WEBVTT - How Traders Used Google Searches To See The Economic Recovery In Real Time 0:00:10.960 --> 0:00:14.360 Hello, and welcome to another episode of the Odd Lots podcast. 0:00:14.400 --> 0:00:19.479 I'm Joe Wisenthal and I'm Tracy Hallaway. So I don't 0:00:19.480 --> 0:00:22.720 know what day people are going to be listening to 0:00:22.840 --> 0:00:26.240 this episode, but um, you know, the stock market hit 0:00:26.239 --> 0:00:29.760 a record high yesterday. Yeah, it's true. So all the 0:00:29.800 --> 0:00:33.040 losses that we saw during the COVID crisis have basically 0:00:33.080 --> 0:00:36.720 been raised and markets are back where they were before 0:00:37.040 --> 0:00:42.920 all of this happened. Yeah, it's essentially six months from 0:00:42.920 --> 0:00:45.280 the pre crisis peak to this one. So I think 0:00:45.320 --> 0:00:49.479 the SUP peaked at um February on February fifteen, and 0:00:49.520 --> 0:00:53.640 then we saw the the new peak yesterday, August eighteen. 0:00:54.160 --> 0:00:57.920 And in a sense, it really feels like we've compressed 0:00:57.960 --> 0:01:01.760 this sort of gigantic cycle into an extremely short period 0:01:01.760 --> 0:01:04.160 of time. Yeah, that's true. And I was looking at 0:01:04.200 --> 0:01:07.760 the latest fund managers survey from Bank of America and 0:01:07.800 --> 0:01:11.560 it showed that I think fund managers have completely flipped 0:01:11.959 --> 0:01:15.040 from thinking that we're in a recession to thinking that 0:01:15.080 --> 0:01:17.959 we're in the early stages of a fresh economic cycle. 0:01:18.040 --> 0:01:21.319 And if they're right to your point, it does suggest 0:01:21.319 --> 0:01:23.280 that we've just seen, you know, one of the shortest 0:01:23.360 --> 0:01:26.600 recessions of all time. Yeah. I mean, you could make 0:01:26.640 --> 0:01:29.360 the argument that the recession recession in terms of the 0:01:29.440 --> 0:01:31.640 shrinking of growth was done by the end of March, 0:01:32.000 --> 0:01:35.880 when most data points started turning up. And while the 0:01:36.040 --> 0:01:39.319 overall level of economic activity is still very depressed, and 0:01:39.360 --> 0:01:42.560 of course unemployment rate is still above ten percent, so 0:01:42.640 --> 0:01:46.720 hardly time to be declaring victory. We have seen steady 0:01:46.760 --> 0:01:50.360 improvement on a host of economic data points basically since 0:01:50.520 --> 0:01:54.559 end of March early April. That's true, but I also 0:01:54.600 --> 0:01:57.400 feel like there's something kind of weird going on with 0:01:57.440 --> 0:02:01.120 the data. Like there's the old stock first flow argument, 0:02:01.200 --> 0:02:04.320 which we're seeing everywhere, but particularly in p M I. 0:02:04.480 --> 0:02:07.240 So even when we get a big rebound in p 0:02:07.440 --> 0:02:09.720 m I s, it doesn't necessarily mean that we're getting 0:02:09.720 --> 0:02:13.320 back to the levels that we saw pre crisis. But 0:02:13.400 --> 0:02:17.560 you're also seeing just sort of weird indicators that are 0:02:17.600 --> 0:02:20.600 happening simultaneously. And I think one of our colleagues pointed 0:02:20.600 --> 0:02:23.880 out a really good one recently, and that was intentions 0:02:24.000 --> 0:02:27.520 to buy a house surging at the same time as 0:02:27.600 --> 0:02:32.520 mortgage delinquencies, which I mean never happens in an economic crisis. 0:02:33.000 --> 0:02:35.720 Now it's really weird. But I think because of all 0:02:35.760 --> 0:02:38.880 the weirdness that we're seeing this sort of contrary indicators, 0:02:39.240 --> 0:02:43.760 because there's this weird gap between pieces of change which 0:02:43.800 --> 0:02:46.840 have been very fast and unexpected versus levels which are 0:02:46.840 --> 0:02:49.320 still very bad levels. And then also just the fact 0:02:49.320 --> 0:02:53.359 that it's so compressed, there's probably never been more demand 0:02:53.680 --> 0:02:57.080 for sort of alternative real time data points and this 0:02:57.200 --> 0:02:59.919 feeling that the official economic data points that we get 0:03:00.360 --> 0:03:04.800 monthly jobs report, monthly retail sales report, they just there's 0:03:04.880 --> 0:03:07.560 not enough of them. They're not timely enough to get 0:03:07.560 --> 0:03:10.360 a sense of what's going on, given how fast the 0:03:10.480 --> 0:03:14.040 changes have been both on the downturn and the rebound. Yeah. Absolutely, 0:03:14.080 --> 0:03:16.320 and I mean, just on a very simple basis, everyone 0:03:16.360 --> 0:03:19.040 wants to know what's going on with the recovery, right, 0:03:19.080 --> 0:03:22.760 and everyone's tracking to what degree the economy has reopened, 0:03:22.840 --> 0:03:25.799 and some of the most useful indicators for that are 0:03:25.919 --> 0:03:31.760 arguably alternative economic indicators like um like open table reservations, 0:03:32.520 --> 0:03:35.040 things like that. Yeah, totally. I mean, that's like one 0:03:35.080 --> 0:03:36.960 of the things we've been watching the most. It's like 0:03:37.400 --> 0:03:41.280 open table they could keep track of people making reservations 0:03:41.400 --> 0:03:43.560 or doing in seeding dining, so if you want to 0:03:43.560 --> 0:03:47.360 sort of understand how behavior has changed or how people 0:03:47.440 --> 0:03:51.200 are doing different things. Um, due to the virus, that's 0:03:51.240 --> 0:03:53.800 been one of the sort of key data points, not 0:03:54.040 --> 0:03:57.840 something that people were really tracking before as far as 0:03:57.840 --> 0:04:00.440 I know, on a meaningful level. So I think that's 0:04:00.480 --> 0:04:03.640 really important. I mean, I think obviously real time alternative 0:04:03.720 --> 0:04:07.600 data has never been more in demand than what we've 0:04:07.640 --> 0:04:10.440 seen over the last six months. But I don't think 0:04:10.480 --> 0:04:12.680 it's going away now. It's kind of another one of 0:04:12.720 --> 0:04:16.560 these things where real time data points of a range 0:04:16.560 --> 0:04:18.920 of things will sort of be part of the conversation 0:04:19.080 --> 0:04:22.240 for a long time, even if and when we get 0:04:22.240 --> 0:04:25.400 back to something resembling a normal economy. Yeah, I think 0:04:25.400 --> 0:04:28.520 that's right. So today we're going to be talking all 0:04:28.560 --> 0:04:32.599 about alternative data, what it's showing, and more importantly, how 0:04:32.720 --> 0:04:36.840 investors actually use it in their process. And so we're 0:04:36.880 --> 0:04:39.200 going to be speaking with Ben Brightholtz. He's a data 0:04:39.320 --> 0:04:42.880 scientist at Arbor Data Science, which is part of Arbor 0:04:42.960 --> 0:04:46.359 Research and Trading. I've been following their stuff. They do 0:04:46.400 --> 0:04:51.000 some really interesting things with looking at Google search trends 0:04:51.040 --> 0:04:54.640 for lots of different keywords and trying to divine an 0:04:54.720 --> 0:04:58.359 economic significance from them. So let's talk more about that. Ben, 0:04:58.360 --> 0:05:00.520 thank you very much for joining us. Yeah, thank you 0:05:00.640 --> 0:05:03.039 very much. Joe, happy to be here. So let's just 0:05:03.080 --> 0:05:05.919 start a big picture. What do you do? What is 0:05:06.120 --> 0:05:09.720 arbor data science? Talk to us a little bit about 0:05:09.760 --> 0:05:15.320 your work. Sure, So over the years we've gotten more 0:05:15.360 --> 0:05:18.920 and more into essentially this idea of filling the gaps 0:05:19.000 --> 0:05:22.600 between latent economic data and the econ data that can 0:05:22.600 --> 0:05:25.680 be distorted like we've seen with unemployment data as of late, 0:05:25.920 --> 0:05:29.279 and also really trying to help our customers and the 0:05:29.320 --> 0:05:32.720 investment space in general deal with surveys that have been 0:05:32.839 --> 0:05:36.080 more or less leading indicators for quite some time. They've 0:05:36.160 --> 0:05:38.159 kind of fallen flat on their face. And this is 0:05:38.160 --> 0:05:42.280 something that's taken place well before UM even the current 0:05:42.279 --> 0:05:46.120 episode we're going through now, looking back to the financial crisis, 0:05:46.120 --> 0:05:51.200 with really the polarization of the country and the world 0:05:51.279 --> 0:05:54.279 on a political space, and really the advent of social 0:05:54.320 --> 0:05:59.240 media has created really this bifurcation and in sentiment it 0:05:59.279 --> 0:06:01.480 could be republic and democrat or it can be more 0:06:01.560 --> 0:06:05.640 or less group think. Based on UM, the use of Facebook, Twitter, 0:06:05.800 --> 0:06:09.039 we create all these small microcosms we essentially live within 0:06:09.520 --> 0:06:13.400 and that is ultimately distorted the ability of survey data, 0:06:13.440 --> 0:06:16.840 for example, to have this leading nature that it used 0:06:16.839 --> 0:06:19.720 to have really for decades um. And that's posed a 0:06:19.760 --> 0:06:22.880 significant problem for investors that are in putting this either 0:06:22.920 --> 0:06:25.640 on a subjective level or within their own modeling to 0:06:25.760 --> 0:06:28.200 then project board where they think financial markets will go 0:06:28.600 --> 0:06:31.600 in the future. I have a really basic question, which 0:06:31.640 --> 0:06:35.240 is what's the difference between big data and a large 0:06:35.400 --> 0:06:39.720 set of data? So big data is such a misnomer 0:06:39.839 --> 0:06:42.920 and um nasty term, you know most I think big 0:06:43.000 --> 0:06:45.920 data is a term that's kind of slowly gone away 0:06:46.520 --> 0:06:49.000 that I think the initial idea is that it's it's 0:06:49.120 --> 0:06:53.280 unstructured data. That's for example, you can find all this 0:06:53.360 --> 0:06:56.680 wonderful information on a Bloomberg terminal, all right, and it 0:06:56.800 --> 0:06:59.279 comes you can download it via a p I or 0:06:59.320 --> 0:07:02.680 access at via via your your Windows or your terminal, 0:07:02.760 --> 0:07:05.840 all nice, clean and easy to use, ready to input. 0:07:06.360 --> 0:07:09.320 And big data um to me. Uh, this day and age, 0:07:09.400 --> 0:07:11.680 especially with alternative data, has to do with more or 0:07:11.760 --> 0:07:15.480 less unstructured kind of ugly data. So this, for example 0:07:15.520 --> 0:07:18.120 could be all just like us talking right now, or 0:07:18.160 --> 0:07:20.280 when you are all on TV, you have all of 0:07:20.280 --> 0:07:23.600 this this text, this closed captioning that exists out there, 0:07:23.800 --> 0:07:26.520 and let's say, for example, it's in fifteen second increments, 0:07:26.880 --> 0:07:29.480 and it can be ugly, it can be have plenty 0:07:29.480 --> 0:07:32.480 of errors within the data within the closed captioning um 0:07:32.560 --> 0:07:37.080 And essentially we have to use algorithms and different processes 0:07:37.160 --> 0:07:39.320 in order to take that unstructured data and make it 0:07:39.440 --> 0:07:42.280 something useful and really turn it into something that's more 0:07:42.320 --> 0:07:46.280 or less numerical in order to benchmark against financial markets, 0:07:46.320 --> 0:07:49.760 econ data, overall sentiment and so on. So, you know, 0:07:50.200 --> 0:07:52.920 big data is kind of a word. I think that's 0:07:52.960 --> 0:07:55.559 somewhat going away. But to me, again, it means somewhat 0:07:55.600 --> 0:07:59.600 of an unstructured data set. So I'm thinking about what 0:07:59.680 --> 0:08:04.600 you scribed as the problem with surveys, and uh, you know, 0:08:04.760 --> 0:08:07.840 I think it's either the University of Michigan Consumer Sentiment 0:08:07.880 --> 0:08:10.640 survey or the conference board one. There's one of these 0:08:10.920 --> 0:08:13.000 data points that we have it on the Bloomberg terminal, 0:08:13.280 --> 0:08:15.080 and it's like they say, it is now a good 0:08:15.120 --> 0:08:17.800 time to buy a washing machine? Is now a good 0:08:17.840 --> 0:08:20.360 time to buy a car. There's even one that's one 0:08:20.360 --> 0:08:22.080 of my favorites. Is now a good time to buy 0:08:22.200 --> 0:08:25.880 a vacuum cleaner. But I guess what you're doing is 0:08:26.000 --> 0:08:28.280 you don't have to ask people is now a good 0:08:28.320 --> 0:08:31.760 time to buy a vacuum cleaner, because in if you 0:08:31.840 --> 0:08:34.200 know how to find the data, you can just look 0:08:34.200 --> 0:08:37.600 at searches for vacuum cleaners and that's presumably a lot 0:08:37.640 --> 0:08:40.160 more reliable than asking people into survey whether or now 0:08:40.240 --> 0:08:42.880 is a good time to buy a vacuum cleaner. Right, So, 0:08:42.920 --> 0:08:46.599 the within surveys, there's and there's a plenty of studies 0:08:46.800 --> 0:08:51.319 on this as of late showing that respondents will not 0:08:51.520 --> 0:08:56.080 provide really honest answers relating to their financial hardships. So 0:08:56.120 --> 0:08:59.360 there's there's large gaps and you know, our things better 0:08:59.480 --> 0:09:02.280 now or worse? Are you going to spend do you 0:09:02.320 --> 0:09:05.319 have the money to spend here moving forward on a vacuum, 0:09:05.400 --> 0:09:08.000 on a new washing machine? And so on? And there's 0:09:08.000 --> 0:09:11.079 always been a gap, for example, example, between the web 0:09:11.120 --> 0:09:14.320 based responses and phone based and we saw this too 0:09:14.320 --> 0:09:17.160 with the election. That's a whole another other topic, but 0:09:17.559 --> 0:09:20.760 um on a web based survey and individuals are typically 0:09:20.840 --> 0:09:24.679 much more honest than they are regarding financial hardship than 0:09:24.720 --> 0:09:28.080 they are on the telephone or basically being put on 0:09:28.120 --> 0:09:33.320 the spot. So the idea here between behind search activity 0:09:33.360 --> 0:09:35.400 and this is something that I think that has improved 0:09:35.960 --> 0:09:38.320 in most recent years, is yes, we can get ahead 0:09:38.400 --> 0:09:42.160 of this intention of consumers and we're not necessarily we're 0:09:42.160 --> 0:09:45.679 not really going to lie to that little window on Google. Um. 0:09:45.720 --> 0:09:48.360 You know, we might lie maybe sometimes to our girlfriends 0:09:48.480 --> 0:09:51.840 or our boyfriends or husbands or wives. Um, but you know, 0:09:51.880 --> 0:09:54.160 what we put into that search window is really truly 0:09:54.200 --> 0:09:56.079 what we're seeking and what we're actually trying to query. 0:09:56.120 --> 0:09:59.280 There's no no one really looking over our shoulder. So 0:09:59.400 --> 0:10:02.840 our belief is that search activity, um, really, over the 0:10:02.880 --> 0:10:07.160 past five six years has become kind of a great 0:10:07.640 --> 0:10:11.280 estimate or indication of the consumers intentions of what they 0:10:11.320 --> 0:10:14.280 plan to do. Am I going to buy a wash machine? 0:10:14.400 --> 0:10:16.280 Or if I'm in distress, what does it mean that 0:10:16.360 --> 0:10:19.400 by default on my credit card payment or I don't 0:10:19.400 --> 0:10:21.360 pay my credit card payment? Or what if I need 0:10:21.400 --> 0:10:23.839 to go out and search and find a bankruptcy lawyer. 0:10:24.160 --> 0:10:26.440 These are the type of things we can pick up on, uh, 0:10:26.760 --> 0:10:29.480 you know, within this information to then create a kind 0:10:29.480 --> 0:10:32.800 of um, you know, overall look at the consumer. And 0:10:32.800 --> 0:10:34.360 this can be all the way from the you know 0:10:34.480 --> 0:10:37.000 up towards the United States, the complete um, you know, 0:10:37.040 --> 0:10:39.640 country level, it can be worldwide, and it can be 0:10:39.720 --> 0:10:42.880 drilled down all the way down to a metropolitan area UM. 0:10:43.200 --> 0:10:45.679 And again, the whole idea there is trying to get 0:10:45.679 --> 0:10:49.240 the most honest representation of the individual. And I'll also 0:10:49.320 --> 0:10:53.440 say that the growth UM in the Internet and really 0:10:53.520 --> 0:10:56.360 access to the Internet, both mobile and on the PC, 0:10:56.559 --> 0:10:59.560 has been a big boon for search activity, so that 0:10:59.600 --> 0:11:04.319 you now have fifty of the world having Internet access 0:11:04.559 --> 0:11:06.679 and using it on an active basis. That's more than 0:11:06.720 --> 0:11:10.120 four and a half billion individuals, which has really doubled, 0:11:10.120 --> 0:11:12.439 if not tripled, since the financial crisis. So I think 0:11:12.480 --> 0:11:16.480 early efforts of using search activity UM is, I know, 0:11:16.520 --> 0:11:19.080 a lot of it pre crisis kind of fell on 0:11:19.160 --> 0:11:21.480 its face and kind of faded away. Google used to 0:11:21.520 --> 0:11:24.520 have these curated indices um. I think they had twenty 0:11:24.600 --> 0:11:27.160 five of them, kind of showing how the economy, economy 0:11:27.280 --> 0:11:30.360 was moving um here and there. I think that that 0:11:31.559 --> 0:11:33.920 what didn't work as well because we didn't have the 0:11:34.000 --> 0:11:37.520 ubiquity of Google searches and really Internet access. And as 0:11:37.559 --> 0:11:41.480 that improves, this type of information becomes that much more important. 0:11:41.640 --> 0:11:45.200 I think to the investing process. How much do you 0:11:45.200 --> 0:11:50.000 think the the unusual or the extreme circumstances surrounding the 0:11:50.080 --> 0:11:55.319 coronavirus crisis are are distorting survey responses? And I asked 0:11:55.320 --> 0:11:57.760 that because again, I've seen a lot of criticism of 0:11:57.800 --> 0:12:00.040 the p m I s recently, and one of of 0:12:00.160 --> 0:12:03.880 things people are saying about those surveys at the moment 0:12:04.080 --> 0:12:08.040 is that respondents aren't really judging their experiences on a 0:12:08.080 --> 0:12:11.440 month to month basis, but they're sort of responding by 0:12:11.480 --> 0:12:16.000 comparing now to a period of relative normality. So everything's 0:12:16.040 --> 0:12:19.240 getting skewed. Do you think the unusual nous of of 0:12:19.240 --> 0:12:24.240 our current circumstances might be skewing survey data as well? Yes, 0:12:24.440 --> 0:12:26.440 I think so. I think it's it's a combination. Like 0:12:26.480 --> 0:12:29.640 you said earlier with stock flow, it's what type of 0:12:29.760 --> 0:12:32.600 reaction have we had over the past couple of months. 0:12:32.640 --> 0:12:35.920 UM I think has been more reflective within the survey data, 0:12:35.920 --> 0:12:39.199 and we're seeing that breakdown between surge activity UM and surveys. 0:12:39.200 --> 0:12:41.679 And we also have this big group think are almost 0:12:41.720 --> 0:12:44.680 circular reference that occurs within a lot of the sentiment data. 0:12:44.679 --> 0:12:47.040 So we all look to the equity market. We all 0:12:47.040 --> 0:12:49.640 know that we can use the equity market essentially forecast 0:12:49.720 --> 0:12:53.640 where consumer a confidence will be for the next month UM, 0:12:54.200 --> 0:12:56.720 and I think a lot of that's feeding into some 0:12:56.800 --> 0:12:59.839 of the more rosy consumer confidence numbers as well as 0:12:59.840 --> 0:13:03.280 the UM eyes and again that's some somewhat of distortion 0:13:03.559 --> 0:13:06.560 UM and why we seem to't like to rely on 0:13:06.600 --> 0:13:25.240 the search activity for the most part. So let's talk 0:13:25.280 --> 0:13:28.960 a little bit more about that search activity. How do 0:13:29.080 --> 0:13:31.960 you take how do you get the data? First of all, 0:13:32.080 --> 0:13:34.920 what does Google make available? And then how do you 0:13:35.040 --> 0:13:38.640 present it in a form so that it's usable because 0:13:38.640 --> 0:13:43.120 there's obviously seasonality factors the you know, you can't just 0:13:43.200 --> 0:13:45.959 look at searches for a vacation and see whether they 0:13:45.960 --> 0:13:48.319 go up or down because people don't vacation at the 0:13:48.360 --> 0:13:51.320 same uh at the same pace all year round. So 0:13:51.400 --> 0:13:53.640 how do you get the data from Google? What's that 0:13:53.720 --> 0:13:56.240 process like? And then what do you do to actually 0:13:56.280 --> 0:13:59.440 put it in a format such that it's not just 0:14:00.000 --> 0:14:04.079 OLiS for investors, like just describe it overall? How work? Sure, 0:14:04.320 --> 0:14:07.000 so we are able to access just like anybody else 0:14:07.120 --> 0:14:10.040 via Google Trends, which there is an API to be 0:14:10.120 --> 0:14:13.480 able to grab that information, and what we do is 0:14:13.520 --> 0:14:16.280 we avoid using the specific search terms. So if we're 0:14:16.320 --> 0:14:19.240 just going to say wash machine or vacuum UM, that 0:14:19.280 --> 0:14:23.120 will include specifically that exact term UM. And we know 0:14:23.160 --> 0:14:26.680 that there can be multiple variations of those actual text terms, 0:14:26.880 --> 0:14:28.360 and so we want to pick up on that. The 0:14:28.400 --> 0:14:32.240 beauty is Google curates and creates two different types of 0:14:32.320 --> 0:14:36.280 groupings of search activity. And they do this for you know, 0:14:36.440 --> 0:14:38.840 each and every country essentially, which is going to take 0:14:38.880 --> 0:14:41.480 care of the major language barriers and issues that we'd 0:14:41.520 --> 0:14:44.920 run into as well. And so that is they create categories, 0:14:45.640 --> 0:14:49.760 which there are roughly und forty plus different categories, everything 0:14:49.800 --> 0:14:52.760 from accounting services all the way out to urban transportation 0:14:52.760 --> 0:14:55.440 which would be things like uber and lift. And then 0:14:55.480 --> 0:14:59.480 they have topics and that can be anything from inflation 0:14:59.840 --> 0:15:03.480 or those talking about disinflation, or gold bugs or bitcoin um. 0:15:03.520 --> 0:15:05.960 And that's going to then be more encompassing and based 0:15:06.000 --> 0:15:09.720 on their mapping of a numerous new it could be hundreds, 0:15:09.800 --> 0:15:13.000 if not thousands of thousands in certain cases, of different 0:15:13.000 --> 0:15:17.040 search terms and phrases that then get housed underneath those 0:15:17.080 --> 0:15:20.680 individual UM topics. We can I stop you and ask 0:15:20.680 --> 0:15:23.080 you a quick question right there, Sure the data the 0:15:23.160 --> 0:15:26.360 year able to draw. Just to make clear, is that 0:15:26.440 --> 0:15:30.840 the granular within those hundreds or thousands of terms, you're 0:15:30.880 --> 0:15:32.640 able to get data for each one of those. You 0:15:32.640 --> 0:15:35.760 could see beyond just the sort of general category. Yeah, 0:15:35.880 --> 0:15:37.640 so we can drill down. There are ways to drill 0:15:37.680 --> 0:15:41.320 down within the individual categories. We understand what the actual 0:15:41.400 --> 0:15:45.760 searches are within those categories, but in order to create 0:15:45.800 --> 0:15:50.600 a more encompassing indication of what the consumer business is 0:15:50.640 --> 0:15:53.960 looking for or thinking about, we do then pull in 0:15:54.040 --> 0:15:57.960 that search trend. Essentially, that's going to be an aggregation 0:15:58.280 --> 0:16:01.520 of all of all those searches underneath a given topic 0:16:02.040 --> 0:16:05.400 or underneath a given category. And like I said that, 0:16:05.520 --> 0:16:08.080 one of the greatest things about the way that Google 0:16:08.160 --> 0:16:10.560 set this up is that you are then able to say, 0:16:10.640 --> 0:16:14.920 let's look at urban transportation uber and LIFT, and let's 0:16:14.920 --> 0:16:17.160 look at it not just here in the US. Let's 0:16:17.200 --> 0:16:21.600 go to UH somewhere like Germany, let's go to Australia, 0:16:21.720 --> 0:16:25.480 or let's go to Japan UM and they take care of, 0:16:25.600 --> 0:16:29.000 fortunately a lot of the language barriers in that urban 0:16:29.000 --> 0:16:33.320 transportation that is translated into into um, you know, Japanese 0:16:33.440 --> 0:16:37.360 or um you know whatever is being German, and so on. UM, 0:16:37.560 --> 0:16:40.640 so getting into the course of how we then digest 0:16:40.720 --> 0:16:42.880 and use that information. Like you said, is there's a 0:16:42.920 --> 0:16:45.280 high degree of seasonality. Of course, it could be like 0:16:45.320 --> 0:16:48.040 with clothing with back to school, or can be accounting 0:16:48.080 --> 0:16:51.680 services coming into March, April and October. UM. So we 0:16:51.760 --> 0:16:56.480 do decomposition where we'll we'll break down each individual topic 0:16:56.760 --> 0:17:00.520 or categories search activity into three components, and that is 0:17:00.640 --> 0:17:02.720 it's overall trend component. You think of it as like 0:17:02.760 --> 0:17:06.879 as kind of a slower moving average trend of that 0:17:06.920 --> 0:17:09.640 search activity. And then we have the seasonality that we're 0:17:09.680 --> 0:17:12.000 able to then strip out. And then we also have 0:17:12.160 --> 0:17:15.119 this thing we call the residual or the shock. What's 0:17:15.240 --> 0:17:18.919 interesting about the experience that we've seen here in UM 0:17:18.920 --> 0:17:21.919 with COVID nineteen is we were never so interested in 0:17:22.000 --> 0:17:25.560 the shock component and the very quick um shifts in 0:17:25.640 --> 0:17:29.320 search activity either positive or negative until COVID hit, when 0:17:29.320 --> 0:17:33.480 we saw it's just substantial breaks from these trends and 0:17:33.520 --> 0:17:36.359 what would be expected by seasonality. That can be anything 0:17:36.400 --> 0:17:39.959 from the searching for you know, physical policy news, economic news, 0:17:40.359 --> 0:17:43.399 how individuals are searching on the line, then for groceries 0:17:43.880 --> 0:17:47.639 UM and making those type of consumer staples purchases. But 0:17:47.760 --> 0:17:49.600 getting back to it, the idea is to break it 0:17:49.640 --> 0:17:51.720 down into those three components that we get idea of 0:17:51.760 --> 0:17:54.360 what is the you know, the long term trend UM 0:17:54.440 --> 0:17:57.800 and shift really potentially in search activity. How does that 0:17:57.880 --> 0:18:00.440 relate then to to what we're seeing within financial markets 0:18:00.520 --> 0:18:03.840 and overall economic data. And then what are these shock 0:18:03.920 --> 0:18:08.600 components and regarding those big distortions or shifts away from 0:18:08.600 --> 0:18:11.160 those underlying trends, what does that have to tell us 0:18:11.760 --> 0:18:14.600 about how things may be abruptly changing in the near 0:18:14.720 --> 0:18:17.199 term UM and what that can mean? I mean, of 0:18:17.240 --> 0:18:21.240 course for potential volatility UM and equity markets, uncertainty in 0:18:21.280 --> 0:18:24.600 general UM from the consumer base UM and so on. 0:18:25.000 --> 0:18:27.160 And so what we do is we pull down those 0:18:27.160 --> 0:18:30.200 three pieces of information that then gets used within our 0:18:30.720 --> 0:18:33.320 written content as well within our own models and our 0:18:33.320 --> 0:18:37.280 clients models UM and so on. So correct me if 0:18:37.280 --> 0:18:40.040 I'm wrong. But the data that you're using is mostly 0:18:40.640 --> 0:18:44.240 public data. If investors all have access to the same data, 0:18:44.760 --> 0:18:49.919 how are they using that to actually generate outperformance? How 0:18:49.920 --> 0:18:52.840 do they differentiate how they're using the data versus how 0:18:53.000 --> 0:18:56.920 another fund or another investor might be using the data? Right? 0:18:57.040 --> 0:18:59.119 So I mean that's that's the question we get. We 0:18:59.280 --> 0:19:01.320 probably get the most most is since we do deal 0:19:01.400 --> 0:19:04.040 mainly again with with public forms of data, there's plenty 0:19:04.080 --> 0:19:06.880 of alternative data that is private and the credit card 0:19:06.920 --> 0:19:10.360 space and spending UM and so on. Is we try 0:19:10.400 --> 0:19:13.920 to uncover data we think that is underutilized UM. And 0:19:13.960 --> 0:19:16.560 in this case with all of our dealings specifically with 0:19:16.640 --> 0:19:20.960 fixed income portfolio managers, pension fund managers UM and the like, 0:19:21.640 --> 0:19:24.520 the use of search activity on a broader scale, on 0:19:24.600 --> 0:19:27.080 a country by country, even a metro by metro level, 0:19:27.720 --> 0:19:31.360 we believe has been under appreciated UM and non internalized 0:19:31.400 --> 0:19:33.600 the extent that it could be. Now, like you said, once, 0:19:33.640 --> 0:19:37.720 think something like this gets over used or gets used 0:19:37.760 --> 0:19:41.479 as a key benchmark potentially to filling the latent gaps 0:19:41.520 --> 0:19:45.399 between economic data. Potentially some that alpha creation could UM 0:19:45.560 --> 0:19:47.960 evaporate UM and that would mean we have to move 0:19:47.960 --> 0:19:51.080 on to some additional data sources for this time being. 0:19:51.480 --> 0:19:57.120 In all our communications, the front offices of of investment managers, 0:19:57.600 --> 0:20:01.719 banks and so on have not been heavy users of 0:20:01.760 --> 0:20:04.879 the search activity. I think that early uses of it 0:20:05.359 --> 0:20:08.080 prior to the crisis and during the crisis kind of 0:20:08.119 --> 0:20:13.360 fell flat. Again. Maybe the ubiquity of actual Internet usage 0:20:13.440 --> 0:20:16.159 and those young too old that we're using Google was 0:20:16.200 --> 0:20:19.760 not there as of yet. And what we've seen over 0:20:19.840 --> 0:20:22.639 the years, really since two thousand eleven two thousand and twelve, 0:20:22.680 --> 0:20:25.200 search activities ability to fill the gap and really take 0:20:25.240 --> 0:20:29.720 the place of surveys has improved markedly year after year UM. 0:20:29.720 --> 0:20:33.200 And that's something we can measure UM statistically and VR 0:20:33.320 --> 0:20:37.520 modeling for essentially those turning points as to when maybe 0:20:37.920 --> 0:20:41.440 search activity loses its flare loses its ability UH to 0:20:41.600 --> 0:20:45.000 then forecast and now cast via g d P retail 0:20:45.080 --> 0:20:48.720 sales inflation UM and the like. But we're not there yet. 0:20:49.680 --> 0:20:54.080 So obviously the demand for this data, and you mentioned 0:20:54.119 --> 0:20:57.880 maybe search data is sort of relatively newly being incorporated 0:20:57.880 --> 0:21:00.720 into investment processes, but for years we've been hearing about 0:21:01.760 --> 0:21:05.960 satellite looking at parking lots in Walmart, or satellite looking 0:21:05.960 --> 0:21:09.560 at trained or credit card data that's been out there 0:21:09.640 --> 0:21:13.639 is a thing for a while. How intense is the 0:21:13.800 --> 0:21:18.480 search basically for new data sources, either on the bi side, 0:21:18.520 --> 0:21:21.480 the investor side, or you as sort of a data 0:21:21.560 --> 0:21:25.120 vendor so to speak, to just constantly be coming up 0:21:25.160 --> 0:21:29.720 with something that's relatively underappreciated. What does that process look 0:21:29.760 --> 0:21:34.399 like the use of alternative data within the investment world. 0:21:34.480 --> 0:21:36.919 You know, really the investment world was very late to 0:21:37.119 --> 0:21:41.760 using alternative data UM, you know, compared to healthcare, even education, 0:21:42.400 --> 0:21:44.679 UM and the like. And we initially saw this, you know, 0:21:45.000 --> 0:21:47.840 in our routines of going out to big banks, for example, 0:21:48.200 --> 0:21:51.120 and discussing with their teams, UM, you know, how they're 0:21:51.200 --> 0:21:54.520 utilizing alternative data. It was almost always in the back office. 0:21:54.960 --> 0:21:56.760 So it could be UM, you know, anything to do 0:21:56.880 --> 0:22:01.520 with their customer relations. It could chatbots in terms of 0:22:01.560 --> 0:22:04.920 creating natural better language, natural language processing and ployment data 0:22:04.960 --> 0:22:07.919 for that. It could be trade matching all kinds of 0:22:07.960 --> 0:22:10.359 different things that were done in the back office. They 0:22:10.359 --> 0:22:12.920 were trying to basically bring in machine learning, bringing better 0:22:13.040 --> 0:22:15.639 data to create better predictions. And it could have to 0:22:15.640 --> 0:22:18.080 do again with their customers, which customers to call and 0:22:18.080 --> 0:22:21.879 not call, who's going to potentially provide the best UM, 0:22:22.320 --> 0:22:25.320 best avenue for new business UM and so on. But 0:22:25.359 --> 0:22:30.040 what we've seen, i'd say, you know, starting roughly inen, 0:22:30.119 --> 0:22:33.000 we started to see a UM with the advent of 0:22:33.720 --> 0:22:38.359 more alternative data available via numerous vendors, the increase in 0:22:38.400 --> 0:22:43.119 transfer to the front office has happened rapidly UM. And 0:22:43.160 --> 0:22:46.600 I would say that now with COVID nineteen and the 0:22:46.680 --> 0:22:50.400 inability for econ data to keep up with the actual 0:22:50.720 --> 0:22:54.440 UM happenings of the economy UM, and really the needs 0:22:54.440 --> 0:22:56.679 of investors to understand that just what's going on with 0:22:56.880 --> 0:23:01.200