send you to WhatsApp or Facebook and you can post it in or send it to whoever you want. And that's how we've been able to create these viral surveys. Yeah. However, maybe. Oh, sorry. Sorry, Gaby. No, go ahead. No, I was just saying that this is, we have talked a lot about the viral and service through social media. However, we not only work on that platform. Sometimes this, this method and this platform also works for in person data collection for collection, their collection through email, because in general the promise is less time and a fun experience and it works. It have worked for us. It had worked for us in every latitude in several continents. So we trust very much this methodology. So that's what we, we had to share. Thank you very much for, for your time. I don't know if we, if you want us to go more in depth with some of the questions here. We have plenty of time for questions. So feel free to unmute yourself. And, and we can have a dialogue. Thank you so much. Anna, Anna Maria and Gabriel, that was a beautiful, beautiful presentation. And I can see why you have such a high response rate. So audience over to you. Please put your hand up so we can select. Oh, Mark, Mark, would you like to unmute yourself and, and ask the presenters. Sorry. Yeah. Well, currently we, we don't have like a software service. Platform that you can manage for yourself. Not, not yet. We have a service in which you, you provide us the questionnaire we, we program it in a couple of days. Then we run a pilot, which usually takes about a week. And after the pilot, we, we can start the data collection process. Then it, depending on how big the sample, who are the responses, what type of relationship you have with them. If they're not regular citizens or your clients, it's a, it's different. The data collection times, but I can tell you that with, we run a survey in Chile for the clients of a bank. They, they send the link to 44,000 people and they got 9,000 surveys in a week. So this is an example of, of what is possible. Also through social media, we know that it can take between a week and month, depending on the number of survey responses that you want. And here there's another question about sampling methods. Any online question survey, sorry, any online survey is not a random probabilistic survey, you know. So what we do to have a proper sampling is that we control quotes or like, can you please answer this question? Yeah. So when, what we do is, of course, it depends on the customer and the research. Sometimes if it's a causal research, they don't need a representative sample. And as Anna said, an online survey, it's going to be self-selected. And that's the case if you use this platform, or if you use SurveyMonkey, or if you use your email or whatever, or Google Forms. But what we can do and we do is that we put minimum quotas to ensure the distribution of key demographics. And we can also do positive certification weighting that allows us to have a similar distribution on demographics from our sample and the country or whatever geographical unit you have. Again, we don't only do online surveys, although we haven't done anything in, you know, the Asia-Pacific region physically. We've only collected data online, but what we've done in Colombia is we can do sampling, regular sampling methods with our platform. And what we also have done with one customer is that we provide the training and the design of our survey and then you guys could provide the devices and the people to collect data physically wherever you need it as a household survey or as an interception survey. So that could also be done. There was another question. Yeah, about the quotas, because I think it was a misunderstanding because we don't collect, we don't do the pre-demographics before the survey on Facebook. We normally advertise to everyone and then we control the quotas in this dashboard. And we monitor how many adults or any age group or sex or whatever quota you want to control. We do this while the data collection process. Also it's very interesting that sometimes we have target people not using Facebook service of targeting, but through communication. For example, we launch an app for everyone in a region, but the message is if you have been diagnosed with depression, or Alzheimer's dementia, or any other mental illness, please tell us how has been the experience during the pandemic. That was a real life example. We didn't have to filter in Facebook, please show this only to people who have mental illnesses, but through communication, people start to select themselves if they are interested. You had a question about SMS survey data. Could you clarify what that means? What is SMS survey data? Is this manner? Would you like to unmute yourself? That's okay. If you can just answer my question through the chat, maybe I could answer better your own question. But in the meantime, what we have done also in the past is that people, clients through their communication means if they have an SMS, through the SMS they send a link to the survey and people can access our platform and answer the survey there. I'm not sure if that answers your question. I'm not sure if you're thinking of these surveys that are like inside the SMS where you answer yes or no inside, and that's different. Okay, yeah, so we could distribute that. Still, if there is no 3G or 4G, our platform does require some internet connection. So the first moment of entry will take longer if it's a slow internet connection. When you did it, I'm assuming it didn't take longer than, I don't know, 10 seconds, maybe 30 seconds if your internet was not that good. If you're doing it in a 2G network, it will take longer. I'm not sure how much, but it will take longer and it will also take longer to send at the end when you're processing. It's going to take a little bit longer on sending. So that would be a problem. But you can use SMS to distribute the survey. The problem would be answering the survey. Yeah, but also sometimes in regions where, for example, you have a social project in a rural area where there's no internet. What we've done is that we can install some devices which have a platform in the offline version, because we have an offline version of the platform that can collect data with no internet at all. So you can deliver some smartphones to the project or the region and there people can report monitoring surveys or whatever data you need. Yes, and then they would have to come back and send it to our server. Yeah, at some point they have to come back to internet and upload everything. That's correct. Oh, sorry. I'm trying to raise my hand on the chat feature, but I can't. So I'm just going to speak. Is that all right? Sure. Okay. So Anna or Gabriel, thank you so much. This is such an excellent presentation and very interesting. I have a couple of questions. Firstly is can you put your URL on the chat so that I can, because I'm trying to find you online and I'm not sure if I've got the right website, but also the second one is in terms of cost. I'm really keen to find out like roughly how much would it cost. I know there's probably, we can probably get a quote or something for something specific, just in general, if you could give us an idea, because I'm working for a government organization where we have a cost recovery model. And so when we are costing our evaluation or research projects, for instance, I need to be able to tell them this is how much it's going to cost and this is how many days it's going to take. So if you can comment on that, that would be super helpful. Thank you. Well, it's really complicated because it depends on the sample size and the way you're getting to people. If you already have their emails or their phone numbers, it might be cheaper than if we have to go out to social media to collect them. And as we haven't worked in Australia or New Zealand before, we don't know exactly how much it's going to cost. But I can tell a project can range from like $1,000 to $50,000 or $100,000 depending on the size of the project. So it's very flexible. And we always figure out ways to work very fast because that's part of our philosophy. We're like a mixture of researchers and IT people. And we believe that research should learn a lot of, like, startups and technology because they work faster, they work lean, and that's how we used to work. Yeah, thank you. Super helpful. It's very expensive, but that's all right. But yeah, it depends on things like question or length, survey size, recruitment method. Of course. So if you want to have a quote, you just come to email Anna Maria there. Yeah. And we'll figure something out. Yeah, great. Thank you. That's benefits. I think I'm just going to add one other comment. I think the other thing too is good quality data doesn't come cheap. No, of course. Just as a general, you know, with anything. And that's always a challenge when we talk to these commissioners. If you want decent data, you're going to have to pay for it. That's right. You'll get what you pay for. Yeah. Yeah. Thanks, Scott, for saying that. I have a, I have a question. Have you used open ended questions in any of your surveys? And then do you do like coding of that? And if so, how has that gone. On your platform. It's a great question. Yeah, we have, we have done that. That's very. Sometimes it's interesting what has happened. Usually we do this for the option other, you know, this question when they have this or that or other. And then what, what other. And it's, it's interesting to, to know what people are thinking about other. And we normally use like this. You know, you know, You know, you can use the qualitative and structural methodologies to, to analyze that. But you can't, you can't. See this information in the, in the. In the dashboard. It's not a specialty to do, to collect that type of information. But if I could answer a follow-up question to. something I was thinking through earlier is there are so many different ways that you can approach how you structure a survey depending on what exactly it is you're trying to find out. And so sometimes you have a very clear understanding of what you're dealing with, so it's quite easy to segment and identify a number of quite specific questions in the manner that you were describing earlier. I was wondering if there have been projects or contexts or situations that haven't lent themselves so well to this. I'm thinking we might be a little bit more inductive discovery type work that would often, you know, which can often happen with research as well. Yeah, I'm keen to know about that. Yeah, if you want and I can answer this. We have had cases where we have to come back and forth between the client in what they want, but also sometimes we have to go and do more than one pilot survey. So we run one pilot survey and we realize that a lot of people are skipping a lot of questions, for example, that we're not capturing what we want to capture because everyone is like answering one or two of the options and we're not. And we have had to do that inductively. And I think you're right that that's just sometimes part of the research process when there is something that is complex and not so clear in terms or maybe it is clear, but it's not so easy for the respondents to answer, you know, there's different reasons what that could happen. But that's why for us the pilot stage of the process is so important. We never skip that. We have always done it and we take very seriously the analysis of the pilot data because we believe it's key to get good quality data to get to have a good pilot. Thanks, Gabrielle. Looks like Marie Mott have accidentally exited the meeting. Yeah, maybe she's having connection issues. So I don't know if anyone else has any other questions. We are really looking forward to hearing more from any of you and I already. I think Marie is back. Yeah. Yeah, do not hesitate to ask. We can, we will be happy to think about your project. As I told you before, we're not yet a software service, maybe in the future we will be. But also we love like thinking about how can we leverage current technology to improve research and we really enjoy helping our clients to find out better solutions to their problems. So really thank you for this space. Marie, you're back. I'm so sorry. My internet connection, this is one of my nightmares to have my internet broke off during a meeting before it was an international seminar. I was just going to say as well, it was lovely. I mean, I've seen a lot of people saying what a great presentation it is. And I agree. And I think one of the things that's a challenge for all of us who you know, got deep roots and commitment to research practice, knowing that we have to move with the times and some of the I suppose what we've been taught and we hold very close and aren't really, you know, find really hard to let go of. It was, it was, I enjoyed listening to how you guys have clearly wrestled with that and have figured out what to let go of or just to get more flexible with. But, but still, you know, understanding that there's there's still that credibility and robustness that needs, you know, you can't sort of throw the baby up with the bathwater. Yeah, so it was, I loved listening to that. Whereas, you know, a typical marketing firm, I'm always a bit skeptical of because it tends to be the easiest path, maybe not necessarily the best. And that's, you know, not to be unkind. But yeah, so I really enjoyed listening to your particular perspective in that way. Thank you, Cara. I'm glad to hear that. I would like to, sorry, Gabriel, go ahead. Go ahead. I was just gonna thank you, Cara. I was gonna say that we've been able to using our new paradigm, we've been able to replicate in Colombia, for example, similar findings than our census bureau on some topics, right? So it's not, it is, we do something different, but we do believe in good data quality. And we do think it is important to actually capture what we want to capture. So we don't, we don't, we, we change things, but we don't cut quality of our data. Are there any more questions for Anna and Gabriel? There, you have a lot of new fans now, Anna and Gabriel. Yeah. We have to think to move closer to Australia and New Zealand. Everything is possible virtually. Thank you. Thank you so much, Anna Maria and Gabriel, for your brilliant presentation. I found it extremely informative and it was a pleasure to watch, you know, it was a pleasure to, to understand your process, your data storage mechanisms, how you define things display on someone's smartphone. I think it's ingenious. And also congratulations for your successful startup during COVID. You know, this is quite a pioneering kind of effort to have a successful business during lockdown and have it grow through the, through the months. So congratulations for, for your brilliant achievement and we look forward to staying in touch. And this is, thank you everyone for, for attending and for your thoughtful and intelligent questions. It really added some very good value. And yeah, yes, this, this is recorded and we will put it on YouTube. It normally takes about a month to upload it, but we'll try to get it up sooner and we will circulate the PowerPoints as well. So thank you so much everyone.