WEBVTT 1 00:00:02.940 --> 00:00:03.440 All right. 2 00:00:03.440 --> 00:00:04.100 Good morning. 3 00:00:04.100 --> 00:00:08.990 [AUDIO OUT] to Cloud BI Trends, The Hybrid Cloud is Dead. 4 00:00:08.990 --> 00:00:13.160 Today we'll be hearing from Snowflake, Sigma and Olivela, 5 00:00:13.160 --> 00:00:15.680 who uses both Snowflake and Sigma. 6 00:00:15.680 --> 00:00:17.420 If you have any questions as you're 7 00:00:17.420 --> 00:00:19.790 going through and listening to the webinar, 8 00:00:19.790 --> 00:00:23.960 feel free to type them into the questions section. 9 00:00:23.960 --> 00:00:27.770 We'll be addressing questions at the end of the webinar, 10 00:00:27.770 --> 00:00:30.950 but feel free to ask questions as we go along. 11 00:00:30.950 --> 00:00:32.990 For now, I'm going to hand it off 12 00:00:32.990 --> 00:00:37.870 to Ross Perez from Snowflake to get us started. 13 00:00:37.870 --> 00:00:39.130 Well, thanks so much, Erica. 14 00:00:39.130 --> 00:00:42.430 And a really big thank you to Olivela, of course, 15 00:00:42.430 --> 00:00:45.850 for jumping on the webinar as well. 16 00:00:45.850 --> 00:00:50.380 We really are appreciative of them using both our products. 17 00:00:50.380 --> 00:00:52.780 And it'll be exciting to be hearing a little bit more 18 00:00:52.780 --> 00:00:54.430 about their story today. 19 00:00:54.430 --> 00:00:56.440 So, yes, I'm Ross Perez, Senior Director 20 00:00:56.440 --> 00:00:57.880 of Marketing at Snowflake. 21 00:00:57.880 --> 00:01:03.280 And what I'm here to talk about is really the impact 22 00:01:03.280 --> 00:01:07.990 that a true first party Cloud solution for data warehousing 23 00:01:07.990 --> 00:01:09.040 can have. 24 00:01:09.040 --> 00:01:14.740 And I really like the title of this webinar and the theme 25 00:01:14.740 --> 00:01:19.360 of avoiding the hybrid Cloud and going fully 26 00:01:19.360 --> 00:01:20.300 on the public Cloud. 27 00:01:20.300 --> 00:01:24.070 And if you look at the history of data platforms and data 28 00:01:24.070 --> 00:01:28.840 warehouses, you can see that this is really the trend 29 00:01:28.840 --> 00:01:31.900 and where the trend is going for data warehousing, 30 00:01:31.900 --> 00:01:35.560 and certainly for analytics as well. 31 00:01:35.560 --> 00:01:37.780 We've gone through a lot of changes, 32 00:01:37.780 --> 00:01:42.520 really, in the past 30 years in the world of data warehousing, 33 00:01:42.520 --> 00:01:46.060 where we went from the first relational databases 34 00:01:46.060 --> 00:01:50.170 to the first data warehousing appliances 35 00:01:50.170 --> 00:01:52.520 through a whole slew of different technologies, 36 00:01:52.520 --> 00:01:55.240 including open source technologies 37 00:01:55.240 --> 00:01:56.830 in the early 2000s. 38 00:01:56.830 --> 00:02:01.810 And now we've kind of arrived to the place where it's certainly 39 00:02:01.810 --> 00:02:05.675 first party Cloud and, we believe, 40 00:02:05.675 --> 00:02:07.300 the data warehouse built for the Cloud, 41 00:02:07.300 --> 00:02:09.082 which of course is Snowflake. 42 00:02:13.710 --> 00:02:19.320 Now if we kind of take a look at what the goals are for somebody 43 00:02:19.320 --> 00:02:21.300 who's using a data warehouse for analytics, 44 00:02:21.300 --> 00:02:23.490 they're really quite straightforward. 45 00:02:23.490 --> 00:02:25.890 Of course you want to use your data warehouse 46 00:02:25.890 --> 00:02:28.920 to develop your own customer experience, 47 00:02:28.920 --> 00:02:33.840 to look at what's going on in whatever business 48 00:02:33.840 --> 00:02:36.030 that you may be running, and understanding 49 00:02:36.030 --> 00:02:38.850 how your customers are interacting with you. 50 00:02:38.850 --> 00:02:41.192 It's quality assurance, knowing that the product 51 00:02:41.192 --> 00:02:42.900 that you're putting forward, whether it's 52 00:02:42.900 --> 00:02:45.330 a service or an actual product, is something 53 00:02:45.330 --> 00:02:50.330 that people can really understand and use properly 54 00:02:50.330 --> 00:02:52.050 and integrate with properly. 55 00:02:52.050 --> 00:02:53.970 Operational efficiency, making sure everything 56 00:02:53.970 --> 00:02:55.470 is running properly. 57 00:02:55.470 --> 00:02:57.870 And innovation, actually taking the data 58 00:02:57.870 --> 00:03:00.300 that you have in your data warehouse, 59 00:03:00.300 --> 00:03:03.430 using a powerful PI tool like Sigma 60 00:03:03.430 --> 00:03:09.330 to understand what you can do to improve what you're doing 61 00:03:09.330 --> 00:03:12.910 and come up with new ideas and opportunities for the future 62 00:03:12.910 --> 00:03:13.410 as well. 63 00:03:17.140 --> 00:03:17.870 Next slide. 64 00:03:17.870 --> 00:03:18.370 Great. 65 00:03:18.370 --> 00:03:22.210 So the capabilities that Snowflake delivers 66 00:03:22.210 --> 00:03:24.430 are actually really quite straightforward. 67 00:03:24.430 --> 00:03:27.160 So we're a data warehouse built for the Cloud. 68 00:03:27.160 --> 00:03:28.810 And what that means is that, instead 69 00:03:28.810 --> 00:03:33.130 of utilizing older, 30 years old, data warehousing 70 00:03:33.130 --> 00:03:35.890 architecture, Snowflake is actually built for the Cloud 71 00:03:35.890 --> 00:03:39.440 from the ground up with a completely new architecture. 72 00:03:39.440 --> 00:03:42.550 This architecture utilizes the power of Cloud infrastructure 73 00:03:42.550 --> 00:03:45.340 providers, such as AWS and Azure, 74 00:03:45.340 --> 00:03:48.610 to completely separate the storage 75 00:03:48.610 --> 00:03:51.970 and compute components of the data warehouse 76 00:03:51.970 --> 00:03:55.240 and enable you to scale them completely independently. 77 00:03:55.240 --> 00:03:58.390 What this means is that you have a SAS first party service, 78 00:03:58.390 --> 00:04:01.300 so it's relatively easy to manage. 79 00:04:01.300 --> 00:04:03.580 It enables you to store limited amounts of data. 80 00:04:03.580 --> 00:04:06.880 So you can now have all of your data in one system. 81 00:04:06.880 --> 00:04:08.560 Of course you can scale on demand 82 00:04:08.560 --> 00:04:11.230 and use the Cloud to basically match 83 00:04:11.230 --> 00:04:14.080 the scale of what you're deploying with Snowflake 84 00:04:14.080 --> 00:04:15.790 to whatever your need is. 85 00:04:15.790 --> 00:04:19.240 And the kind of great part about it 86 00:04:19.240 --> 00:04:23.500 is that this is not a open source technology that's 87 00:04:23.500 --> 00:04:24.370 difficult to use. 88 00:04:24.370 --> 00:04:27.450 You can use standard SQL to interact with Snowflake 89 00:04:27.450 --> 00:04:29.190 and work with it day to day. 90 00:04:33.630 --> 00:04:36.720 So let's dig into these individual components 91 00:04:36.720 --> 00:04:37.930 in a little bit more detail. 92 00:04:37.930 --> 00:04:40.170 So what do we mean by easy management? 93 00:04:40.170 --> 00:04:42.570 Well, with traditional data warehouses, 94 00:04:42.570 --> 00:04:45.040 they were designed at a time that was much, 95 00:04:45.040 --> 00:04:47.250 much further ago, about 30 years ago, 96 00:04:47.250 --> 00:04:50.070 and sometimes you can experience a significant amount 97 00:04:50.070 --> 00:04:51.878 of management overhead. 98 00:04:51.878 --> 00:04:53.670 And it's important to point out that that's 99 00:04:53.670 --> 00:04:57.540 even true with data warehouses that are technically 100 00:04:57.540 --> 00:05:02.220 in the Cloud, but in many cases are older technologies 101 00:05:02.220 --> 00:05:05.460 have simply been kind of warmed over and put into the Cloud. 102 00:05:05.460 --> 00:05:07.800 But really in the background there's 103 00:05:07.800 --> 00:05:10.980 still older technologies that require a lot of management. 104 00:05:10.980 --> 00:05:12.480 Well, with Snowflake, you don't have 105 00:05:12.480 --> 00:05:13.897 to worry about infrastructure, you 106 00:05:13.897 --> 00:05:15.360 don't have to worry about tuning, 107 00:05:15.360 --> 00:05:17.460 there's no optimization or indexing, 108 00:05:17.460 --> 00:05:20.130 you don't have to partition and worry about how storage 109 00:05:20.130 --> 00:05:24.090 is going to work, certainly no vacuuming or sorting, 110 00:05:24.090 --> 00:05:25.260 or workload management. 111 00:05:25.260 --> 00:05:27.690 So the tasks that you would traditionally 112 00:05:27.690 --> 00:05:31.822 associate with managing and optimizing a data warehouse 113 00:05:31.822 --> 00:05:33.030 aren't required in Snowflake. 114 00:05:33.030 --> 00:05:37.530 It's a SAS data warehouse that really manages the optimization 115 00:05:37.530 --> 00:05:40.410 and tuning for you, so that you don't have to worry about it. 116 00:05:45.440 --> 00:05:47.700 Another capability that the Cloud really affords us 117 00:05:47.700 --> 00:05:51.860 is to account for newer data types that are becoming 118 00:05:51.860 --> 00:05:53.653 significantly more common. 119 00:05:53.653 --> 00:05:55.070 For instance, semi-structured data 120 00:05:55.070 --> 00:05:57.980 like JSON, Avro, XML, Parquet, and ORC 121 00:05:57.980 --> 00:05:59.990 can be stored in Snowflake natively, 122 00:05:59.990 --> 00:06:02.205 alongside CSV and text data. 123 00:06:02.205 --> 00:06:03.830 You also notice that this aligns really 124 00:06:03.830 --> 00:06:05.640 well with the capabilities in Sigma 125 00:06:05.640 --> 00:06:07.880 to query semi-structured data. 126 00:06:07.880 --> 00:06:09.830 So with Snowflake and Sigma together, 127 00:06:09.830 --> 00:06:12.830 you can load semi-structured data into a variant type 128 00:06:12.830 --> 00:06:14.690 column in Snowflake, and then you 129 00:06:14.690 --> 00:06:17.780 can query that data, using either Snowflake or Sigma. 130 00:06:17.780 --> 00:06:20.740 In Snowflake you can use dot notation to create this data. 131 00:06:20.740 --> 00:06:23.030 In Sigma you can natively connect and use 132 00:06:23.030 --> 00:06:26.720 the semi-structured data through the UI. 133 00:06:26.720 --> 00:06:30.260 So regardless of how you want to interact with your data, 134 00:06:30.260 --> 00:06:31.790 the really important thing here is 135 00:06:31.790 --> 00:06:34.370 that you can use Snowflake and Sigma together 136 00:06:34.370 --> 00:06:37.850 to query any type of data that you may 137 00:06:37.850 --> 00:06:39.863 be collecting and utilizing. 138 00:06:39.863 --> 00:06:41.780 The other thing that's important to point out, 139 00:06:41.780 --> 00:06:44.900 and really a huge advantage to using the public Cloud 140 00:06:44.900 --> 00:06:48.295 for services like this, is that you can store as much data 141 00:06:48.295 --> 00:06:48.920 as you need to. 142 00:06:48.920 --> 00:06:51.800 You never have to worry about the capabilities 143 00:06:51.800 --> 00:06:54.260 or the actual infrastructure or hardware sitting 144 00:06:54.260 --> 00:06:57.080 behind the data that you're loading because you know 145 00:06:57.080 --> 00:07:00.350 that, at the end of the day, you can store petabytes of data 146 00:07:00.350 --> 00:07:03.792 if you need to and you can start loading it immediately. 147 00:07:07.540 --> 00:07:08.040 All right. 148 00:07:08.040 --> 00:07:09.810 So this slide builds. 149 00:07:09.810 --> 00:07:13.410 And I think you know it's really important to get into here, 150 00:07:13.410 --> 00:07:18.570 is the reason why on demand usage is so important. 151 00:07:18.570 --> 00:07:21.960 And this, again, is a huge differentiator 152 00:07:21.960 --> 00:07:24.420 between the hybrid Cloud and the public Cloud. 153 00:07:24.420 --> 00:07:28.260 So if we look at data warehousing as a use case, 154 00:07:28.260 --> 00:07:30.930 and you can click next here. 155 00:07:30.930 --> 00:07:34.090 So data warehousing usage varies. 156 00:07:34.090 --> 00:07:35.730 It goes up and down over time. 157 00:07:35.730 --> 00:07:37.710 In a traditional warehouse, which 158 00:07:37.710 --> 00:07:41.460 we can see in the next click, is completely 159 00:07:41.460 --> 00:07:44.950 inflexible to the change in usage over time. 160 00:07:44.950 --> 00:07:49.470 In other words, if you have a huge rush of people using Sigma 161 00:07:49.470 --> 00:07:52.620 on Monday morning, well traditional data warehouses 162 00:07:52.620 --> 00:07:55.360 are unable to scale to meet this demand. 163 00:07:55.360 --> 00:07:57.360 And if you look at hybrid Cloud data warehouses, 164 00:07:57.360 --> 00:07:59.070 it's exactly the same. 165 00:07:59.070 --> 00:08:02.010 But Snowflake uses the public Cloud 166 00:08:02.010 --> 00:08:06.370 to enable as much elasticity as you need at any given time. 167 00:08:06.370 --> 00:08:09.300 So instead of matching compute to whatever your highest 168 00:08:09.300 --> 00:08:12.630 amount of use could be in the future, 169 00:08:12.630 --> 00:08:14.970 you can match your compute at any given moment 170 00:08:14.970 --> 00:08:16.890 to exactly what you're using. 171 00:08:16.890 --> 00:08:19.650 This means that you don't have to pay more for when you're not 172 00:08:19.650 --> 00:08:21.660 using the data warehouse. 173 00:08:21.660 --> 00:08:24.720 It also means that you can avoid purchasing more capacity 174 00:08:24.720 --> 00:08:27.990 than you need, so that you can support 175 00:08:27.990 --> 00:08:30.042 times of maximum demand. 176 00:08:30.042 --> 00:08:31.750 And, of course, it allows you to scale up 177 00:08:31.750 --> 00:08:36.280 and down transparently and automatically. 178 00:08:36.280 --> 00:08:38.350 The other advantage here with Snowflake, 179 00:08:38.350 --> 00:08:40.210 and the last one that I'll talk to today, 180 00:08:40.210 --> 00:08:43.280 is the ability to use standard skills and tools. 181 00:08:43.280 --> 00:08:48.585 So, for instance, most people are familiar with SQL, 182 00:08:48.585 --> 00:08:51.720 and with Snowflake you can utilize standard SQL 183 00:08:51.720 --> 00:08:54.250 to be able to query and interact with the data. 184 00:08:57.730 --> 00:08:59.830 You can also use tools like Sigma, 185 00:08:59.830 --> 00:09:03.910 and other ETL tools that are available to you, 186 00:09:03.910 --> 00:09:07.300 to be able to connect with Snowflake. 187 00:09:07.300 --> 00:09:11.010 And because Snowflake utilizes standard SQL, 188 00:09:11.010 --> 00:09:12.670 the compatibility with other tools 189 00:09:12.670 --> 00:09:15.920 is really high, making it easier to integrate with the systems 190 00:09:15.920 --> 00:09:17.420 and tools that you're already using. 191 00:09:20.320 --> 00:09:22.780 And it's highly available, meaning 192 00:09:22.780 --> 00:09:26.110 that you don't have to worry about Snowflake. 193 00:09:26.110 --> 00:09:28.840 It basically, out of the box, it's usable in the way 194 00:09:28.840 --> 00:09:31.330 that you need to be and it's available 195 00:09:31.330 --> 00:09:32.950 and it's highly secure as well. 196 00:09:38.610 --> 00:09:39.420 All right. 197 00:09:39.420 --> 00:09:41.920 So at this point, you may be wondering, 198 00:09:41.920 --> 00:09:46.500 hopefully you're wondering, how to be able to not only utilize 199 00:09:46.500 --> 00:09:50.520 Snowflake, but also utilize these other ecosystem 200 00:09:50.520 --> 00:09:53.070 tools that I've been referring to, such as Sigma. 201 00:09:53.070 --> 00:09:56.880 And you may also be interested in working 202 00:09:56.880 --> 00:09:59.820 with an ETL vendor that can help you start getting data 203 00:09:59.820 --> 00:10:01.170 into Snowflake. 204 00:10:01.170 --> 00:10:03.750 Well we had a lot of customers who were asking us you, 205 00:10:03.750 --> 00:10:05.850 "Well, hey we're enjoying using Snowflake 206 00:10:05.850 --> 00:10:08.240 and we understand you have this broad ecosystem, 207 00:10:08.240 --> 00:10:10.530 but how am I going to get data into it? 208 00:10:10.530 --> 00:10:13.590 How am I going to stage this data? 209 00:10:13.590 --> 00:10:17.520 And how am I going to be able to start analyzing the data/" 210 00:10:17.520 --> 00:10:21.990 Well, of course, with Partner Connect what you can do 211 00:10:21.990 --> 00:10:25.920 is log into Snowflake and you can very easily 212 00:10:25.920 --> 00:10:29.760 start up a trial with any of the partner tools 213 00:10:29.760 --> 00:10:32.430 that are available through Partner Connect, including 214 00:10:32.430 --> 00:10:36.690 Sigma and ETL vendors such as Fivetran. 215 00:10:36.690 --> 00:10:39.180 So you can go in to Snowflake, and without having 216 00:10:39.180 --> 00:10:43.020 to integrate these tools on your own, 217 00:10:43.020 --> 00:10:44.520 you can actually use Partner Connect 218 00:10:44.520 --> 00:10:45.933 to get started right away. 219 00:10:45.933 --> 00:10:47.100 And it makes it really easy. 220 00:10:47.100 --> 00:10:50.190 And actually if you're interested in seeing 221 00:10:50.190 --> 00:10:52.180 how you can connect Snowflake with Sigma, 222 00:10:52.180 --> 00:10:57.330 there's a YouTube video that's available online. 223 00:10:57.330 --> 00:11:00.870 You can search for Sigma and Snowflake Partner Connect 224 00:11:00.870 --> 00:11:04.320 and it'll show you the actual steps for connecting Snowflake 225 00:11:04.320 --> 00:11:06.960 to Sigma quite easily. 226 00:11:12.190 --> 00:11:14.260 And here you can see the interface 227 00:11:14.260 --> 00:11:16.705 for actually going through that connection. 228 00:11:16.705 --> 00:11:17.580 And that's it for me. 229 00:11:22.650 --> 00:11:24.040 Hi, everyone. 230 00:11:24.040 --> 00:11:25.140 My name is Ali Sayeed. 231 00:11:25.140 --> 00:11:28.680 I'm a Product Solutions Engineer with Sigma. 232 00:11:28.680 --> 00:11:31.350 So I'm here to kind of talk through the product, 233 00:11:31.350 --> 00:11:33.180 guide you through it, give you a demo, 234 00:11:33.180 --> 00:11:35.478 then answer any technical questions. 235 00:11:41.100 --> 00:11:45.360 So this new Cloud architecture, that Ross has talked about, 236 00:11:45.360 --> 00:11:47.040 has really created a new technology 237 00:11:47.040 --> 00:11:49.330 with that new opportunity. 238 00:11:49.330 --> 00:11:51.910 And now modern analytics now takes advantage 239 00:11:51.910 --> 00:11:54.410 of your Cloud data warehouse. 240 00:11:54.410 --> 00:11:58.000 BI has now evolved to be Cloud built, right. 241 00:11:58.000 --> 00:12:00.040 And now can handle semi-structured data, 242 00:12:00.040 --> 00:12:01.750 such as JSON. 243 00:12:01.750 --> 00:12:04.480 Even allows you to query billions of records of data 244 00:12:04.480 --> 00:12:09.350 and ask tough questions, all without writing a line of SQL. 245 00:12:09.350 --> 00:12:12.040 What that really does, it kind of provides 246 00:12:12.040 --> 00:12:15.258 that direct and secure and lasting performance, 247 00:12:15.258 --> 00:12:17.050 that Ross just talked about, but now anyone 248 00:12:17.050 --> 00:12:18.300 in [INAUDIBLE] can now use it. 249 00:12:23.910 --> 00:12:24.420 So, Sigma. 250 00:12:24.420 --> 00:12:28.170 Sigma has reinvented the IN analytics for the Cloud, right. 251 00:12:28.170 --> 00:12:31.260 So now it's really easy and powerful to use. 252 00:12:31.260 --> 00:12:34.920 It's secure and governed, right. 253 00:12:34.920 --> 00:12:38.130 It's got a usage pricing based model. 254 00:12:38.130 --> 00:12:40.110 And, just like Snowflake, it requires 255 00:12:40.110 --> 00:12:43.650 no upfront work for you and your organization to get started. 256 00:12:43.650 --> 00:12:45.720 So really gone are the days of having 257 00:12:45.720 --> 00:12:48.660 to model and map the data beforehand, or ask the data 258 00:12:48.660 --> 00:12:50.915 team to work with the new data. 259 00:12:50.915 --> 00:12:52.290 The traditional surveyor that was 260 00:12:52.290 --> 00:12:55.854 made famous by traditional tools, such as [INAUDIBLE].. 261 00:13:00.330 --> 00:13:01.448 All right. 262 00:13:01.448 --> 00:13:02.490 So let's get to the demo. 263 00:13:14.250 --> 00:13:15.620 OK. 264 00:13:15.620 --> 00:13:17.950 So this is Sigma. 265 00:13:17.950 --> 00:13:21.190 It's a cloud native app, so browser based. 266 00:13:21.190 --> 00:13:24.580 Simply log onto app.sigmacomputing.com. 267 00:13:24.580 --> 00:13:27.310 This is the first screen you see, the Welcome page. 268 00:13:27.310 --> 00:13:30.250 And we're going to use Sigma to connect to data on a Snowflake 269 00:13:30.250 --> 00:13:31.260 instance. 270 00:13:31.260 --> 00:13:33.760 So the first thing I'm going to show you is our team spaces. 271 00:13:33.760 --> 00:13:35.790 For example, when you connect to data 272 00:13:35.790 --> 00:13:38.750 you can expose data and dashboards at different levels. 273 00:13:38.750 --> 00:13:42.100 So, for example, I can expose it at the team company wide level, 274 00:13:42.100 --> 00:13:44.680 here; at a team level, for example, I'm 275 00:13:44.680 --> 00:13:47.660 part of the marketing and product solutions teams. 276 00:13:47.660 --> 00:13:50.710 So what these team spaces work as is they're 277 00:13:50.710 --> 00:13:52.210 kind of collective work environments 278 00:13:52.210 --> 00:13:56.720 that your team can work on specific data sets together, 279 00:13:56.720 --> 00:13:59.513 additionally to provide another kind of permissions layer 280 00:13:59.513 --> 00:14:01.180 on top of your Snowflake data warehouse. 281 00:14:01.180 --> 00:14:03.315 So for example, unless I'm part of this team, 282 00:14:03.315 --> 00:14:04.690 I'm unable to see the data that's 283 00:14:04.690 --> 00:14:06.273 been exposed at this level. 284 00:14:06.273 --> 00:14:07.690 Then again you're kind of private, 285 00:14:07.690 --> 00:14:10.000 my documents, work area. 286 00:14:10.000 --> 00:14:12.580 I'm going to connect to data and show you 287 00:14:12.580 --> 00:14:16.520 how you can build data and visualizations in Sigma 288 00:14:16.520 --> 00:14:18.300 from scratch. 289 00:14:18.300 --> 00:14:21.582 So we're going to go ahead and click New worksheet. 290 00:14:21.582 --> 00:14:23.290 So you'll notice there's three ways I can 291 00:14:23.290 --> 00:14:25.690 get started with data in Sigma. 292 00:14:25.690 --> 00:14:29.037 So, for example, if I do have a SQL query already written 293 00:14:29.037 --> 00:14:31.120 and I want to start from that, I have that ability 294 00:14:31.120 --> 00:14:33.490 to input that into our editor. 295 00:14:33.490 --> 00:14:36.400 We also have the notion of a reference worksheet. 296 00:14:36.400 --> 00:14:38.500 So that reference worksheet is basically 297 00:14:38.500 --> 00:14:44.055 a non-materialized view that you can curate and expose to users. 298 00:14:44.055 --> 00:14:45.430 For example, if there's users you 299 00:14:45.430 --> 00:14:49.270 don't want directly to use the database. 300 00:14:49.270 --> 00:14:50.730 And then we're directly connected 301 00:14:50.730 --> 00:14:53.630 to a table in the database, and this is what most of our users 302 00:14:53.630 --> 00:14:54.130 will use. 303 00:14:57.410 --> 00:14:59.160 So we're going to connect to our demo data 304 00:14:59.160 --> 00:15:01.550 on our Snowflake instance. 305 00:15:01.550 --> 00:15:04.768 You'll see immediately the schemas that I have access to. 306 00:15:04.768 --> 00:15:07.310 The data I want to connect to is in this Insta Credit schema, 307 00:15:07.310 --> 00:15:09.600 it's called Orders JSON Large. 308 00:15:09.600 --> 00:15:12.220 Select that. 309 00:15:12.220 --> 00:15:15.360 It's going to show you a preview of this data, some information 310 00:15:15.360 --> 00:15:16.200 about it. 311 00:15:16.200 --> 00:15:18.630 So over 18 million rows of data. 312 00:15:18.630 --> 00:15:20.970 Again, this is instantly more than anything 313 00:15:20.970 --> 00:15:24.315 you could do as an analyst in Excel. 314 00:15:24.315 --> 00:15:25.690 When we take a look at this data, 315 00:15:25.690 --> 00:15:30.210 we see OK, it's actually older data with a large nested JSON 316 00:15:30.210 --> 00:15:31.460 field in it. 317 00:15:31.460 --> 00:15:34.030 So we're going to use Sigma to parse 318 00:15:34.030 --> 00:15:37.120 out the relevant fields from this JSON so flatten them out. 319 00:15:37.120 --> 00:15:42.040 And then use that to determine, basically create 320 00:15:42.040 --> 00:15:45.580 customer cohorts and determine what those customer cohorts 321 00:15:45.580 --> 00:15:48.740 performance [INAUDIBLE]. 322 00:15:48.740 --> 00:15:51.980 When I click get started, that actually brings the data 323 00:15:51.980 --> 00:15:53.390 into our development UI. 324 00:15:53.390 --> 00:15:55.820 So it's a spreadsheet based UI, so very 325 00:15:55.820 --> 00:15:57.820 easy to get familiar with. 326 00:15:57.820 --> 00:16:00.650 So you'll notice that there's certain functions that 327 00:16:00.650 --> 00:16:03.740 [INAUDIBLE],, you have the column headings, such as hiding 328 00:16:03.740 --> 00:16:06.550 columns, sorting, aggregating. 329 00:16:06.550 --> 00:16:07.842 There's this formula bar, here. 330 00:16:07.842 --> 00:16:10.300 This is where we're going to write formulas, take advantage 331 00:16:10.300 --> 00:16:11.990 of our functions library to do that. 332 00:16:11.990 --> 00:16:13.820 And on the right, on the Inspector, 333 00:16:13.820 --> 00:16:15.278 this is where we're going to create 334 00:16:15.278 --> 00:16:16.400 groupings and aggregates. 335 00:16:16.400 --> 00:16:20.770 They also can join to the data sources as well. 336 00:16:20.770 --> 00:16:23.270 All right, so I'm going to get started building. 337 00:16:23.270 --> 00:16:25.000 So maybe the first thing I want to do 338 00:16:25.000 --> 00:16:29.870 is go ahead and sort this data in an ascending manner. 339 00:16:29.870 --> 00:16:32.170 The oldest records appear first. 340 00:16:32.170 --> 00:16:33.290 We've done that. 341 00:16:33.290 --> 00:16:35.000 Again, next thing I want to do, maybe 342 00:16:35.000 --> 00:16:36.535 I want to create a column. 343 00:16:36.535 --> 00:16:37.910 So we have quantity, unit, price. 344 00:16:37.910 --> 00:16:39.700 I want to find the total line amount. 345 00:16:39.700 --> 00:16:41.640 I can start to type that in the formula bar, 346 00:16:41.640 --> 00:16:45.550 notice that there's autocomplete built in, like so. 347 00:16:45.550 --> 00:16:48.380 And we'll go ahead and call this amount. 348 00:16:48.380 --> 00:16:49.460 Just like that. 349 00:16:49.460 --> 00:16:53.160 Additionally, I can also format this as a currency. 350 00:16:53.160 --> 00:16:54.960 So notice how fast everything is happening. 351 00:16:54.960 --> 00:16:57.180 Right? 352 00:16:57.180 --> 00:16:59.010 And the reason that it's happening so fast 353 00:16:59.010 --> 00:17:02.550 with such a large data set is because Sigma is actually 354 00:17:02.550 --> 00:17:06.240 creating actual SQL and executing it in the database. 355 00:17:06.240 --> 00:17:09.450 So we're taking a SQL and we're taking advantage of Snowflake 356 00:17:09.450 --> 00:17:12.510 and it's elastic compute to execute these queries. 357 00:17:12.510 --> 00:17:14.550 So I can actually take a look at what 358 00:17:14.550 --> 00:17:17.369 that SQL looks like that Sigma is generating, 359 00:17:17.369 --> 00:17:20.619 that's being pushed to the database. 360 00:17:20.619 --> 00:17:26.140 Because Sigma is creating actual SQL, there's no need to model 361 00:17:26.140 --> 00:17:28.450 or map the data out beforehand. 362 00:17:28.450 --> 00:17:31.570 You're up and running as soon as you connect to your database. 363 00:17:31.570 --> 00:17:34.970 This also means that we're never working on extracts 364 00:17:34.970 --> 00:17:37.502 and we're never pulling out the data to process it. 365 00:17:37.502 --> 00:17:38.960 Processing it in the database, that 366 00:17:38.960 --> 00:17:40.930 means it's never leaving the secure environment 367 00:17:40.930 --> 00:17:42.597 and we're never storing or caching data. 368 00:17:46.720 --> 00:17:48.820 The next thing I want to do is flatten out 369 00:17:48.820 --> 00:17:51.730 some relevant fields I want from this JSON. 370 00:17:51.730 --> 00:17:54.520 Sigma will recognize that JSON in there 371 00:17:54.520 --> 00:17:57.480 and give you the option to extract columns. 372 00:17:57.480 --> 00:17:59.050 So I'll see this nested JSON here, 373 00:17:59.050 --> 00:18:01.300 I can actually choose the columns that I want from it. 374 00:18:01.300 --> 00:18:04.000 For example, maybe I want product ID, product name, 375 00:18:04.000 --> 00:18:04.967 and user ID. 376 00:18:08.550 --> 00:18:11.190 Then Sigma will parse, or flatten those columns out. 377 00:18:11.190 --> 00:18:14.860 I can go ahead and hide this JSON because I don't need it. 378 00:18:14.860 --> 00:18:21.570 And quickly rename the fields, like so. 379 00:18:26.557 --> 00:18:28.890 The next thing I want to do is I want to create a group. 380 00:18:28.890 --> 00:18:30.830 So I want to group my data by the user ID 381 00:18:30.830 --> 00:18:34.880 so that I can then treat my customer cohort. 382 00:18:34.880 --> 00:18:37.700 I can easily do that by taking the user ID column 383 00:18:37.700 --> 00:18:40.916 and putting it on a drooping level. 384 00:18:40.916 --> 00:18:43.140 So now we've created an aggregate there. 385 00:18:43.140 --> 00:18:45.128 I can create a column or calculation 386 00:18:45.128 --> 00:18:46.170 of aggregating be levels. 387 00:18:46.170 --> 00:18:48.480 For example, I want to know the total amount generated 388 00:18:48.480 --> 00:18:49.100 per user. 389 00:18:49.100 --> 00:18:53.440 I can simply do sum of amount. 390 00:18:53.440 --> 00:18:57.970 Call this user [INAUDIBLE] like so. 391 00:18:57.970 --> 00:19:01.460 So, yeah, you can easily do a cross level calculation 392 00:19:01.460 --> 00:19:03.203 like that. 393 00:19:03.203 --> 00:19:04.370 I'll create one more column. 394 00:19:04.370 --> 00:19:08.660 So maybe I want put each customer into a cohort, 395 00:19:08.660 --> 00:19:10.290 when did they become a customer? 396 00:19:10.290 --> 00:19:13.230 I could say, OK let's do that based on their first purchase 397 00:19:13.230 --> 00:19:13.730 date. 398 00:19:13.730 --> 00:19:15.480 So I'll do minimum on their purchase date. 399 00:19:20.400 --> 00:19:22.950 And then, additionally, so you see the data 400 00:19:22.950 --> 00:19:24.780 at its most granular level right now, 401 00:19:24.780 --> 00:19:28.220 I can go ahead and collapse the data to the user 402 00:19:28.220 --> 00:19:31.430 ID grouping you see there. 403 00:19:31.430 --> 00:19:34.250 And I want to see it at the quarter level. 404 00:19:34.250 --> 00:19:39.557 I can truncate the data to the quarter, via the column heading 405 00:19:39.557 --> 00:19:40.640 because it's a JSON field. 406 00:19:40.640 --> 00:19:44.190 So we'll do that like so. 407 00:19:44.190 --> 00:19:45.740 And I can change any kind of formats 408 00:19:45.740 --> 00:19:49.670 if I want to see it in a short day, like that, I can do so. 409 00:19:49.670 --> 00:19:53.950 I can even say let's go ahead and take a look at the month. 410 00:19:53.950 --> 00:19:57.450 Let's go ahead and call this cohort. 411 00:19:57.450 --> 00:19:59.690 I've created the customer cohort. 412 00:19:59.690 --> 00:20:02.000 Again, I wanted to group my customers 413 00:20:02.000 --> 00:20:05.930 into each cohort to see how cohorts perform individually. 414 00:20:05.930 --> 00:20:07.820 So I'll simply take that cohort field 415 00:20:07.820 --> 00:20:13.460 and add it to new aggregate level like so. 416 00:20:13.460 --> 00:20:16.500 Again, I can collapse the data to the aggregate level, 417 00:20:16.500 --> 00:20:19.680 cohort level, so I can see the quarters that are present. 418 00:20:19.680 --> 00:20:23.650 Let's go ahead and do some more calculations at this level now. 419 00:20:23.650 --> 00:20:25.640 So for example, maybe I want to know 420 00:20:25.640 --> 00:20:30.570 the total revenue or the total number of users 421 00:20:30.570 --> 00:20:31.980 that were in each cohort. 422 00:20:31.980 --> 00:20:35.174 I can simply do a count on the user ID. 423 00:20:46.160 --> 00:20:48.380 So what I've seen so far was basically 424 00:20:48.380 --> 00:20:51.730 a lot of the simple functions from our functions library, 425 00:20:51.730 --> 00:20:54.440 but we've pretty much covered 99% of all SQL functions, 426 00:20:54.440 --> 00:20:56.660 anything from date/time functions 427 00:20:56.660 --> 00:21:00.270 such as these differences; text functions, such as contains 428 00:21:00.270 --> 00:21:03.830 or user [INAUDIBLE];; and then quite a few window functions, 429 00:21:03.830 --> 00:21:05.090 such as moving aggregates. 430 00:21:05.090 --> 00:21:07.010 I'm going to use a window function now. 431 00:21:07.010 --> 00:21:10.070 I'm going to use cumulative sum to see how our customers 432 00:21:10.070 --> 00:21:12.545 acquisition has kind of trade on throughout the quarters. 433 00:21:12.545 --> 00:21:15.680 I'll start to type it in to that find that function I want 434 00:21:15.680 --> 00:21:17.690 to use via the auto complete. 435 00:21:17.690 --> 00:21:22.350 I want to do this on the user account, customer account, 436 00:21:22.350 --> 00:21:25.460 like so. 437 00:21:25.460 --> 00:21:27.862 We can now see how our customer's acquisition has 438 00:21:27.862 --> 00:21:29.570 kind of progressed throughout the quarter 439 00:21:29.570 --> 00:21:32.780 starting with 3,000 all the way to 250,000 customers we 440 00:21:32.780 --> 00:21:34.110 have now. 441 00:21:34.110 --> 00:21:37.980 Let's just call this cumulative sum. 442 00:21:37.980 --> 00:21:40.917 I'll create one more metric, here. 443 00:21:40.917 --> 00:21:43.250 We'll call this, maybe you want to do performance metric 444 00:21:43.250 --> 00:21:44.917 that we want to chart, for example, what 445 00:21:44.917 --> 00:21:49.430 was the median of user revenue. 446 00:21:49.430 --> 00:21:51.360 Again, a cross level calculation. 447 00:21:54.390 --> 00:21:58.420 Call this median revenue. 448 00:21:58.420 --> 00:22:01.120 So now I've arrived at the data set that I wanted. 449 00:22:01.120 --> 00:22:04.160 I can go ahead and now click the Publish button. 450 00:22:04.160 --> 00:22:07.240 And what that does is actually create a new version 451 00:22:07.240 --> 00:22:09.010 of this worksheet for me. 452 00:22:09.010 --> 00:22:10.360 So version control is built in. 453 00:22:10.360 --> 00:22:12.110 If I want to revert to a previous version, 454 00:22:12.110 --> 00:22:14.620 I can actually do so by clicking over here. 455 00:22:14.620 --> 00:22:19.810 You have the ability to download the data as a CSV or Excel. 456 00:22:19.810 --> 00:22:22.660 You can also send this data on a scheduled basis. 457 00:22:22.660 --> 00:22:25.090 For example, a list of emails, such as your boss, 458 00:22:25.090 --> 00:22:28.480 could get this data as an Excel or PDF every Monday morning 459 00:22:28.480 --> 00:22:29.046 at 9AM. 460 00:22:32.268 --> 00:22:33.810 But I'm not done yet, I actually want 461 00:22:33.810 --> 00:22:36.240 to start to build a charter visualization. 462 00:22:36.240 --> 00:22:39.240 So I can go ahead and click the visualization tab 463 00:22:39.240 --> 00:22:41.422 and open that up. 464 00:22:41.422 --> 00:22:44.350 So you notice it supports quite a bit of charts, 465 00:22:44.350 --> 00:22:47.910 anything from bar charts, line charts, scatter plots. 466 00:22:47.910 --> 00:22:50.500 And we make it very easy to create those or fill it. 467 00:22:50.500 --> 00:22:52.960 So, for example, on my x-axis I want the cohort. 468 00:22:52.960 --> 00:22:55.600 I can just drop it in there. 469 00:22:55.600 --> 00:22:59.080 I want to chart the customer account. 470 00:22:59.080 --> 00:23:02.087 I can simply put it in as a value there. 471 00:23:02.087 --> 00:23:03.670 If I want to add an additional series, 472 00:23:03.670 --> 00:23:05.110 such as the median revenue, I can 473 00:23:05.110 --> 00:23:08.060 go ahead and add that is well. 474 00:23:08.060 --> 00:23:10.980 And then I can decide, this doesn't look quite right 475 00:23:10.980 --> 00:23:15.380 as a bar chart, let's see it as a grouped bar chart. 476 00:23:15.380 --> 00:23:17.100 I want the revenue value to show more 477 00:23:17.100 --> 00:23:24.000 so we can actually move that to a second axis, on the right. 478 00:23:24.000 --> 00:23:27.870 And I can notice that what this is maybe not the best 479 00:23:27.870 --> 00:23:29.400 chart for seeing trends. 480 00:23:29.400 --> 00:23:32.300 Let's make this a line chart. 481 00:23:32.300 --> 00:23:35.650 And I can see some kind of clear divergence with our data. 482 00:23:35.650 --> 00:23:39.590 I can notice that around this time frame, 483 00:23:39.590 --> 00:23:41.500 after the July quarter, there seems 484 00:23:41.500 --> 00:23:44.500 to be a divergence between median revenue and customer 485 00:23:44.500 --> 00:23:45.340 account. 486 00:23:45.340 --> 00:23:49.150 And I knew that was because we had started a new marketing 487 00:23:49.150 --> 00:23:49.680 campaign. 488 00:23:49.680 --> 00:23:51.430 So looks like we're getting new customers, 489 00:23:51.430 --> 00:23:52.888 but they're actually spending less. 490 00:23:52.888 --> 00:23:55.220 So maybe not the campaign that we want to address. 491 00:23:55.220 --> 00:23:56.890 So easily getting insight from Sigma. 492 00:23:59.460 --> 00:24:01.870 And, additionally while you're charting, 493 00:24:01.870 --> 00:24:04.630 you're still in the data prep environment. 494 00:24:04.630 --> 00:24:07.490 So, for example, if I want to filter my data down so 495 00:24:07.490 --> 00:24:07.990 we can use. 496 00:24:07.990 --> 00:24:10.650 We have a really robust filtering console, 497 00:24:10.650 --> 00:24:13.750 where I can say I just want to look at the 12 months of data. 498 00:24:13.750 --> 00:24:16.740 I can actually do a relative data filter 499 00:24:16.740 --> 00:24:22.530 on months where I can say, show me the previous 12 months data. 500 00:24:22.530 --> 00:24:24.530 That updates our data and the accompanying chart 501 00:24:24.530 --> 00:24:27.670 as well, just like so. 502 00:24:27.670 --> 00:24:30.505 You can create as many charts as you like. 503 00:24:30.505 --> 00:24:31.880 When you're done creating charts, 504 00:24:31.880 --> 00:24:34.110 you can add them to a dashboard. 505 00:24:34.110 --> 00:24:36.350 Let me give you an example of what a dashboard looks 506 00:24:36.350 --> 00:24:38.054 like in Snowflake. 507 00:24:41.400 --> 00:24:43.030 This is an example of a dashboard. 508 00:24:43.030 --> 00:24:45.440 You notice they all load up really quickly. 509 00:24:45.440 --> 00:24:47.440 And since the dashboard and the charts 510 00:24:47.440 --> 00:24:50.880 are sitting on top of live data on Snowflake, and not extracts, 511 00:24:50.880 --> 00:24:54.745 they're always live and the data is always up to date. 512 00:24:54.745 --> 00:24:57.120 So you can see I can easily format a pivot table, an area 513 00:24:57.120 --> 00:24:58.660 chart. 514 00:24:58.660 --> 00:25:02.510 Kind of an example of a chart that I was working with before. 515 00:25:02.510 --> 00:25:05.380 Additionally, you can add filters to your charts 516 00:25:05.380 --> 00:25:07.160 and have them applied to specific charts. 517 00:25:07.160 --> 00:25:08.660 So, for example, I only want to look 518 00:25:08.660 --> 00:25:10.490 at data that comes from users that 519 00:25:10.490 --> 00:25:13.430 have at least 7,000 revenue. 520 00:25:13.430 --> 00:25:17.460 And you can notice how some charts kind of get 521 00:25:17.460 --> 00:25:18.752 updated like that. 522 00:25:22.360 --> 00:25:23.760 That's the demo of Sigma. 523 00:25:35.650 --> 00:25:36.470 Great. 524 00:25:36.470 --> 00:25:37.610 I think I'm up. 525 00:25:37.610 --> 00:25:39.350 My name is Dominic Go. 526 00:25:39.350 --> 00:25:40.353 And I can get started. 527 00:25:40.353 --> 00:25:41.520 We can go to the next slide. 528 00:25:44.950 --> 00:25:47.150 So like I said, my name is Dominic Go. 529 00:25:47.150 --> 00:25:49.260 And I'm the Director of Analytics 530 00:25:49.260 --> 00:25:51.370 at Schoola and Olivela. 531 00:25:51.370 --> 00:25:53.880 And I thought I would just give a quick overview 532 00:25:53.880 --> 00:25:56.700 of what these companies do and how I got involved. 533 00:25:56.700 --> 00:26:01.080 So Schoola up was actually the first brand 534 00:26:01.080 --> 00:26:03.150 and it was launched in 2013. 535 00:26:03.150 --> 00:26:08.040 And it's an e-commerce platform that sells used clothing. 536 00:26:08.040 --> 00:26:13.200 I joined the company as the first analytics hire in 2015. 537 00:26:13.200 --> 00:26:14.880 And I was asked to basically set up 538 00:26:14.880 --> 00:26:17.250 the analytics infrastructure. 539 00:26:17.250 --> 00:26:19.800 Since then we've launched a second brand, which 540 00:26:19.800 --> 00:26:22.860 deals with high fashion goods, so it's 541 00:26:22.860 --> 00:26:25.590 a completely different website, but sitting on 542 00:26:25.590 --> 00:26:29.010 the same existing infrastructure and there's 543 00:26:29.010 --> 00:26:32.300 a lot of overlapping business processes and stuff. 544 00:26:32.300 --> 00:26:39.420 And that launched in the summer of 2017. 545 00:26:39.420 --> 00:26:40.920 So we can move on to the next slide, 546 00:26:40.920 --> 00:26:43.230 but that's just a quick overview of what we do. 547 00:26:46.770 --> 00:26:50.530 So I wanted to give a quick look of where we were coming from. 548 00:26:50.530 --> 00:26:54.100 So essentially, when I joined Schoola, as I mentioned, 549 00:26:54.100 --> 00:26:56.880 I was the first analytics hire. 550 00:26:56.880 --> 00:27:01.420 And so I had the invariable question of build or buy. 551 00:27:01.420 --> 00:27:04.390 And given our low resources, I decided 552 00:27:04.390 --> 00:27:08.320 that build was going to be the easier avenue for us 553 00:27:08.320 --> 00:27:09.430 at that time. 554 00:27:09.430 --> 00:27:12.610 So I essentially built our entire data warehouse 555 00:27:12.610 --> 00:27:17.050 on a single EC2 analytics server. 556 00:27:17.050 --> 00:27:21.340 As you can see here, we had a number of local data sources. 557 00:27:21.340 --> 00:27:25.000 So you see the Schoola database and Olivela databases, 558 00:27:25.000 --> 00:27:27.640 as well as our operations databases. 559 00:27:27.640 --> 00:27:29.710 And then we had a number of Cloud software 560 00:27:29.710 --> 00:27:36.040 as a service to SAS sources to cover various other business 561 00:27:36.040 --> 00:27:40.690 needs, so things like Zendesk for customer service, Silver 562 00:27:40.690 --> 00:27:45.280 Pop was our ESP at the time, Quickbooks, et cetera. 563 00:27:45.280 --> 00:27:47.290 And so we integrated all this data 564 00:27:47.290 --> 00:27:49.870 and brought it into a single source, as you want 565 00:27:49.870 --> 00:27:51.250 to do for a data warehouse. 566 00:27:51.250 --> 00:27:54.100 And then sort it all into a single schema. 567 00:27:54.100 --> 00:27:58.120 And then, on top of that, we built a web application 568 00:27:58.120 --> 00:28:01.660 to basically surface the data to our users. 569 00:28:01.660 --> 00:28:06.818 So you can go to the next slide to show just a quick look 570 00:28:06.818 --> 00:28:07.610 at what that looks. 571 00:28:07.610 --> 00:28:11.700 So this was just a completely homegrown application. 572 00:28:11.700 --> 00:28:14.000 As you can see, the screenshot, to the left, 573 00:28:14.000 --> 00:28:15.590 there is kind of like a repository 574 00:28:15.590 --> 00:28:18.260 of the existing reports and things that we've built. 575 00:28:18.260 --> 00:28:20.360 And then an example of one of the reports 576 00:28:20.360 --> 00:28:25.100 here from one of our old revenue reports. 577 00:28:25.100 --> 00:28:27.110 So that's where we were. 578 00:28:27.110 --> 00:28:33.170 Now I want to just talk a little bit about why this was painful. 579 00:28:33.170 --> 00:28:39.110 So we actually, as a company, had a very good data 580 00:28:39.110 --> 00:28:41.060 driven culture. 581 00:28:41.060 --> 00:28:46.790 And what I mean by that is that from our leadership team all 582 00:28:46.790 --> 00:28:49.340 the way down to the warehouse associates, 583 00:28:49.340 --> 00:28:52.340 everyone was interested in what the data could teach us 584 00:28:52.340 --> 00:28:58.070 about how to improve our operations or our marketing 585 00:28:58.070 --> 00:28:59.700 initiatives, et cetera. 586 00:28:59.700 --> 00:29:04.130 But in the infrastructure that I showed you, 587 00:29:04.130 --> 00:29:08.240 getting data to the business users 588 00:29:08.240 --> 00:29:11.520 required significant technical skills. 589 00:29:11.520 --> 00:29:16.400 So even if we were able to get it to them in a raw CSV, 590 00:29:16.400 --> 00:29:21.500 they would need very strong Excel skills to get around 591 00:29:21.500 --> 00:29:23.850 and do things with it. 592 00:29:23.850 --> 00:29:27.020 In addition to that, as our data sources grew 593 00:29:27.020 --> 00:29:31.100 and as the company grew, we had very large queries, 594 00:29:31.100 --> 00:29:35.270 very complex queries, that stored a lot of business logic. 595 00:29:35.270 --> 00:29:39.950 And then, even after we got those things developed, 596 00:29:39.950 --> 00:29:41.780 they were even harder to maintain. 597 00:29:41.780 --> 00:29:44.750 And then on top of all those things, all those business 598 00:29:44.750 --> 00:29:48.680 related activities, you had things like server space 599 00:29:48.680 --> 00:29:53.120 and compute time and things like that to worry about. 600 00:29:53.120 --> 00:29:55.520 So I love this analogy of If You Give A Mouse A 601 00:29:55.520 --> 00:29:57.320 Cookie because basically any time I 602 00:29:57.320 --> 00:29:59.330 would finish a report for somebody else, 603 00:29:59.330 --> 00:30:01.310 there were about three other questions that 604 00:30:01.310 --> 00:30:03.470 came out of that report that required 605 00:30:03.470 --> 00:30:05.720 additional development. 606 00:30:05.720 --> 00:30:10.880 And so in the time leading up to the launch of our second brand 607 00:30:10.880 --> 00:30:13.820 of Olivela, I saw a great kind of turning point 608 00:30:13.820 --> 00:30:16.790 for the company to make a change. 609 00:30:16.790 --> 00:30:20.480 So we can move on to the next slide. 610 00:30:20.480 --> 00:30:21.930 And so it was at that time that I 611 00:30:21.930 --> 00:30:24.840 decided we needed something more sustainable, 612 00:30:24.840 --> 00:30:28.500 something more easy to collaborate on, 613 00:30:28.500 --> 00:30:31.320 for both our business users and for our growing data 614 00:30:31.320 --> 00:30:35.830 team to maintain and grow our ETL jobs 615 00:30:35.830 --> 00:30:38.700 and set some more business standards about how 616 00:30:38.700 --> 00:30:39.790 to do things. 617 00:30:39.790 --> 00:30:41.850 So there were a number of things that I 618 00:30:41.850 --> 00:30:43.690 had to consider while making this decision. 619 00:30:43.690 --> 00:30:47.610 So the first one, and the biggest one, was cost. 620 00:30:47.610 --> 00:30:51.390 So there's tons of BI tools out on the market, from ETL 621 00:30:51.390 --> 00:30:55.290 to user level tools to choose from. 622 00:30:55.290 --> 00:30:56.910 I've had the fortunate experience 623 00:30:56.910 --> 00:31:00.090 of working with a lot of them firsthand. 624 00:31:00.090 --> 00:31:03.360 But we're a small company, we're still a startup. 625 00:31:03.360 --> 00:31:05.940 I mean, even though we started 2013, 626 00:31:05.940 --> 00:31:08.850 we're still very much a startup, growing, 627 00:31:08.850 --> 00:31:12.630 and so we need to manage costs very carefully. 628 00:31:12.630 --> 00:31:17.520 And so I wanted something that could provide immediate value, 629 00:31:17.520 --> 00:31:20.320 but also be cost effective. 630 00:31:20.320 --> 00:31:22.260 In similar, and related to that, is 631 00:31:22.260 --> 00:31:23.712 that I need I knew that it needed 632 00:31:23.712 --> 00:31:25.170 to be able to grow with the company 633 00:31:25.170 --> 00:31:28.410 because we have a very rapid growth trajectory right now. 634 00:31:28.410 --> 00:31:30.638 And then, in addition to that, I wanted 635 00:31:30.638 --> 00:31:33.180 to make sure it was something that people would actually use. 636 00:31:33.180 --> 00:31:36.060 I've seen, far too many times, people 637 00:31:36.060 --> 00:31:39.540 will be introduced to a BI tool that they use for a week 638 00:31:39.540 --> 00:31:45.050 and then just forget about and switch back to downloading CSV, 639 00:31:45.050 --> 00:31:47.130 so that they can put it into Excel. 640 00:31:47.130 --> 00:31:49.320 And then finally I wanted something, 641 00:31:49.320 --> 00:31:51.240 like I mentioned earlier, that was easier 642 00:31:51.240 --> 00:31:54.960 to maintain after we get put all that complex business 643 00:31:54.960 --> 00:31:56.640 logic into place. 644 00:31:56.640 --> 00:32:01.240 So we can move on to the next slide. 645 00:32:01.240 --> 00:32:03.660 So this is where we moved to. 646 00:32:03.660 --> 00:32:06.510 So you've got the same sort of list of data sources 647 00:32:06.510 --> 00:32:10.290 and a number of others working there on the left. 648 00:32:10.290 --> 00:32:13.110 And then for our ETL tool we actually 649 00:32:13.110 --> 00:32:15.360 chose one of the other Partner Connect 650 00:32:15.360 --> 00:32:19.500 tools, that wasn't listed above, but it's called Matilion. 651 00:32:19.500 --> 00:32:24.750 And then we put all of that into a Snowflake data warehouse. 652 00:32:24.750 --> 00:32:31.110 And then on top of that, we have Sigma is our primary business 653 00:32:31.110 --> 00:32:32.670 user tool. 654 00:32:32.670 --> 00:32:34.420 So we can move to our next slide. 655 00:32:37.600 --> 00:32:40.450 So, again, it addresses all the main issues here 656 00:32:40.450 --> 00:32:41.650 that I was talking about. 657 00:32:41.650 --> 00:32:48.550 So Snowflake and Sigma are both completely usage based 658 00:32:48.550 --> 00:32:49.780 and compute usage based. 659 00:32:49.780 --> 00:32:52.270 So that meant as I was rolling it out 660 00:32:52.270 --> 00:32:57.340 there were really only me and one other user, an analyst, 661 00:32:57.340 --> 00:33:00.040 working on the platforms as we set up the infrastructure. 662 00:33:00.040 --> 00:33:04.372 Which meant that our monthly costs were very, very low 663 00:33:04.372 --> 00:33:06.080 as we were getting things up and running, 664 00:33:06.080 --> 00:33:07.622 which was great because that meant we 665 00:33:07.622 --> 00:33:09.070 had time to validate the data. 666 00:33:09.070 --> 00:33:14.220 And then we could really set our business standards well. 667 00:33:14.220 --> 00:33:18.190 And that allowed us then to also onboard people, one at a time 668 00:33:18.190 --> 00:33:21.820 or a small team at a time, so that we could ensure 669 00:33:21.820 --> 00:33:27.160 that the people received the training and the knowledge 670 00:33:27.160 --> 00:33:30.430 that they need to actually make the tool useful for them. 671 00:33:30.430 --> 00:33:33.460 And then, because we had all of that time leading up 672 00:33:33.460 --> 00:33:35.170 before we had to scale and introduce 673 00:33:35.170 --> 00:33:38.500 to the entire organization, that meant 674 00:33:38.500 --> 00:33:41.620 we could set up standards to make maintenance very simple 675 00:33:41.620 --> 00:33:43.030 for us. 676 00:33:43.030 --> 00:33:46.630 So we can move on to the next one. 677 00:33:46.630 --> 00:33:54.480 So as I was preparing for this, I 678 00:33:54.480 --> 00:33:58.140 was thinking about this specific example of a win for us. 679 00:33:58.140 --> 00:34:00.540 And one really big one was we have 680 00:34:00.540 --> 00:34:03.030 a dashboard that we call the Cohort 681 00:34:03.030 --> 00:34:06.240 Dashboard at our company, which allows us to look at-- 682 00:34:06.240 --> 00:34:08.610 actually very similar to the example 683 00:34:08.610 --> 00:34:12.600 they ran through earlier, we cohort our customers 684 00:34:12.600 --> 00:34:16.770 by the month that they joined and through the acquisition 685 00:34:16.770 --> 00:34:18.179 channel that they joined. 686 00:34:18.179 --> 00:34:24.420 And so before this cohort dashboard 687 00:34:24.420 --> 00:34:27.690 was over 700 lines of code and PHP 688 00:34:27.690 --> 00:34:33.210 and another 500 lines of code and manually maintained SQL. 689 00:34:33.210 --> 00:34:37.110 Now it's a Matillion job with about five nodes. 690 00:34:37.110 --> 00:34:39.810 And then a Sigma dashboard, which 691 00:34:39.810 --> 00:34:43.949 took a little bit, maybe a week, to put together 692 00:34:43.949 --> 00:34:46.800 after collecting all the business requirements, 693 00:34:46.800 --> 00:34:50.520 but then an update is as simple as five minutes. 694 00:34:50.520 --> 00:34:52.080 So, yeah. 695 00:34:52.080 --> 00:34:56.070 So we can move on to the next slide. 696 00:34:56.070 --> 00:35:02.190 So as I was thinking about my experience with the tool, 697 00:35:02.190 --> 00:35:06.210 Sigma in particular has grown and adapted a lot 698 00:35:06.210 --> 00:35:08.818 in there they're constantly making great improvements 699 00:35:08.818 --> 00:35:09.360 to the tools. 700 00:35:09.360 --> 00:35:12.090 But some things that I thought I'd 701 00:35:12.090 --> 00:35:13.840 share that I've learned along the way. 702 00:35:13.840 --> 00:35:17.640 So the biggest thing is that technology is never 703 00:35:17.640 --> 00:35:20.610 going to-- well maybe someday I can't speak forever, 704 00:35:20.610 --> 00:35:23.070 but right now technology is not solving 705 00:35:23.070 --> 00:35:27.720 for the good organizational and logical conventions. 706 00:35:27.720 --> 00:35:31.020 So those sorts of things still very much matter. 707 00:35:31.020 --> 00:35:36.150 So while traditional things in data warehousing, like really 708 00:35:36.150 --> 00:35:40.860 clearly defining dimensions and facts and things like that, 709 00:35:40.860 --> 00:35:45.060 there's no longer the technical need for those things to exist. 710 00:35:45.060 --> 00:35:46.890 As they mentioned earlier, Snowflake 711 00:35:46.890 --> 00:35:49.950 is and no index environment. 712 00:35:49.950 --> 00:35:54.150 But the logical organization that those traditional data 713 00:35:54.150 --> 00:35:56.820 warehouse standards provided really 714 00:35:56.820 --> 00:36:00.600 helped me create a good environment where our business 715 00:36:00.600 --> 00:36:04.770 users could do something directly 716 00:36:04.770 --> 00:36:07.380 without a ton of guidance. 717 00:36:07.380 --> 00:36:11.850 So I list out a few things here as just 718 00:36:11.850 --> 00:36:15.630 like really precise things that I've done. 719 00:36:15.630 --> 00:36:19.285 One in particular is using explicit ID column names 720 00:36:19.285 --> 00:36:20.160 and things like that. 721 00:36:20.160 --> 00:36:22.470 Because that just helps your business users, 722 00:36:22.470 --> 00:36:26.730 if they decide to build their own worksheet and do a join, 723 00:36:26.730 --> 00:36:30.700 it will help them to be successful. 724 00:36:30.700 --> 00:36:33.000 The second big thing that I learned 725 00:36:33.000 --> 00:36:36.540 was that you need to train the users with relevant data. 726 00:36:36.540 --> 00:36:40.080 So when I was first starting with Sigma, 727 00:36:40.080 --> 00:36:41.850 I was very, very excited about the tool. 728 00:36:41.850 --> 00:36:44.790 I wanted to get it out there as quickly as possible. 729 00:36:44.790 --> 00:36:49.740 But, as we mentioned earlier, on our merchandising team, 730 00:36:49.740 --> 00:36:51.600 I showed her some sales data, which 731 00:36:51.600 --> 00:36:55.300 is not very relevant to what she is doing. 732 00:36:55.300 --> 00:36:59.950 She was much more interested in our inventory levels. 733 00:36:59.950 --> 00:37:01.950 And so it just didn't stick. 734 00:37:01.950 --> 00:37:04.980 I hadn't cleaned up the data warehouse columns or anything 735 00:37:04.980 --> 00:37:05.520 yet. 736 00:37:05.520 --> 00:37:08.850 And so it ended up being a wasted effort 737 00:37:08.850 --> 00:37:11.620 until a few months later when I had 738 00:37:11.620 --> 00:37:16.940 set those good conventions and such so 739 00:37:16.940 --> 00:37:18.990 that it was more relevant to her and could 740 00:37:18.990 --> 00:37:20.940 provide immediate value. 741 00:37:20.940 --> 00:37:25.500 And then the final piece here is that, unlike other BI 742 00:37:25.500 --> 00:37:28.260 tools on the market right now, there's not really 743 00:37:28.260 --> 00:37:32.760 a data modeling layer. 744 00:37:32.760 --> 00:37:36.240 Which is a really, really important consideration 745 00:37:36.240 --> 00:37:38.970 as you decide to roll this out because a lot 746 00:37:38.970 --> 00:37:43.710 of the kind of traditional guardrails are not there. 747 00:37:43.710 --> 00:37:46.020 What it does allow and why I think it's better 748 00:37:46.020 --> 00:37:51.270 is that it allows for the business users 749 00:37:51.270 --> 00:37:57.120 to be much closer to the actual data and how it's produced. 750 00:37:57.120 --> 00:37:58.770 And the reason that's important is 751 00:37:58.770 --> 00:38:01.650 because they're the ones often that are producing it 752 00:38:01.650 --> 00:38:03.600 or that know it best. 753 00:38:03.600 --> 00:38:05.640 And so I think that this actually 754 00:38:05.640 --> 00:38:10.410 is a great innovation by Sigma, but it also 755 00:38:10.410 --> 00:38:12.720 means that it does lack those guardrails that 756 00:38:12.720 --> 00:38:17.030 used to be there with kind of more traditional modeling 757 00:38:17.030 --> 00:38:17.530 pieces. 758 00:38:17.530 --> 00:38:20.790 So that's just one other thing to consider as you're rolling 759 00:38:20.790 --> 00:38:23.040 this out to your organization. 760 00:38:23.040 --> 00:38:25.170 And I think that's it for me, so. 761 00:38:29.090 --> 00:38:29.590 All right. 762 00:38:29.590 --> 00:38:33.420 So now is when we're going to jump into some questions. 763 00:38:33.420 --> 00:38:36.720 So if you have any questions, feel free to type it in 764 00:38:36.720 --> 00:38:41.910 and I will communicate those questions to our panelists. 765 00:38:41.910 --> 00:38:46.850 We have one question to get us started. 766 00:38:46.850 --> 00:38:48.860 So this is a question about Sigma. 767 00:38:48.860 --> 00:38:51.890 And the question is, how does the functionality in Sigma 768 00:38:51.890 --> 00:38:52.715 compare to Excel? 769 00:38:59.270 --> 00:39:00.060 Hi, Erica. 770 00:39:02.488 --> 00:39:04.030 So this question was kind of like how 771 00:39:04.030 --> 00:39:08.380 does the functionality in Sigma compared to Excel. 772 00:39:08.380 --> 00:39:11.210 So it's very similar. 773 00:39:11.210 --> 00:39:13.000 So we definitely try to model what 774 00:39:13.000 --> 00:39:16.630 we do around that spreadsheet interface. 775 00:39:16.630 --> 00:39:20.270 So you can find all the typical arithmetic functions. 776 00:39:20.270 --> 00:39:23.360 For example, if you want to do if statements, fee supportive 777 00:39:23.360 --> 00:39:27.640 statements, then you can nest any of our other functions 778 00:39:27.640 --> 00:39:31.130 within an if statement to make a more complex if statement just 779 00:39:31.130 --> 00:39:32.110 like that. 780 00:39:37.500 --> 00:39:40.620 And a question for Snowflake. 781 00:39:40.620 --> 00:39:44.400 Why do people tend to move from their current data warehouse 782 00:39:44.400 --> 00:39:45.570 over to Snowflake? 783 00:39:45.570 --> 00:39:47.680 Like what's the main difference? 784 00:39:47.680 --> 00:39:50.500 Like what's the pain point that brings? 785 00:39:50.500 --> 00:39:53.320 Well, I'd say that the most common pain 786 00:39:53.320 --> 00:39:55.180 point that people are running into 787 00:39:55.180 --> 00:39:57.400 is that they just don't have flexibility. 788 00:39:57.400 --> 00:40:02.140 So let's say that they have a couple hundred or maybe even 789 00:40:02.140 --> 00:40:07.210 just a couple dozen BI users who are accessing the data 790 00:40:07.210 --> 00:40:11.200 warehouse through their BI implementation through Sigma. 791 00:40:11.200 --> 00:40:15.070 Well, if they're using a traditional data warehouse, 792 00:40:15.070 --> 00:40:17.170 even when it's in the Cloud, oftentimes you'll 793 00:40:17.170 --> 00:40:19.997 find that they have a lot of concurrency issues 794 00:40:19.997 --> 00:40:22.330 where you have too many people trying to access the data 795 00:40:22.330 --> 00:40:24.100 warehouse at the same time. 796 00:40:24.100 --> 00:40:25.900 And you simply don't have the throughput 797 00:40:25.900 --> 00:40:29.460 that you'd need to be able to respond to that problem. 798 00:40:29.460 --> 00:40:31.450 Well, with Snowflake you can scale up and down 799 00:40:31.450 --> 00:40:34.270 to meet that demand and very easily satisfy 800 00:40:34.270 --> 00:40:37.510 all of those people who are trying to access data. 801 00:40:37.510 --> 00:40:39.310 With traditional data warehouses you really 802 00:40:39.310 --> 00:40:42.962 can't do that, even data warehouses in the Cloud. 803 00:40:42.962 --> 00:40:44.920 Another thing that's really nice with Snowflake 804 00:40:44.920 --> 00:40:46.253 is that it's really predictable. 805 00:40:46.253 --> 00:40:48.970 So you know exactly how much you're going to be paying. 806 00:40:48.970 --> 00:40:51.280 You know exactly how long a query is going to take. 807 00:40:51.280 --> 00:40:55.930 You know exactly what you can do to be able to get that query 808 00:40:55.930 --> 00:40:58.420 to be satisfied quicker by using more resources 809 00:40:58.420 --> 00:41:00.790 or using resources in a different way. 810 00:41:00.790 --> 00:41:03.220 So it's predictable and it's something 811 00:41:03.220 --> 00:41:07.080 that you can very easily adapt to your situation. 812 00:41:11.860 --> 00:41:14.170 This next question is about Sigma. 813 00:41:14.170 --> 00:41:17.020 And the question is, where does the data processing in Sigma 814 00:41:17.020 --> 00:41:17.530 occur? 815 00:41:17.530 --> 00:41:21.160 Are you moving the data into Sigma and then working on it? 816 00:41:21.160 --> 00:41:23.470 Or is it staying in the warehouse? 817 00:41:39.190 --> 00:41:40.180 OK. 818 00:41:40.180 --> 00:41:42.810 It looks like we may have lost Ali, 819 00:41:42.810 --> 00:41:47.520 so we can come back to that question. 820 00:41:47.520 --> 00:41:49.850 The next question is for-- 821 00:41:52.857 --> 00:41:53.940 the next question is for-- 822 00:41:53.940 --> 00:41:54.440 Sorry. 823 00:41:54.440 --> 00:41:55.640 I can get to that question. 824 00:41:55.640 --> 00:41:56.765 I can get to that question. 825 00:41:56.765 --> 00:41:58.460 I'm sorry. 826 00:41:58.460 --> 00:42:02.150 Yeah, so basically all of the data 827 00:42:02.150 --> 00:42:03.590 as processed within the database, 828 00:42:03.590 --> 00:42:06.470 it's processed within Snowflake. 829 00:42:06.470 --> 00:42:10.340 So all that Sigma does is it creates a SQL query, 830 00:42:10.340 --> 00:42:12.257 pushes that to Snowflake and executes it 831 00:42:12.257 --> 00:42:13.090 within the database. 832 00:42:13.090 --> 00:42:15.382 So we're not we're actually pulling a full data set out 833 00:42:15.382 --> 00:42:17.552 of Snowflake to be executed. 834 00:42:21.480 --> 00:42:21.980 All right. 835 00:42:21.980 --> 00:42:24.080 Thank you. 836 00:42:24.080 --> 00:42:29.230 The next question is for Dominic about Olivela. 837 00:42:29.230 --> 00:42:32.500 The question was about the data modeling 838 00:42:32.500 --> 00:42:37.840 and how Sigma doesn't have the data model. 839 00:42:37.840 --> 00:42:41.910 How does that work in your implementation? 840 00:42:41.910 --> 00:42:43.550 Yeah, great question. 841 00:42:43.550 --> 00:42:48.320 So in our implementation, it really 842 00:42:48.320 --> 00:42:53.120 is dependent on the level of expertise of the business user, 843 00:42:53.120 --> 00:42:57.320 if it was like an Excel power user I like to think about, 844 00:42:57.320 --> 00:43:02.480 then I would expose more of the data access capabilities 845 00:43:02.480 --> 00:43:06.440 to the end users and explain how the joins work. 846 00:43:06.440 --> 00:43:08.740 They do a really good job of explaining, 847 00:43:08.740 --> 00:43:11.930 in the interface, how joining tables work 848 00:43:11.930 --> 00:43:16.670 and what you can expect from the resulting set. 849 00:43:16.670 --> 00:43:19.220 So, for me, it was always a training choice 850 00:43:19.220 --> 00:43:22.550 to assess the level of technical expertise 851 00:43:22.550 --> 00:43:25.040 and to know whether or not I could expose 852 00:43:25.040 --> 00:43:28.260 that area of the tool to them. 853 00:43:28.260 --> 00:43:31.400 In a lot of cases, I was able to do that. 854 00:43:31.400 --> 00:43:35.270 And then for others, who are more like consumers or very 855 00:43:35.270 --> 00:43:40.010 light manipulators , people who will filter, sort columns, 856 00:43:40.010 --> 00:43:43.460 those sorts of things, we would tend to have the data team 857 00:43:43.460 --> 00:43:47.150 build pre-existing worksheets or dashboards that they could then 858 00:43:47.150 --> 00:43:52.000 save a copy for themselves to manipulate et cetera. 859 00:43:52.000 --> 00:43:55.190 And so it is it's always a very kind of calculated training 860 00:43:55.190 --> 00:43:59.360 choice, depending on the team and everything, 861 00:43:59.360 --> 00:44:01.190 knowing that those guardrails aren't there. 862 00:44:01.190 --> 00:44:03.110 But like I mentioned earlier, I think 863 00:44:03.110 --> 00:44:08.780 that it's very valuable to have that closeness of the business 864 00:44:08.780 --> 00:44:12.470 user to the actual data that is output by the systems 865 00:44:12.470 --> 00:44:14.780 that they're using because they'll understand 866 00:44:14.780 --> 00:44:16.490 better what they're doing. 867 00:44:16.490 --> 00:44:18.650 When there are data quality issues, 868 00:44:18.650 --> 00:44:22.520 it's not some data analysts trying to put in some business 869 00:44:22.520 --> 00:44:23.720 logic to solve for that. 870 00:44:23.720 --> 00:44:25.850 Instead, it gets exposed to the end user 871 00:44:25.850 --> 00:44:29.510 and they can see what happened with the data quality 872 00:44:29.510 --> 00:44:33.080 and why they need to improve their processes 873 00:44:33.080 --> 00:44:34.650 or whatever the case may be. 874 00:44:34.650 --> 00:44:37.910 So, in my experience, it's been a win. 875 00:44:37.910 --> 00:44:44.060 But it does mean that it shifts a lot of priority 876 00:44:44.060 --> 00:44:45.065 on that training aspect. 877 00:44:49.680 --> 00:44:50.180 All right. 878 00:44:50.180 --> 00:44:50.680 Thank you. 879 00:44:50.680 --> 00:44:54.570 This dovetails well into the next question, 880 00:44:54.570 --> 00:44:56.960 which is for Sigma, so for Ali. 881 00:44:56.960 --> 00:45:01.070 And the question was, do all Sigma users 882 00:45:01.070 --> 00:45:03.670 have access to the entire database 883 00:45:03.670 --> 00:45:08.220 or are there levels of access that can be set up? 884 00:45:08.220 --> 00:45:11.530 Yes, so there's definitely levels of access 885 00:45:11.530 --> 00:45:14.280 that can be set up. 886 00:45:14.280 --> 00:45:16.108 So the user is part of a specific team, 887 00:45:16.108 --> 00:45:17.900 like I showed at the beginning of the demo, 888 00:45:17.900 --> 00:45:19.950 as part of the product solutions team. 889 00:45:19.950 --> 00:45:25.500 You can determine which database or data connection, if you want 890 00:45:25.500 --> 00:45:27.923 them to have a authoring usage, meaning they 891 00:45:27.923 --> 00:45:29.340 can connect directly to the table. 892 00:45:29.340 --> 00:45:31.255 Or if you want to just be a reader, meaning 893 00:45:31.255 --> 00:45:32.880 they can only work with worksheets that 894 00:45:32.880 --> 00:45:34.880 have been curated for them. 895 00:45:34.880 --> 00:45:40.280 And so it is possible that for one data connection 896 00:45:40.280 --> 00:45:44.190 they can have author access to user tables and for another one 897 00:45:44.190 --> 00:45:46.555 they can only have reader access where they only 898 00:45:46.555 --> 00:45:47.680 work with curated versions. 899 00:45:54.310 --> 00:45:54.810 All right. 900 00:45:54.810 --> 00:45:58.460 And it looks like that is all of the questions 901 00:45:58.460 --> 00:46:01.350 that we've gotten from our attendees. 902 00:46:01.350 --> 00:46:06.320 So thank you so much to our panelists for working us 903 00:46:06.320 --> 00:46:09.390 through everything with Snowflakes, 904 00:46:09.390 --> 00:46:13.310 Sigma, and how Olivela has implemented. 905 00:46:13.310 --> 00:46:15.920 And that's the end of our webinar. 906 00:46:15.920 --> 00:46:19.490 So have a great day everyone. 907 00:46:19.490 --> 00:46:21.120 Thank you.