1 00:00:05,768 --> 00:00:07,793 [Announcer]: Welcome to the Analytics Power Hour. 2 00:00:08,474 --> 00:00:12,724 [Announcer]: Analytics topics covered conversationally and sometimes with explicit language. 3 00:00:14,648 --> 00:00:16,573 [Michael Helbling]: Hey everybody, it's the Analytics Power Hour. 4 00:00:16,793 --> 00:00:19,018 [Michael Helbling]: This is episode 286. 5 00:00:19,437 --> 00:00:21,861 [Michael Helbling]: It's Tuesday and I know what you're thinking. 6 00:00:22,562 --> 00:00:26,668 [Michael Helbling]: I sure hope revenue and active customers still mean the same thing as they did yesterday. 7 00:00:26,688 --> 00:00:30,313 [Michael Helbling]: A lot of you know firsthand the pain I'm describing. 8 00:00:30,434 --> 00:00:36,703 [Michael Helbling]: There's data ping-pongs around the business taking on shapes and definitions that were never really intended. 9 00:00:36,683 --> 00:00:39,827 [Michael Helbling]: Well, the semantic layer was supposed to take care of all that. 10 00:00:40,008 --> 00:00:44,313 [Michael Helbling]: And to be fair, there are some nice, tidy businesses out there doing a great job. 11 00:00:45,034 --> 00:00:50,782 [Michael Helbling]: But most of us are still trying to figure out where it should live, what it should be written with, and who should own it. 12 00:00:50,922 --> 00:00:52,665 [Michael Helbling]: So I think we should dig into it. 13 00:00:53,206 --> 00:00:55,869 [Michael Helbling]: But first, let me introduce my co-host, Moee Kisss. 14 00:00:56,450 --> 00:00:57,091 [Michael Helbling]: How are you going? 15 00:00:58,272 --> 00:00:59,334 [Moe Kiss]: I'm going great. 16 00:00:59,454 --> 00:01:00,936 [Moe Kiss]: I'm really excited about this. 17 00:01:01,085 --> 00:01:04,535 [Michael Helbling]: I'm excited to, and excited to do the show with you. 18 00:01:04,796 --> 00:01:07,224 [Michael Helbling]: And Tim Wilson, howdy. 19 00:01:07,244 --> 00:01:11,115 [Michael Helbling]: I think it's just all semantics. 20 00:01:11,878 --> 00:01:13,643 [Michael Helbling]: It's all, oh, so good. 21 00:01:15,749 --> 00:01:19,673 [Michael Helbling]: Well, that's an interesting potential cop out. 22 00:01:19,953 --> 00:01:20,914 [Michael Helbling]: Okay, no. 23 00:01:20,934 --> 00:01:25,939 [Michael Helbling]: And I'm like, well, to really get into this topic, I think we found the perfect guest. 24 00:01:26,600 --> 00:01:30,624 [Michael Helbling]: Cindi Howson is the Chief Data and AI Strategy Officer at ThoughtSpot. 25 00:01:30,824 --> 00:01:35,228 [Michael Helbling]: She was previously Vice President at Gartner, along with many other distinguished roles throughout her career. 26 00:01:35,809 --> 00:01:41,094 [Michael Helbling]: She is the host of the Data Chief podcast and has authored many books on BI and data. 27 00:01:41,234 --> 00:01:43,176 [Michael Helbling]: And today she is our guest. 28 00:01:43,196 --> 00:01:44,878 [Michael Helbling]: Welcome to the show, Cindi. 29 00:01:45,365 --> 00:01:46,918 [Cindi Howson]: Thank you for having me, everyone. 30 00:01:46,939 --> 00:01:48,996 [Cindi Howson]: I'm so excited to be here. 31 00:01:49,178 --> 00:01:51,161 [Michael Helbling]: I am excited that you're here too. 32 00:01:51,181 --> 00:01:55,847 [Michael Helbling]: You're kind of, to me, sort of an OG of the data space. 33 00:01:56,048 --> 00:02:13,332 [Michael Helbling]: And so I love people who can provide as much depth and background and historical perspective on all the things we're struggling with in the world of data today that were struggles 20 years ago and still remain with us today, but with different tools and names and things like that. 34 00:02:13,412 --> 00:02:16,857 [Michael Helbling]: But today we're talking about sort of 35 00:02:16,972 --> 00:02:17,714 [Tim Wilson]: Oh, go ahead, Tim. 36 00:02:18,596 --> 00:02:29,499 [Tim Wilson]: Well, I was going to say, I mean, I, having done my little forensic sleuthing that I saw Cindi speak at a TDWI summit back in 2004, which I think is amazing since we're both like 35 years old. 37 00:02:31,463 --> 00:02:33,107 [Tim Wilson]: We just look at him. 38 00:02:33,712 --> 00:02:48,042 [Tim Wilson]: It was like my entry moving from technical writing to Markham, slightly into web analytics, and then it was like my entree into the world of analytics and BI was that conference. 39 00:02:48,595 --> 00:02:51,880 [Tim Wilson]: Yeah, so I just felt like I had an appetite. 40 00:02:51,900 --> 00:02:52,781 [Tim Wilson]: He's a big fat. 41 00:02:52,801 --> 00:02:53,923 [Tim Wilson]: I'm gonna say that. 42 00:02:53,943 --> 00:02:54,384 [Cindi Howson]: He's a big fat. 43 00:02:54,404 --> 00:02:55,125 [Cindi Howson]: It's the summary. 44 00:02:55,906 --> 00:03:05,121 [Cindi Howson]: Tim, now I feel like I have to send you an original BI scorecard black bear that I used to use as giveaways for class participation. 45 00:03:05,361 --> 00:03:06,503 [Cindi Howson]: We'll see if I still have one. 46 00:03:08,125 --> 00:03:14,535 [Tim Wilson]: Now, I couldn't remember the topic, but I feel like I was coming back and I was like, there were some other 47 00:03:15,241 --> 00:03:24,525 [Tim Wilson]: lady who spoke and it wasn't Jill Deshaix and the other person was like the only I cannot remember her name because we wound up actually Claudia Imhoff. 48 00:03:24,590 --> 00:03:34,343 [Tim Wilson]: Yes, like two months later called and had Claudia like come out and just spend two days explaining data warehouses to our team. 49 00:03:34,523 --> 00:03:40,592 [Tim Wilson]: So like doing the conference circuit and as a consultant, you're like, oh yeah, nobody's really just going to call you up. 50 00:03:40,652 --> 00:03:42,534 [Tim Wilson]: Like we literally, she's like, what do you want me to do? 51 00:03:42,675 --> 00:03:43,636 [Tim Wilson]: And we're like, you're smart. 52 00:03:44,177 --> 00:03:47,501 [Tim Wilson]: Please come just sit in a room and answer our questions for two days. 53 00:03:47,761 --> 00:03:48,943 [Tim Wilson]: And it was, 54 00:03:48,923 --> 00:03:51,127 [Tim Wilson]: Our dev team was like thrilled. 55 00:03:51,708 --> 00:03:54,252 [Tim Wilson]: And I'm so glad that you can remember that was who it was. 56 00:03:54,372 --> 00:03:55,033 [Tim Wilson]: Okay. 57 00:03:55,053 --> 00:03:56,355 [Tim Wilson]: Yeah. 58 00:03:56,375 --> 00:03:56,916 [Tim Wilson]: Okay. 59 00:03:56,936 --> 00:03:57,157 [Tim Wilson]: Michael. 60 00:03:57,197 --> 00:03:57,658 [Tim Wilson]: Okay. 61 00:03:57,738 --> 00:04:00,022 [Tim Wilson]: But now we're going to do the episode too. 62 00:04:00,242 --> 00:04:00,582 [Michael Helbling]: All right. 63 00:04:00,623 --> 00:04:02,145 [Michael Helbling]: Fast forward just a few years. 64 00:04:02,686 --> 00:04:04,168 [Michael Helbling]: Yeah. 65 00:04:04,189 --> 00:04:06,132 [Michael Helbling]: So nowadays, no. 66 00:04:07,674 --> 00:04:09,097 [Michael Helbling]: So let's talk about Star Schema. 67 00:04:09,297 --> 00:04:09,758 [Michael Helbling]: No. 68 00:04:10,138 --> 00:04:12,282 [Michael Helbling]: Well, I mean, we can start there. 69 00:04:12,616 --> 00:04:19,089 [Michael Helbling]: So semantic layers, Cindi, obviously everyone talks about those, but there's a history here. 70 00:04:19,189 --> 00:04:27,045 [Michael Helbling]: And maybe just to get us started and for people who aren't as familiar with the concept, maybe just a quick primer on sort of, what does that even mean? 71 00:04:27,405 --> 00:04:28,087 [Michael Helbling]: What are those? 72 00:04:28,428 --> 00:04:30,692 [Michael Helbling]: And we can kind of use that as a launching point. 73 00:04:31,550 --> 00:04:31,971 [Cindi Howson]: Sure. 74 00:04:32,091 --> 00:04:51,118 [Cindi Howson]: So in the simplest terms, a semantic layer provides a representation of the business model in business terms to the physical structures in your whatever, data warehouse, data lake, cloud data platform, whatever you want to call it. 75 00:04:51,999 --> 00:04:56,646 [Cindi Howson]: And it is important that it is in business terms. 76 00:04:56,626 --> 00:05:06,100 [Cindi Howson]: So if I think about, my German language has only served me well when I look at SAP original tables. 77 00:05:06,401 --> 00:05:11,108 [Cindi Howson]: VBAP was the customer table in SAP R3. 78 00:05:12,289 --> 00:05:17,797 [Cindi Howson]: So you could never show a field VBAP to a business person. 79 00:05:18,118 --> 00:05:20,862 [Cindi Howson]: Instead, you would say, this is customer name. 80 00:05:21,955 --> 00:05:27,205 [Moe Kiss]: And I keep hearing the word when people talk about semantic layers of context. 81 00:05:27,926 --> 00:05:32,234 [Moe Kiss]: And that seems like when you say business terms, is that what you mean? 82 00:05:32,295 --> 00:05:37,344 [Moe Kiss]: Like the context of the data, how it relates to each other, what the definition is. 83 00:05:38,025 --> 00:05:38,947 [Moe Kiss]: Is that the same thing? 84 00:05:39,007 --> 00:05:42,073 [Moe Kiss]: Or when you say business terms, are you thinking about something different? 85 00:05:42,425 --> 00:06:00,214 [Cindi Howson]: No, I think it's both because if we get precise, something like revenue, well, revenue in an inventory and supply chain context, I'm going to look at revenue based on when somebody placed an order. 86 00:06:00,194 --> 00:06:16,677 [Cindi Howson]: But if I take it in terms of finance, the Office of Finance, they're going to want to know when that invoice was paid, or if I'm doing, you know, if I'm a cash basis or a cruel basis. 87 00:06:16,757 --> 00:06:20,562 [Cindi Howson]: So the context of that revenue field matters. 88 00:06:21,484 --> 00:06:22,906 [Cindi Howson]: Did that answer your question, Moe? 89 00:06:23,707 --> 00:06:24,568 [Cindi Howson]: Yeah, absolutely. 90 00:06:24,588 --> 00:06:25,289 [Cindi Howson]: Okay, totally. 91 00:06:26,170 --> 00:06:28,834 [Tim Wilson]: Not to head, I struggle with 92 00:06:29,388 --> 00:06:32,665 [Tim Wilson]: And revenue is a great example because it's... 93 00:06:32,932 --> 00:06:47,485 [Tim Wilson]: part of the challenge, it's like we're trying to find a technology or a tool or a process to solve for something where the business, the person in finance, when they think revenue, they're always thinking in a revenue recognition world. 94 00:06:47,505 --> 00:07:00,036 [Tim Wilson]: And when somebody's an inventory, they're always thinking of it another way and don't even necessarily, and they both may complain that I see reports from the other department and they have revenue and it's wrong. 95 00:07:00,136 --> 00:07:02,938 [Tim Wilson]: It becomes a, 96 00:07:02,918 --> 00:07:13,276 [Tim Wilson]: business understanding challenge that data processing technique mechanism is trying to solve. 97 00:07:13,296 --> 00:07:15,720 [Tim Wilson]: Or am I just being cynical about that? 98 00:07:15,780 --> 00:07:17,383 [Cindi Howson]: No, it is. 99 00:07:17,870 --> 00:07:21,575 [Cindi Howson]: I could get annoying and say, yes, it is semantics, Tim. 100 00:07:25,981 --> 00:07:35,173 [Cindi Howson]: But this is where, let's say, and sometimes people conflate data literacy with technical literacy, which is wrong. 101 00:07:35,253 --> 00:07:41,481 [Cindi Howson]: But we're really talking about what does the data mean in a business context? 102 00:07:41,461 --> 00:07:44,005 [Cindi Howson]: And where does the data originate from? 103 00:07:44,025 --> 00:07:51,936 [Cindi Howson]: And if I'm talking about an order system versus an invoicing system, sometimes that's different. 104 00:07:52,958 --> 00:08:01,330 [Cindi Howson]: And so a finance person is always going to assume I am talking about when it was invoiced. 105 00:08:02,271 --> 00:08:08,380 [Cindi Howson]: A salesperson is going to come from the context of when is my commission going to get paid. 106 00:08:08,360 --> 00:08:17,375 [Cindi Howson]: And so we come to the data already thinking about the data through our own lens, our own business function. 107 00:08:17,856 --> 00:08:20,701 [Cindi Howson]: And yet they may have very, very different meanings. 108 00:08:21,322 --> 00:08:25,609 [Cindi Howson]: Even somebody that I was working with 109 00:08:25,589 --> 00:08:28,674 [Cindi Howson]: I won't name him, but it was hysterical. 110 00:08:28,834 --> 00:08:33,542 [Cindi Howson]: We're both working off the same dataset you would have thought. 111 00:08:34,323 --> 00:08:37,048 [Cindi Howson]: I'm like, why are your numbers different than mine? 112 00:08:38,570 --> 00:08:41,875 [Cindi Howson]: Here's what I thought the number was, and you're coming up with a different number. 113 00:08:42,777 --> 00:08:52,693 [Cindi Howson]: Yet, in his dataset, he only included software licensing and did not include professional services. 114 00:08:53,180 --> 00:08:55,304 [Cindi Howson]: I was like, why would you exclude that? 115 00:08:55,444 --> 00:09:00,332 [Cindi Howson]: I'm really just looking for total revenues related to this particular segment. 116 00:09:01,053 --> 00:09:03,838 [Moe Kiss]: Can I then ask, do we risk? 117 00:09:03,938 --> 00:09:08,826 [Moe Kiss]: I'm going to definitely come full circle on this because it's definitely been a topic that's on my mind a lot. 118 00:09:09,616 --> 00:09:20,858 [Moe Kiss]: One of the things that I've seen play out is this very precise, I don't want to say business domain, but this very specific interpretation of a metric by a particular area of the business. 119 00:09:21,299 --> 00:09:27,651 [Moe Kiss]: And I'm going to give the typical example in my world, which is like, let's say you have 12 different products. 120 00:09:28,092 --> 00:09:31,438 [Moe Kiss]: And so then one team is like, well, we're going to talk about 121 00:09:31,418 --> 00:09:32,860 [Moe Kiss]: video MAUs. 122 00:09:32,940 --> 00:09:35,765 [Moe Kiss]: And another team is like, we're going to talk about search MAUs. 123 00:09:35,825 --> 00:09:39,851 [Moe Kiss]: And then another area of the business is, I don't know, template MAUs. 124 00:09:39,911 --> 00:09:42,795 [Moe Kiss]: I'm making up all stuff that's relevant to my world, of course. 125 00:09:43,436 --> 00:09:51,669 [Moe Kiss]: And then we come up with this fundamental problem of if you summed every department's version of their metric, we would never end up with all MAUs. 126 00:09:52,109 --> 00:09:58,419 [Moe Kiss]: But we also end up with these very precise definitions that might work within the business context that they're in. 127 00:09:59,080 --> 00:10:00,582 [Moe Kiss]: But then like, 128 00:10:00,562 --> 00:10:10,715 [Moe Kiss]: I think the thing that I struggle from that viewpoint is sometimes I feel like we over orchestrate things for a specific domain. 129 00:10:10,956 --> 00:10:15,562 [Moe Kiss]: Then we can't roll up and think about what's the bigger picture across the whole company. 130 00:10:15,622 --> 00:10:17,905 [Moe Kiss]: When we say, MAU, what do we mean? 131 00:10:17,945 --> 00:10:21,469 [Moe Kiss]: Because they might have had interactions with lots of different products, for example. 132 00:10:22,090 --> 00:10:28,138 [Moe Kiss]: I feel like semantic layers in this are like there is an overlap here. 133 00:10:28,118 --> 00:10:29,159 [Moe Kiss]: I'm sure we'll get to. 134 00:10:29,239 --> 00:10:32,324 [Moe Kiss]: But do you see that problem playing out a lot? 135 00:10:32,344 --> 00:10:38,051 [Moe Kiss]: And is that part of why semantic layers are becoming the new hot topic of the moment? 136 00:10:38,612 --> 00:10:41,996 [Cindi Howson]: Well, let's go back and say, you just used a term, Moe. 137 00:10:42,197 --> 00:10:47,884 [Cindi Howson]: If I was a new employee at Canva or at any kind of SaaS startup, what the heck is it? 138 00:10:48,205 --> 00:10:48,585 [Moe Kiss]: Is it MAU? 139 00:10:48,605 --> 00:10:49,366 [Cindi Howson]: Or what the hell? 140 00:10:50,047 --> 00:10:50,428 [Moe Kiss]: Oh, fair. 141 00:10:50,448 --> 00:10:51,749 [Cindi Howson]: A male. 142 00:10:51,770 --> 00:10:54,413 [Cindi Howson]: A male active user. 143 00:10:55,135 --> 00:11:12,992 [Cindi Howson]: Well, and maybe even really going to split hairs here and say, well, if I only clicked on the video and so it was a one-second interaction, are you going to count me there as a user or should I have actually watched at least two minutes of the content? 144 00:11:12,972 --> 00:11:17,784 [Cindi Howson]: So we can parse these definitions a lot of different ways. 145 00:11:18,406 --> 00:11:29,293 [Cindi Howson]: But I want to come back to why did semantic layers start more than 30 years ago and why are they coming back now? 146 00:11:29,273 --> 00:11:46,590 [Cindi Howson]: 30 years ago, prior to semantic layers and really business objects patented and won in courts, the first semantic layer and Cognos at the time within Promptu had to actually pay a license fee to them. 147 00:11:47,231 --> 00:11:50,814 [Cindi Howson]: And prior to this, you had to code your own SQL. 148 00:11:51,415 --> 00:11:56,580 [Cindi Howson]: You would have to say some cryptic name, 149 00:11:57,303 --> 00:12:03,150 [Cindi Howson]: by VBAP L333 from this table, and that was terrible. 150 00:12:03,730 --> 00:12:14,382 [Cindi Howson]: The semantic layer gave report writers a way to click on business terminology to generate the SQL. 151 00:12:14,983 --> 00:12:18,226 [Cindi Howson]: That was the first purpose of the semantic layer. 152 00:12:18,947 --> 00:12:24,053 [Cindi Howson]: Now, as the industry moved to, let's say, in memory tools, 153 00:12:24,033 --> 00:12:34,739 [Cindi Howson]: With the likes of Click and Tableau, there's a whole generation of let's say 10 years, maybe 15 years, where people didn't think about this. 154 00:12:35,301 --> 00:12:40,253 [Cindi Howson]: They just loaded their data, did maybe one big SQL extract. 155 00:12:40,233 --> 00:12:43,138 [Cindi Howson]: loaded it into an in-memory file. 156 00:12:43,218 --> 00:12:45,562 [Cindi Howson]: And so they were only working with their subset of data. 157 00:12:45,602 --> 00:12:49,549 [Cindi Howson]: And so, of course, the MAUs meant what I wanted it to mean. 158 00:12:50,411 --> 00:12:56,301 [Cindi Howson]: And there was this loss of knowledge about what semantic layers are. 159 00:12:56,281 --> 00:13:04,533 [Cindi Howson]: Now, here we are in 2025, and we're all trying to build agentic AI systems. 160 00:13:05,294 --> 00:13:14,326 [Cindi Howson]: And what we're learning is that without this context or clear business definitions, we have hallucinations. 161 00:13:14,427 --> 00:13:16,530 [Cindi Howson]: We get incorrect results. 162 00:13:16,550 --> 00:13:22,558 [Cindi Howson]: So the more context you give the LLM, the more accurate 163 00:13:22,825 --> 00:13:24,667 [Cindi Howson]: your answers will be. 164 00:13:24,707 --> 00:13:34,759 [Cindi Howson]: And that is why I think, well, I think semantic layers have become more important because of agentic AI. 165 00:13:35,540 --> 00:13:45,812 [Cindi Howson]: But also, let's say before that, cloud data platforms and the whole modern data stack have given rise 166 00:13:45,792 --> 00:13:48,957 [Cindi Howson]: to, hey, I don't have to subset my data. 167 00:13:49,177 --> 00:13:54,365 [Cindi Howson]: I don't have to load just a small data set into an in-memory engine. 168 00:13:54,805 --> 00:13:59,853 [Cindi Howson]: Let me get to all of it, whether it's in Snowflake or Databricks or Google BigQuery. 169 00:14:00,173 --> 00:14:01,575 [Cindi Howson]: Let me get to all of it. 170 00:14:02,196 --> 00:14:09,707 [Cindi Howson]: And so people don't want to move the data, but they do want to trust it, no matter if they're doing agentic or not. 171 00:14:09,839 --> 00:14:19,739 [Tim Wilson]: Well, so this whole notion of context and using agentic AI as an example, is it moving down the path? 172 00:14:19,880 --> 00:14:27,315 [Tim Wilson]: Will a semantic layer help AI demand some explicit context? 173 00:14:27,435 --> 00:14:30,902 [Tim Wilson]: If I ask for, tell me how many customers 174 00:14:31,118 --> 00:14:36,329 [Tim Wilson]: we had last month, will a semantic layer start to say, that's not enough? 175 00:14:36,850 --> 00:14:51,220 [Tim Wilson]: I know, Tim, what role you're in, I can guess what your definition of a customer is, but I'm going to require that you give me more business context in order for me to find the right 176 00:14:52,550 --> 00:14:54,413 [Tim Wilson]: to pull the right information. 177 00:14:54,513 --> 00:15:07,792 [Tim Wilson]: Is that a feature of the semantic layer or is that something that's got to be built in the intermediary tool that's using the semantic layer to interface with a business user? 178 00:15:08,273 --> 00:15:09,976 [Cindi Howson]: Yeah, I follow you, Tim. 179 00:15:10,136 --> 00:15:15,604 [Cindi Howson]: And this is where I think what people want 180 00:15:16,462 --> 00:15:19,029 [Cindi Howson]: is one semantic layer to rule them all. 181 00:15:19,891 --> 00:15:23,100 [Cindi Howson]: And I just think that's a fallacy. 182 00:15:23,140 --> 00:15:25,808 [Cindi Howson]: Will I ever see that? 183 00:15:26,690 --> 00:15:31,483 [Cindi Howson]: I don't ever see that. 184 00:15:31,463 --> 00:15:40,859 [Cindi Howson]: What the industry, at least right now, is trying to get to, and I will also say this is the second attempt, maybe the third attempt in the industry. 185 00:15:42,001 --> 00:15:55,885 [Cindi Howson]: With snowflakes open semantic interchange, is it least let there be a common set of standards so that everyone can interoperate? 186 00:15:56,438 --> 00:15:59,842 [Cindi Howson]: And that already would be a huge sea change. 187 00:16:00,443 --> 00:16:04,308 [Cindi Howson]: Otherwise, everyone's building proprietary integrations. 188 00:16:05,009 --> 00:16:09,274 [Cindi Howson]: Even, I mean, I will say, working for ThoughtSpot. 189 00:16:09,334 --> 00:16:14,301 [Cindi Howson]: ThoughtSpot integrates with the Looker metrics layer in LookML. 190 00:16:15,602 --> 00:16:22,992 [Cindi Howson]: ThoughtSpot integrates with the DBT semantic layer, and that has changed different incarnations. 191 00:16:24,093 --> 00:16:25,635 [Cindi Howson]: There's a few others. 192 00:16:25,902 --> 00:16:32,950 [Cindi Howson]: that some have built integrations with KubeJS, some have built integrations with AtScale, there's others. 193 00:16:33,030 --> 00:16:34,912 [Cindi Howson]: But let's just take those. 194 00:16:35,032 --> 00:16:37,976 [Cindi Howson]: Well, those are all point solutions. 195 00:16:38,036 --> 00:16:45,064 [Cindi Howson]: We have to keep up with what is DBT's latest protocol, what is Looker's latest protocol. 196 00:16:45,805 --> 00:16:52,332 [Cindi Howson]: And it would be great if we all just say, here are the approaches that we're going to use. 197 00:16:52,312 --> 00:16:56,881 [Cindi Howson]: And so it's all common rather than point solutions. 198 00:16:57,302 --> 00:17:03,615 [Cindi Howson]: So that is the vision and the hope of snowflakes open semantic interchange. 199 00:17:04,056 --> 00:17:11,451 [Cindi Howson]: However, so this is a very long-winded answer, but we will have separate incarnations. 200 00:17:11,684 --> 00:17:20,718 [Cindi Howson]: And that, I have to say, like every customer conversation I've had about this in the last month, they're like, we only have to have one instantiation. 201 00:17:20,738 --> 00:17:30,233 [Cindi Howson]: And I'm like, no, you don't, you do need separate instantiations because every downstream tool 202 00:17:30,213 --> 00:17:34,645 [Cindi Howson]: and even backend database, they have their own limitations. 203 00:17:35,387 --> 00:17:46,517 [Cindi Howson]: So if I'm going to create something, a metric called top 10 customers, well, there's some databases that don't support a ranking function. 204 00:17:47,223 --> 00:17:51,928 [Cindi Howson]: So even like Denoto virtualization tried to do this for a while. 205 00:17:51,988 --> 00:17:53,730 [Cindi Howson]: And it's like, great. 206 00:17:53,910 --> 00:17:57,034 [Cindi Howson]: In ThoughtSpot, we have an object called top 10. 207 00:17:57,494 --> 00:18:00,057 [Cindi Howson]: Well, if I hit the Snowflake database, it's working. 208 00:18:00,798 --> 00:18:07,665 [Cindi Howson]: If on the back end it's hitting, I'm going to forget which database didn't support it, some variation of SQL server or whatever didn't support it. 209 00:18:08,046 --> 00:18:12,150 [Cindi Howson]: Well, then Denoto is like not working, not giving an answer. 210 00:18:12,130 --> 00:18:18,799 [Cindi Howson]: Or in Looker, we have a very cool visualization, my favorite visualization, a KPI chart. 211 00:18:19,268 --> 00:18:22,973 [Cindi Howson]: it's too complicated for the looker metrics layer. 212 00:18:23,614 --> 00:18:38,596 [Cindi Howson]: So you're always going to have these separate instantiations of a semantic layer because nobody is going to want to dumb down their semantic layer for the least common denominator. 213 00:18:39,176 --> 00:18:39,357 [Cindi Howson]: Okay. 214 00:18:39,397 --> 00:18:41,119 [Moe Kiss]: I've got to make sure I'm following this though. 215 00:18:41,339 --> 00:18:41,680 [Moe Kiss]: Okay. 216 00:18:41,700 --> 00:18:45,225 [Moe Kiss]: So what we're saying 217 00:18:45,526 --> 00:18:54,100 [Moe Kiss]: is that I guess the thing is like what I'm observing is that folks seem to want to be pulling their semantic layer further and further up in the chain, right? 218 00:18:54,461 --> 00:19:02,995 [Moe Kiss]: So like you want it to sit less in a downstream tool and more like internally and I obviously have a biased view. 219 00:19:02,975 --> 00:19:11,148 [Moe Kiss]: But wanting to bring things like semantic layer in-house so that you also have your options open about which way you go with whatever AI you choose to leverage. 220 00:19:11,688 --> 00:19:23,106 [Moe Kiss]: But what you're saying is it's unavoidable that we're going to end up with a semantic sandwich or cake or whatever you want to call it, where you might have to have something at one layer. 221 00:19:23,407 --> 00:19:29,556 [Moe Kiss]: And then when you go to a BI tool or some other type of tool or integration, you might end up having 222 00:19:29,536 --> 00:19:34,722 [Moe Kiss]: to have a second layer just because they have different features or attributes that you want to leverage. 223 00:19:35,003 --> 00:19:36,384 [Moe Kiss]: Am I hearing that correctly? 224 00:19:36,605 --> 00:19:37,506 [Cindi Howson]: Yes. 225 00:19:37,686 --> 00:19:45,195 [Cindi Howson]: And if by downstream and upstream you mean the database, people want it closer to the database because that's where the data lives. 226 00:19:46,597 --> 00:19:56,809 [Cindi Howson]: But then as you get closer to the business decisions, you're going to have derivations and metrics 227 00:19:57,093 --> 00:20:00,079 [Cindi Howson]: and context that may not exist in the database. 228 00:20:00,861 --> 00:20:06,112 [Cindi Howson]: And I would also say, we also have to think about how these things get defined. 229 00:20:07,215 --> 00:20:13,107 [Cindi Howson]: So working with one team, they're like, okay, we're going to build everything out in the database. 230 00:20:13,909 --> 00:20:14,550 [Cindi Howson]: I'm like, great. 231 00:20:14,711 --> 00:20:17,717 [Cindi Howson]: So your DBA is going to do all this. 232 00:20:18,035 --> 00:20:22,963 [Cindi Howson]: Or here I have a really strong SME. 233 00:20:23,003 --> 00:20:39,929 [Cindi Howson]: And if we bring data mesh operating principles and domain ownership, so I have this great marketing person and they know the differences between a video, MAU, or a web click MAU. 234 00:20:39,970 --> 00:20:44,717 [Cindi Howson]: And I'm gonna want them to add a little more context to it. 235 00:20:44,697 --> 00:20:46,980 [Cindi Howson]: So I'm going to want an easier interface. 236 00:20:47,181 --> 00:20:47,741 [Cindi Howson]: And guess what? 237 00:20:47,781 --> 00:20:52,588 [Cindi Howson]: That interface does not exist in the one that was designed for the DBA. 238 00:20:53,389 --> 00:20:59,278 [Moe Kiss]: OK, but I don't want to take things in a totally different direction, but I'm like dying a little bit. 239 00:21:00,960 --> 00:21:02,623 [Cindi Howson]: I feel like a dymo. 240 00:21:03,424 --> 00:21:07,730 [Moe Kiss]: I opened another can of worms, I can tell. 241 00:21:09,733 --> 00:21:14,620 [Moe Kiss]: The thing I'm really struggling with with this whole discussion about semantic layers 242 00:21:14,954 --> 00:21:20,986 [Moe Kiss]: It doesn't feel new, and I feel like what you've written about it makes that very clear. 243 00:21:21,928 --> 00:21:28,021 [Moe Kiss]: But part of me is also really grappling with, is it actually the fact that it's not net new? 244 00:21:28,041 --> 00:21:32,710 [Moe Kiss]: Is it the fact that the way we want to use agentic AI on top of our data? 245 00:21:32,690 --> 00:21:40,340 [Moe Kiss]: Or is it the fact that we have gone towards this data mesh approach with less, I don't know if structured is the right word. 246 00:21:40,520 --> 00:21:46,228 [Moe Kiss]: Cindi, you can definitely insert better terminology because you are the queen of exceptional terminology. 247 00:21:46,288 --> 00:21:50,514 [Moe Kiss]: But we used to have such structured datasets. 248 00:21:50,534 --> 00:21:52,677 [Moe Kiss]: We had store schemas, they had context. 249 00:21:53,478 --> 00:21:59,886 [Moe Kiss]: Is part of this just our own doing because we wanted to move faster and have less structure in our data? 250 00:22:00,046 --> 00:22:02,089 [Moe Kiss]: And so this is just the consequence. 251 00:22:04,246 --> 00:22:06,608 [Cindi Howson]: So that was a two-part question. 252 00:22:06,728 --> 00:22:07,789 [Cindi Howson]: So is it new? 253 00:22:07,849 --> 00:22:10,091 [Cindi Howson]: Is the semantic layer new? 254 00:22:10,632 --> 00:22:11,213 [Cindi Howson]: It's not new. 255 00:22:11,973 --> 00:22:14,876 [Cindi Howson]: It has gotten more robust over time. 256 00:22:16,277 --> 00:22:20,181 [Cindi Howson]: And not all semantic layers are created equal. 257 00:22:20,922 --> 00:22:27,127 [Cindi Howson]: So I can show you one semantic layer, and it only supports a single star schema. 258 00:22:28,108 --> 00:22:31,992 [Cindi Howson]: Or maybe even worse, it only supports one big table. 259 00:22:32,731 --> 00:22:42,694 [Cindi Howson]: I can show you another semantic layer and it supports multiple fact tables, different design approaches, star schema, snowflake schema. 260 00:22:43,295 --> 00:22:48,327 [Cindi Howson]: It even includes capabilities 261 00:22:48,307 --> 00:22:57,839 [Cindi Howson]: for aggregate table navigation or query compilation so that the most efficient query path is taken. 262 00:22:58,319 --> 00:23:01,163 [Cindi Howson]: So not all semantic layers are created equal. 263 00:23:01,723 --> 00:23:04,327 [Cindi Howson]: And I do think that has changed over time. 264 00:23:04,367 --> 00:23:09,673 [Cindi Howson]: And for sure, the openness has changed over time. 265 00:23:10,312 --> 00:23:20,026 [Cindi Howson]: So if I go back to the original query tools, whether again, business objects, Cognos, whatever, those were largely closed. 266 00:23:20,186 --> 00:23:26,135 [Cindi Howson]: Some boutique consultancies had open APIs to access them. 267 00:23:26,836 --> 00:23:31,343 [Cindi Howson]: OBI EE, their model was open and nobody used it. 268 00:23:31,443 --> 00:23:36,631 [Cindi Howson]: You could expose it as an ODBC connector to other BI tools, but nobody used it. 269 00:23:36,691 --> 00:23:38,293 [Cindi Howson]: Performance was not good. 270 00:23:38,273 --> 00:23:52,290 [Cindi Howson]: So what we have now is definitely more openness, but I do believe it is the agentic part of why we're demanding, why we need them more. 271 00:23:52,491 --> 00:23:54,533 [Cindi Howson]: It'll just make AI better. 272 00:23:54,573 --> 00:24:00,501 [Cindi Howson]: The second part of your question was then, are we decentralizing these things? 273 00:24:00,901 --> 00:24:03,584 [Cindi Howson]: And yes, I think that's part of it too. 274 00:24:03,604 --> 00:24:06,508 [Tim Wilson]: This makes me feel like there are 275 00:24:06,860 --> 00:24:10,185 [Tim Wilson]: This is either going to be just so obvious that it's dumb. 276 00:24:10,966 --> 00:24:26,067 [Tim Wilson]: Are there people out there who, if some generic person came in and looked at it, they would say, you have built a wonderful semantic layer, and the people who built it would say, I built something that functioned for what I need. 277 00:24:26,047 --> 00:24:27,830 [Tim Wilson]: I didn't know that's what it was called. 278 00:24:28,771 --> 00:24:35,221 [Tim Wilson]: On the flip side, that has me thinking that semantic layer, it sounds cool. 279 00:24:35,682 --> 00:24:37,645 [Tim Wilson]: It gets treated as this binary. 280 00:24:37,685 --> 00:24:40,549 [Tim Wilson]: If you have one, things are good. 281 00:24:40,790 --> 00:24:42,472 [Tim Wilson]: If you don't have one, they're bad. 282 00:24:42,633 --> 00:24:44,916 [Tim Wilson]: It sounds like what you're saying is, 283 00:24:44,896 --> 00:24:52,047 [Tim Wilson]: you could try to boil the ocean with one grand semantic layer, and it would probably be bad. 284 00:24:52,087 --> 00:24:56,253 [Tim Wilson]: We treat it as though there's this label, and if you have it, then things are fixed. 285 00:24:56,354 --> 00:25:03,164 [Tim Wilson]: But there's always the gradations of whether you do it well, well-architected and appropriately. 286 00:25:03,805 --> 00:25:07,090 [Tim Wilson]: That probably happens with everything that gets a fancy new label. 287 00:25:07,998 --> 00:25:08,499 [Cindi Howson]: Yeah. 288 00:25:08,739 --> 00:25:09,180 [Cindi Howson]: Yeah. 289 00:25:09,300 --> 00:25:15,268 [Cindi Howson]: So I don't, I don't know, Tim, like, do I want one mega semantic layer? 290 00:25:15,529 --> 00:25:16,530 [Cindi Howson]: Oh, please not. 291 00:25:16,630 --> 00:25:34,596 [Cindi Howson]: So because it becomes overwhelming to maintain and, and it becomes now maybe, maybe if I'm just using natural language, um, to ask questions, I don't care what it's hitting on the back end, but I, I, 292 00:25:35,369 --> 00:25:39,178 [Cindi Howson]: I would be skeptical that that would work. 293 00:25:40,340 --> 00:25:49,922 [Cindi Howson]: There is a belief that in the industry we're going to go towards verticalization of some of these semantic layers. 294 00:25:50,002 --> 00:25:53,530 [Cindi Howson]: So there will be 295 00:25:53,763 --> 00:25:57,428 [Cindi Howson]: And maybe this is, I kind of bristle it. 296 00:25:57,528 --> 00:26:00,772 [Cindi Howson]: We throw new terms out there, ontologies. 297 00:26:01,813 --> 00:26:03,615 [Cindi Howson]: Well, can we just talk about domains? 298 00:26:03,716 --> 00:26:06,579 [Cindi Howson]: That makes more sense to me, and that aligns with the data mesh. 299 00:26:09,062 --> 00:26:14,950 [Cindi Howson]: But could we have an insurance industry semantic layer? 300 00:26:15,711 --> 00:26:22,960 [Cindi Howson]: Could we have a marketing web analytics semantic layer? 301 00:26:23,446 --> 00:26:24,408 [Cindi Howson]: I think we could. 302 00:26:24,428 --> 00:26:25,651 [Cindi Howson]: I think we could. 303 00:26:26,152 --> 00:26:28,055 [Cindi Howson]: We would get to common metrics. 304 00:26:30,060 --> 00:26:41,944 [Cindi Howson]: The physical pointers ultimately back to which table is it hitting might change a little bit, but I think that business representation, we could possibly get to that. 305 00:26:43,493 --> 00:26:55,186 [Moe Kiss]: Okay, one thought, just asking for a friend, of course, that has been on my mind is we can approach this from like a business domain perspective, just like the examples you gave, right? 306 00:26:55,226 --> 00:27:05,697 [Moe Kiss]: So like, you might have one that's more like marketing and acquisition, more the one that's like, I don't know, finance or whatever, and so whatever else business domain. 307 00:27:05,737 --> 00:27:11,003 [Moe Kiss]: And the thing that I kind of keep wrestling with though is, 308 00:27:10,983 --> 00:27:25,275 [Moe Kiss]: Are we just doing this again where we're overlaying our thoughts about like what a domain is versus the business user and how they want to interact with data? 309 00:27:25,656 --> 00:27:30,220 [Moe Kiss]: And what I mean by that is like, if I'm a business user, what's my business question? 310 00:27:30,260 --> 00:27:31,761 [Moe Kiss]: What are the questions that I want to ask? 311 00:27:31,781 --> 00:27:40,028 [Moe Kiss]: And let's say the theme might be I want to ask a question about our users or I want to ask a question about 312 00:27:40,008 --> 00:27:40,849 [Moe Kiss]: I don't know. 313 00:27:40,869 --> 00:27:43,152 [Moe Kiss]: Now I'm going to like struggle to think of a comparative example. 314 00:27:43,793 --> 00:27:48,439 [Moe Kiss]: I might want to ask something about, someone help me with an answer. 315 00:27:48,619 --> 00:27:49,640 [Moe Kiss]: A marketing channel. 316 00:27:50,401 --> 00:27:51,403 [Moe Kiss]: A marketing channel. 317 00:27:51,603 --> 00:27:51,923 [Moe Kiss]: Sure. 318 00:27:51,943 --> 00:27:57,991 [Moe Kiss]: A marketing channel or, yeah, like I want to understand something about how experiments have done. 319 00:27:58,011 --> 00:28:01,155 [Moe Kiss]: Like, are we doing a thing where we're trying to make 320 00:28:01,135 --> 00:28:13,898 [Moe Kiss]: semantic layer is representative of business domains that make maybe business sense, but don't reflect the way that users and our business users want to interact with data when they have questions. 321 00:28:15,261 --> 00:28:23,295 [Cindi Howson]: Well, so to me, if you build a semantic layer that doesn't work that way, what is the point? 322 00:28:23,410 --> 00:28:25,053 [Cindi Howson]: Like, go home. 323 00:28:27,256 --> 00:28:31,703 [Cindi Howson]: Because you know SQL, you want to code your SQL, you don't need a semantic layer. 324 00:28:31,803 --> 00:28:42,219 [Cindi Howson]: You might want it for some reusability, but the semantic layer gives the business user the ability to ask the questions without knowing SQL. 325 00:28:42,880 --> 00:28:51,173 [Cindi Howson]: And then it gives the LLM more context to generate better SQL. 326 00:28:51,153 --> 00:29:00,306 [Cindi Howson]: So all these companies that have tried to do text to SQL without a semantic layer, they're largely failing. 327 00:29:00,326 --> 00:29:01,107 [Cindi Howson]: And guess what? 328 00:29:01,267 --> 00:29:03,811 [Cindi Howson]: They're adding semantic layers so that they work. 329 00:29:04,492 --> 00:29:07,576 [Cindi Howson]: So semantic layers bring reusability. 330 00:29:07,616 --> 00:29:09,919 [Cindi Howson]: That was the original purpose. 331 00:29:10,520 --> 00:29:15,507 [Cindi Howson]: And then it is a business-friendly interface. 332 00:29:15,487 --> 00:29:20,656 [Cindi Howson]: And now in agentic AI, it's the context for the LLMs to ensure accuracy. 333 00:29:21,398 --> 00:29:33,559 [Cindi Howson]: So if you're going to give me a semantic layer that is just a bunch of cryptic names, technical names, and it's not giving it to me in a way that the business sees it, it's a waste of time. 334 00:29:33,840 --> 00:29:36,685 [Cindi Howson]: It's a poorly architected semantic layer. 335 00:29:37,660 --> 00:29:50,842 [Moe Kiss]: So hypothetically, if you just took all your YAML descriptions, that probably wouldn't be good enough because it's been written by a data scientist in their domain for their own specific domain for use by someone who deeply understands their area. 336 00:29:51,283 --> 00:29:57,794 [Cindi Howson]: Well, if they deeply understand their area, there might be a lot of useful context in there. 337 00:29:58,915 --> 00:30:02,301 [Cindi Howson]: But if it's a lot of code, 338 00:30:03,040 --> 00:30:06,669 [Cindi Howson]: and techno babble, then I think it's going to be less useful. 339 00:30:06,989 --> 00:30:10,157 [Tim Wilson]: Back on the, I may blend two things together. 340 00:30:10,217 --> 00:30:15,550 [Tim Wilson]: The referencing snowflakes open semantic, what is it? 341 00:30:15,590 --> 00:30:19,800 [Tim Wilson]: Open semantic interchange does feel like 342 00:30:21,198 --> 00:30:27,532 [Tim Wilson]: That brings to mind the XKCD cartoon about people complaining that there are 13 different standards. 343 00:30:27,572 --> 00:30:32,663 [Tim Wilson]: We need one standard, and then the next panel is, well, now we have 14 competing standards. 344 00:30:34,327 --> 00:30:37,333 [Tim Wilson]: There does need to be a first 345 00:30:37,600 --> 00:30:48,721 [Tim Wilson]: mover or a dominant, is there a race to say, obviously, Snowflake wants to be the owner, the driver of that? 346 00:30:49,042 --> 00:30:52,268 [Tim Wilson]: I guess the same thing when you talk about verticalization, 347 00:30:52,906 --> 00:31:00,684 [Tim Wilson]: say something like digital analytics, and you're like, let's just have one common marketing digital analytics. 348 00:31:00,824 --> 00:31:11,910 [Tim Wilson]: Well, now you're going to have the players in that are all going to say, yes, the way that we think about that data is the way that the industry. 349 00:31:12,700 --> 00:31:27,735 [Tim Wilson]: the effort to try to have some sense of standards not lead to self-interested competition to sort of pull the market towards whoever's on point for defining the standard. 350 00:31:27,815 --> 00:31:29,998 [Tim Wilson]: Or maybe my third example would be the W3C. 351 00:31:30,318 --> 00:31:34,182 [Tim Wilson]: I mean, we go back 30 years trying to define what HTML is supposed to do. 352 00:31:34,802 --> 00:31:40,328 [Tim Wilson]: And Microsoft doesn't even conform to the W3C standards because 353 00:31:40,409 --> 00:31:45,357 [Cindi Howson]: Tim, you just answered that question. 354 00:31:45,557 --> 00:31:51,366 [Cindi Howson]: Microsoft would love us to revert to MDX instead of SQL for the most part. 355 00:31:52,408 --> 00:31:53,730 [Cindi Howson]: But it is true. 356 00:31:53,750 --> 00:31:58,317 [Cindi Howson]: So look at who is not part of that. 357 00:31:58,297 --> 00:31:58,858 [Cindi Howson]: effort. 358 00:31:59,559 --> 00:32:02,564 [Cindi Howson]: Was Databricks invited to the party? 359 00:32:02,604 --> 00:32:05,228 [Cindi Howson]: Was Google BigQuery invited to the party? 360 00:32:05,288 --> 00:32:07,091 [Cindi Howson]: Will they invite themselves? 361 00:32:07,312 --> 00:32:08,634 [Cindi Howson]: Will they become part of it? 362 00:32:08,654 --> 00:32:20,473 [Cindi Howson]: Standards get adopted based on who leads it, but then also who uses it and who asks for it. 363 00:32:20,453 --> 00:32:30,187 [Cindi Howson]: So that's where when I look at how we prioritize our product strategies, we are very much listening to the customers. 364 00:32:31,208 --> 00:32:35,354 [Cindi Howson]: And sometimes we've gone down rabbit holes. 365 00:32:36,195 --> 00:32:38,518 [Cindi Howson]: And I'm like, why did we build that integration? 366 00:32:38,558 --> 00:32:42,023 [Cindi Howson]: So I won't say which integrations to me were a waste of time. 367 00:32:42,577 --> 00:32:44,959 [Cindi Howson]: But some of them, I'm like, why did we do that? 368 00:32:46,200 --> 00:32:50,605 [Cindi Howson]: Because we were trying to, we thought something would have legs to it. 369 00:32:51,305 --> 00:32:55,970 [Cindi Howson]: We listened to the customer and it never really took off. 370 00:32:56,850 --> 00:32:59,493 [Cindi Howson]: And then some will change strategies. 371 00:33:00,234 --> 00:33:10,103 [Cindi Howson]: So, you know, we thought DBT's initial effort would take off and instead then, you know, they're on version two and now 372 00:33:10,538 --> 00:33:15,248 [Cindi Howson]: So Snowflake, hugely influential in the industry. 373 00:33:15,909 --> 00:33:25,770 [Cindi Howson]: We're very proud to be part of the committee defining these standards, but we have to see how broadly adopted they are. 374 00:33:26,692 --> 00:33:28,075 [Cindi Howson]: The market will decide. 375 00:33:29,236 --> 00:33:39,689 [Michael Helbling]: And certainly right now, AI is kind of a forcing function for the industry where maybe that hasn't been or there hasn't been an imperative like that for a lot of companies. 376 00:33:40,090 --> 00:33:40,871 [Michael Helbling]: Does that seem fair? 377 00:33:41,612 --> 00:33:42,573 [Cindi Howson]: I think that seems fair. 378 00:33:42,833 --> 00:33:43,895 [Cindi Howson]: Yeah. 379 00:33:43,935 --> 00:33:57,572 [Cindi Howson]: And there's more willingness to be open and to focus on where you add the value in this data to insight to action change. 380 00:33:59,088 --> 00:34:02,656 [Moe Kiss]: That actually triggers an interesting thought. 381 00:34:03,718 --> 00:34:08,108 [Moe Kiss]: One of the things that I've observed is this push for semantic layers. 382 00:34:08,529 --> 00:34:11,315 [Moe Kiss]: I feel like it's come out of left field. 383 00:34:11,355 --> 00:34:12,457 [Moe Kiss]: I don't know if that's fair or not. 384 00:34:12,517 --> 00:34:15,544 [Moe Kiss]: It just seems to have swirled very quickly. 385 00:34:15,845 --> 00:34:22,452 [Moe Kiss]: And the products maybe aren't at a state of maturity where they need to be for what people. 386 00:34:22,833 --> 00:34:29,160 [Moe Kiss]: I almost feel like a lot of companies are building as they're gathering requirements as customers are trying to build out with them. 387 00:34:29,180 --> 00:34:32,964 [Moe Kiss]: Do you think that's a fair representation? 388 00:34:33,144 --> 00:34:41,153 [Moe Kiss]: Has this happened before with a particular tool that's had to develop very quickly because of the pressure? 389 00:34:41,193 --> 00:34:43,776 [Moe Kiss]: And I feel like AI is the pressure of like, 390 00:34:43,756 --> 00:34:47,761 [Moe Kiss]: Everyone suddenly needs these semantic layers to make AI quote unquote work. 391 00:34:48,862 --> 00:35:00,076 [Moe Kiss]: Do you feel like that's happened with the product development before or is this like a net new thing that data companies are trying to deal with where they're trying to build at pace while customers are wanting to already leverage and use it? 392 00:35:00,497 --> 00:35:01,037 [Cindi Howson]: Yeah. 393 00:35:01,057 --> 00:35:07,165 [Cindi Howson]: So I don't want to sound like a commercial and you can edit this out afterwards. 394 00:35:08,005 --> 00:35:10,490 [Cindi Howson]: All semantic layers are not created equal. 395 00:35:10,750 --> 00:35:24,457 [Cindi Howson]: Now, fortunately, because ThoughtSpot, whether it was purposeful or luck, ThoughtSpot always generated SQL on the backend. 396 00:35:25,098 --> 00:35:29,727 [Cindi Howson]: So the semantic layer was always super robust. 397 00:35:30,314 --> 00:35:33,460 [Cindi Howson]: So, did we get lucky or was it intentional? 398 00:35:33,500 --> 00:35:39,432 [Cindi Howson]: And the cloud data warehouse and agentic AI has just helped that. 399 00:35:40,274 --> 00:35:46,787 [Cindi Howson]: Others have only just started to embark on natural language processing. 400 00:35:46,767 --> 00:35:48,050 [Cindi Howson]: and a gentek AI. 401 00:35:48,591 --> 00:35:52,039 [Cindi Howson]: And they tried to do it without a semantic layer. 402 00:35:52,079 --> 00:35:55,526 [Cindi Howson]: And that's why now they're dabbling in it. 403 00:35:55,887 --> 00:35:59,094 [Cindi Howson]: And they're like, oh, it takes a lot to build this. 404 00:35:59,896 --> 00:36:03,223 [Cindi Howson]: And some of them, they're starting out simple. 405 00:36:03,203 --> 00:36:04,626 [Cindi Howson]: You know one big table. 406 00:36:05,067 --> 00:36:08,273 [Cindi Howson]: That's all they can handle and code based. 407 00:36:08,554 --> 00:36:19,035 [Cindi Howson]: So and I think about I think about a blog actually that our co-founder Amit Prakash wrote about 408 00:36:19,015 --> 00:36:26,604 [Cindi Howson]: four years ago, I think it was, and it was the metrics layer, which is just a subset of the semantic layer. 409 00:36:26,944 --> 00:36:29,147 [Cindi Howson]: The metrics layer has some growing up to do. 410 00:36:29,187 --> 00:36:42,982 [Cindi Howson]: And even as a former Gartner research vice president, I have to give credit to Gartner, they still say that the time to maturity for metrics layers is five to 10 years. 411 00:36:43,463 --> 00:36:44,324 [Cindi Howson]: That's a long time. 412 00:36:45,525 --> 00:36:48,769 [Tim Wilson]: Yeah. 413 00:36:49,390 --> 00:37:05,157 [Tim Wilson]: But so how unfair is this parallel to point to master data management as something that I remember having a moment that was, oh, things are getting fragmented. 414 00:37:05,197 --> 00:37:08,463 [Tim Wilson]: We need to just do an MDM initiative. 415 00:37:08,683 --> 00:37:13,812 [Tim Wilson]: And I guess to my earlier point, it was kind of a binary, if we do MDM, 416 00:37:13,792 --> 00:37:16,600 [Tim Wilson]: all these problems get solved. 417 00:37:16,861 --> 00:37:28,273 [Tim Wilson]: The companies that were already built where they had MDM under the hood anyway because they'd architected their setup well, 418 00:37:28,625 --> 00:37:39,825 [Tim Wilson]: could do MDM, the ones that had built kind of a hot mess and were then trying to just apply a whole bunch of duct tape and bailing wire to do MDM like never really got there. 419 00:37:40,065 --> 00:37:45,435 [Tim Wilson]: Is that a fair parallel or am I too much of a stretch? 420 00:37:45,567 --> 00:37:49,133 [Cindi Howson]: Yeah, well, so is it a fair parallel? 421 00:37:49,153 --> 00:37:53,120 [Cindi Howson]: I would just say it's valid. 422 00:37:54,242 --> 00:37:54,843 [Cindi Howson]: It's valid. 423 00:37:55,003 --> 00:38:01,534 [Cindi Howson]: And I remember, so the first eight years of my life in this industry were at Dow Chemical. 424 00:38:01,514 --> 00:38:05,139 [Cindi Howson]: And we had a master data management system called INCA. 425 00:38:05,339 --> 00:38:08,224 [Cindi Howson]: I don't even remember what it stands for. 426 00:38:08,304 --> 00:38:09,105 [Cindi Howson]: It was homegrown. 427 00:38:09,986 --> 00:38:16,255 [Cindi Howson]: And then I worked at Deloitte, and I was like, wait, you don't have clean product codes? 428 00:38:16,535 --> 00:38:18,759 [Cindi Howson]: You don't have a single product hierarchy? 429 00:38:18,859 --> 00:38:20,681 [Cindi Howson]: You don't have clean customer data? 430 00:38:20,701 --> 00:38:25,128 [Cindi Howson]: It was a foreign concept to me that not everyone had clean master data. 431 00:38:25,949 --> 00:38:30,255 [Cindi Howson]: So I would just say that 432 00:38:30,235 --> 00:38:35,462 [Cindi Howson]: And Moe, you asked this earlier, so I wanted to come back to this point. 433 00:38:35,502 --> 00:38:39,567 [Cindi Howson]: Semantic layers right now are mainly for the structured data. 434 00:38:40,488 --> 00:38:48,578 [Cindi Howson]: But I think there's a time in the not too distant future that it will encompass also the semi-structured data. 435 00:38:48,618 --> 00:38:51,602 [Cindi Howson]: And I would say this data 436 00:38:51,582 --> 00:39:06,859 [Cindi Howson]: is a hot mess, frankly, because we've never applied all of these data governance and data management disciplines that we have been applying to the structured data. 437 00:39:07,360 --> 00:39:18,973 [Cindi Howson]: So I think organizations that had the organizations that are best positioned for the agentic AI era got to cloud. 438 00:39:18,953 --> 00:39:27,881 [Cindi Howson]: had clean data, had good master data, and then of course the culture and the people change management. 439 00:39:28,802 --> 00:39:32,865 [Cindi Howson]: If they already did that, they're already, they have such a leg up. 440 00:39:33,145 --> 00:39:43,875 [Cindi Howson]: Now we're throwing generative AI, agentic AI, semi-structured data, a lot more data that we couldn't get to before. 441 00:39:44,415 --> 00:39:47,458 [Cindi Howson]: And yeah, it's not that easy. 442 00:39:48,316 --> 00:39:51,739 [Michael Helbling]: It's nice to know we'll continue to have jobs going into the future though. 443 00:39:51,879 --> 00:39:52,280 [Cindi Howson]: Yeah. 444 00:39:52,320 --> 00:39:55,182 [Cindi Howson]: That's why I'm like, what's everyone worried about not having jobs? 445 00:39:55,242 --> 00:39:58,926 [Cindi Howson]: They, they just will be different jobs, different jobs. 446 00:39:59,686 --> 00:40:00,887 [Cindi Howson]: Yeah. 447 00:40:00,907 --> 00:40:01,608 [Tim Wilson]: I got one more. 448 00:40:01,788 --> 00:40:18,323 [Tim Wilson]: This is like, this could be a complete non sequitur, but it, but I feel like Cindi could tee off on this and I want more color because it was, it was from the post that you'd written where the quote was, our industry has also now raised the generation of data analysts who never learned proper data modeling. 449 00:40:19,636 --> 00:40:23,166 [Tim Wilson]: And I kind of wanted you to elaborate on that. 450 00:40:23,753 --> 00:40:27,677 [Cindi Howson]: Well, I'm going to say first, tell me if you disagree or not. 451 00:40:28,438 --> 00:40:30,479 [Cindi Howson]: But tell me if you disagree or not. 452 00:40:30,520 --> 00:40:39,588 [Cindi Howson]: But I follow the work of people like Joe Rice and Sonny Rivera, a snowflake superhero. 453 00:40:39,608 --> 00:40:51,820 [Cindi Howson]: And yeah, it's, and I work with a lot of, let's say visualization experts who are just used to 454 00:40:52,424 --> 00:40:57,309 [Cindi Howson]: one-offs, let me load the data and let me visualize it. 455 00:40:57,750 --> 00:41:01,274 [Cindi Howson]: They never really learned proper data modeling techniques. 456 00:41:02,355 --> 00:41:18,573 [Tim Wilson]: I guess my question is that a way of saying that there are analysts who aren't really actually thinking about the structure of the data and the ramifications for how the data fits together. 457 00:41:18,613 --> 00:41:21,216 [Tim Wilson]: They're just trying to get to 458 00:41:21,449 --> 00:41:22,331 [Tim Wilson]: and output. 459 00:41:22,632 --> 00:41:24,617 [Tim Wilson]: I don't know if I agree or disagree. 460 00:41:24,637 --> 00:41:29,790 [Tim Wilson]: I probably agree because I'm just generally negative and that's like a negative statement. 461 00:41:31,715 --> 00:41:33,660 [Cindi Howson]: Let's not take it negative. 462 00:41:34,101 --> 00:41:34,783 [Cindi Howson]: Yeah. 463 00:41:34,983 --> 00:41:37,467 [Cindi Howson]: Let's challenge these people. 464 00:41:37,507 --> 00:41:39,070 [Cindi Howson]: Let, to me, empower them. 465 00:41:39,510 --> 00:41:40,232 [Cindi Howson]: So you know what? 466 00:41:40,412 --> 00:41:47,664 [Cindi Howson]: You're great at visualization and you're great at building dashboards. 467 00:41:48,405 --> 00:41:57,700 [Cindi Howson]: But if you want to continue to have a career on this space, in this space, I want you to learn some data modeling fundamentals. 468 00:41:57,680 --> 00:42:02,251 [Cindi Howson]: And I don't care which methodology you follow. 469 00:42:02,712 --> 00:42:04,175 [Cindi Howson]: Learn some data modeling. 470 00:42:04,256 --> 00:42:05,619 [Cindi Howson]: That's on the technical side. 471 00:42:06,300 --> 00:42:09,107 [Cindi Howson]: But also, we talk about data literacy. 472 00:42:09,147 --> 00:42:11,693 [Cindi Howson]: We also need to bring in business literacy. 473 00:42:11,713 --> 00:42:15,823 [Cindi Howson]: And so to me, it's not just about 474 00:42:16,850 --> 00:42:18,292 [Cindi Howson]: Where is the data coming from? 475 00:42:18,312 --> 00:42:21,176 [Cindi Howson]: It is also, how is it used? 476 00:42:21,937 --> 00:42:25,942 [Cindi Howson]: And that there really might be two different definitions. 477 00:42:26,223 --> 00:42:34,173 [Cindi Howson]: I mean, when I talk to somebody in airlines, I don't even, I'm like, oh wow, I think of on time performance. 478 00:42:35,755 --> 00:42:37,778 [Cindi Howson]: Did it leave the gate on time? 479 00:42:38,247 --> 00:42:39,769 [Cindi Howson]: or did it arrive on time? 480 00:42:40,630 --> 00:42:42,713 [Cindi Howson]: Which one is really more important to you? 481 00:42:43,214 --> 00:42:52,588 [Cindi Howson]: And by the way, when you're crossing international date lines, that it gets a little more complicated still. 482 00:42:53,269 --> 00:42:58,676 [Cindi Howson]: So I would say I want these analysts to learn both the skills. 483 00:42:59,578 --> 00:43:01,200 [Moe Kiss]: I have one last question. 484 00:43:01,467 --> 00:43:08,660 [Moe Kiss]: Just hypothetically, if you were implementing a semantic layer, what would be the top three things you'd want to avoid? 485 00:43:09,341 --> 00:43:11,425 [Cindi Howson]: The top three things? 486 00:43:11,665 --> 00:43:20,141 [Cindi Howson]: Okay, well, I'm going to start with the first thing I would want to do, so I'd have to flip it, avoid it, or what do I want to do? 487 00:43:20,301 --> 00:43:25,450 [Moe Kiss]: Or you can do the top three things to make it successful, either way, whichever your brain works. 488 00:43:25,430 --> 00:43:38,530 [Cindi Howson]: You want to avoid bringing in absolutely everything in the physical storage and exposing that to mere mortals because that'll be overwhelming. 489 00:43:38,930 --> 00:43:43,337 [Cindi Howson]: So I always start with who is going to use this. 490 00:43:44,043 --> 00:43:48,827 [Cindi Howson]: And what are the top questions they're going to want to be able to ask of it? 491 00:43:49,348 --> 00:43:56,133 [Cindi Howson]: Not because I'm going to hard code that, but that I'm going to get an idea of the context in which they're operating. 492 00:43:56,654 --> 00:43:57,334 [Michael Helbling]: Cindi, wow. 493 00:43:58,676 --> 00:44:00,337 [Michael Helbling]: So cool to talk to you. 494 00:44:00,397 --> 00:44:01,258 [Michael Helbling]: Thank you so much. 495 00:44:01,518 --> 00:44:04,401 [Michael Helbling]: This has been really, really good. 496 00:44:04,461 --> 00:44:07,363 [Michael Helbling]: I've got a ton of notes that I've been writing down. 497 00:44:07,463 --> 00:44:13,088 [Michael Helbling]: So I know that our listeners probably also get gaining a lot from this episode. 498 00:44:13,068 --> 00:44:20,881 [Michael Helbling]: All right, well, let me switch gears really quickly because I need to talk about a quick break with our friend Michael Kaminsky from ReCast. 499 00:44:21,382 --> 00:44:27,512 [Michael Helbling]: The media makes marketing and GeoLift platform helping teams forecast accurately and make better decisions. 500 00:44:28,073 --> 00:44:34,604 [Michael Helbling]: Michael's been sharing with us bite-sized marketing science lessons over the last couple of months, and they'll help you measure smarter. 501 00:44:34,944 --> 00:44:36,527 [Michael Helbling]: Okay, over to you, Michael. 502 00:44:38,127 --> 00:44:44,320 [Michael Kaminsky (Recast)]: Multicollinearity strikes fear into the hearts of many analysts and executives, but it's also one of the most commonly misunderstood concepts in analytics. 503 00:44:44,741 --> 00:44:55,022 [Michael Kaminsky (Recast)]: Some amount of correlation across variables is expected in most real-world analyses, so it's critical to understand what multicollinearity is, why it causes issues, and whether or not it's a problem for your particular analysis. 504 00:44:55,002 --> 00:44:58,489 [Michael Kaminsky (Recast)]: Multicollinearity means that two of your variables share some of the same signal. 505 00:44:58,509 --> 00:45:02,937 [Michael Kaminsky (Recast)]: This causes problems for a regression model, which will not know how to allocate credit between the two variables. 506 00:45:03,218 --> 00:45:06,143 [Michael Kaminsky (Recast)]: This can cause challenges when it comes to interpreting the results of your regression. 507 00:45:06,424 --> 00:45:13,197 [Michael Kaminsky (Recast)]: Let's imagine you're modeling the drivers of home prices in some geography, and you want to include home square footage and the number of bedrooms as predictors. 508 00:45:13,177 --> 00:45:16,623 [Michael Kaminsky (Recast)]: These two variables share some amount of signal, namely about the bigness of the house. 509 00:45:16,723 --> 00:45:25,699 [Michael Kaminsky (Recast)]: If you include both variables in a simple linear regression, you'll often get strange results, where one of the two variables is highly impactful with a large coefficient, and the other might be very small or even negative. 510 00:45:26,080 --> 00:45:30,708 [Michael Kaminsky (Recast)]: Slightly different data sets might even cause the variables to flip, which one is positive and which one is negative. 511 00:45:30,888 --> 00:45:36,157 [Michael Kaminsky (Recast)]: This happens because the model doesn't know how to apportion credit for bigness, which is present in both variables. 512 00:45:36,358 --> 00:45:37,800 [Michael Kaminsky (Recast)]: So you get these strange results. 513 00:45:37,780 --> 00:45:45,495 [Michael Kaminsky (Recast)]: So the core problem of multicollinearity is that when there's shared information across variables, a simple regression won't know how to apportion credit between them. 514 00:45:45,655 --> 00:45:51,486 [Michael Kaminsky (Recast)]: This means that you either need to accept more uncertainty in results, or try to change the variables you're using to account for the shared information. 515 00:45:52,327 --> 00:45:52,908 [Michael Helbling]: Thanks, Michael. 516 00:45:53,289 --> 00:45:59,639 [Michael Helbling]: And for those who haven't heard, our friends at ReCast just launched their new incrementality testing platform, GeoLift, by ReCast. 517 00:45:59,659 --> 00:46:09,035 [Michael Helbling]: It's a simple, powerful way for marketing and data deems to measure the true impact of their advertising spend, and even better, you can use it completely free for six months. 518 00:46:09,095 --> 00:46:14,925 [Michael Helbling]: Just visit getrecast.com slash geolift to start your trial today. 519 00:46:15,293 --> 00:46:17,419 [Michael Helbling]: Okay, well, we've got that done. 520 00:46:17,459 --> 00:46:25,262 [Michael Helbling]: One thing we'd love to do is go around the horn and share something we call last call, something of interest that might be of interest to our listeners. 521 00:46:25,723 --> 00:46:26,827 [Michael Helbling]: Cindi, you're our guest. 522 00:46:26,847 --> 00:46:28,712 [Michael Helbling]: Do you have a last call you'd like to share? 523 00:46:29,266 --> 00:46:32,710 [Cindi Howson]: Well, I want to ask a question if I can on the last call. 524 00:46:33,411 --> 00:46:44,063 [Cindi Howson]: And when you think about how quickly our industry is moving and innovating, what do you see as your best method media to keep up? 525 00:46:44,664 --> 00:46:52,513 [Cindi Howson]: Is it listening to podcasts, reading, substack or medium articles, or how do you feel about books? 526 00:46:53,374 --> 00:46:54,936 [Michael Helbling]: Are we supposed to answer that? 527 00:46:55,237 --> 00:47:06,450 [Cindi Howson]: Well, I'm looking for feedback because you know, even though I'm a podcast host, I'm a writer at heart and yet is the industry moving too quickly for another book? 528 00:47:08,292 --> 00:47:08,492 [Moe Kiss]: Yeah. 529 00:47:08,512 --> 00:47:10,134 [Moe Kiss]: I mean, I can speak for myself. 530 00:47:10,154 --> 00:47:13,738 [Moe Kiss]: I listen to podcasts and host a podcast. 531 00:47:13,778 --> 00:47:15,680 [Moe Kiss]: That's a big part of how I stay up to date. 532 00:47:16,441 --> 00:47:17,703 [Moe Kiss]: But I also, I love books. 533 00:47:17,923 --> 00:47:18,744 [Moe Kiss]: I'm a book person. 534 00:47:18,884 --> 00:47:20,766 [Moe Kiss]: Probably books more than articles. 535 00:47:21,407 --> 00:47:23,269 [Moe Kiss]: But you listen to a lot of the books, right? 536 00:47:24,008 --> 00:47:28,834 [Moe Kiss]: Yes, I do, but that's just because of my life stage of being time poor. 537 00:47:28,854 --> 00:47:32,699 [Moe Kiss]: I end up listening to books on Audible a lot. 538 00:47:33,500 --> 00:47:33,961 [Moe Kiss]: Yeah, for sure. 539 00:47:34,882 --> 00:47:35,983 [Michael Helbling]: What about you, House? 540 00:47:36,003 --> 00:47:39,888 [Michael Helbling]: I would say my number one source is articles. 541 00:47:40,929 --> 00:47:45,615 [Michael Helbling]: So in my day-to-day travels, I'll run across an article and then bookmark it and read it later. 542 00:47:45,635 --> 00:47:47,137 [Michael Helbling]: So I'll do that. 543 00:47:47,337 --> 00:47:49,961 [Michael Helbling]: I buy a lot of books and then don't read them. 544 00:47:51,142 --> 00:47:51,783 [Michael Helbling]: Oh boy. 545 00:47:51,803 --> 00:47:53,425 [Michael Helbling]: In fact, that's... 546 00:47:53,725 --> 00:47:54,648 [Michael Helbling]: Right behind me. 547 00:47:55,250 --> 00:47:57,677 [Michael Helbling]: Michael, have you finished the book? 548 00:47:58,440 --> 00:47:59,804 [Michael Helbling]: I have not finished your book, Tim. 549 00:47:59,824 --> 00:48:01,710 [Michael Helbling]: Oh, well, you haven't finished that either. 550 00:48:02,450 --> 00:48:10,699 [Michael Helbling]: Uh, so yeah, but I, so I don't, cause for me reading is sort of like an enjoyable pastime. 551 00:48:10,899 --> 00:48:15,565 [Michael Helbling]: And I, unlike Moe, I can't pay attention if someone's reading it aloud or audio books. 552 00:48:15,605 --> 00:48:17,407 [Michael Helbling]: So I have to sit down and read it. 553 00:48:17,767 --> 00:48:24,194 [Michael Helbling]: And then when I do finally get a chance to read, I end up reading like sci-fi or fantasy novels instead of business books. 554 00:48:24,274 --> 00:48:25,756 [Michael Helbling]: So it's, it's a tough one. 555 00:48:26,196 --> 00:48:29,300 [Michael Helbling]: And then of course, of course podcasts are very important. 556 00:48:29,500 --> 00:48:30,982 [Michael Helbling]: I have to believe that, right? 557 00:48:31,382 --> 00:48:32,203 [Michael Helbling]: So there you go. 558 00:48:32,183 --> 00:48:35,106 [Cindi Howson]: This feels like confessions of a podcast host. 559 00:48:35,887 --> 00:48:36,908 [Michael Helbling]: Yeah, that's right. 560 00:48:36,928 --> 00:48:37,369 [Michael Helbling]: Exactly. 561 00:48:39,712 --> 00:48:40,853 [Michael Helbling]: What do you think Tim? 562 00:48:40,873 --> 00:48:41,874 [Michael Helbling]: I listen to a ton of podcasts. 563 00:48:42,475 --> 00:48:43,175 [Michael Helbling]: Yeah, he does. 564 00:48:43,276 --> 00:48:49,463 [Tim Wilson]: I listen to a ton of podcasts and very few of them are business or data analytics related. 565 00:48:49,583 --> 00:48:54,208 [Tim Wilson]: So I am very much the subscribe to, I mean, a medium substack. 566 00:48:55,268 --> 00:49:00,521 [Tim Wilson]: daily weekly newsletter fiend, which starts to feel a little overwhelming. 567 00:49:00,621 --> 00:49:05,633 [Tim Wilson]: But yeah, so with the occasional book. 568 00:49:06,828 --> 00:49:08,070 [Tim Wilson]: The books feel like a chore, though. 569 00:49:08,651 --> 00:49:11,215 [Tim Wilson]: Well, I feel like if someone else is doing a good job. 570 00:49:11,235 --> 00:49:11,395 [Tim Wilson]: So cool. 571 00:49:11,415 --> 00:49:12,157 [Tim Wilson]: I'm just going to be clear. 572 00:49:12,217 --> 00:49:15,963 [Tim Wilson]: So I don't listen to the podcast, even though I make one. 573 00:49:16,083 --> 00:49:17,025 [Tim Wilson]: And I don't tend to read. 574 00:49:17,445 --> 00:49:19,589 [Tim Wilson]: I struggle to read the books, even though I wrote one. 575 00:49:19,709 --> 00:49:23,014 [Tim Wilson]: So yeah, I'm the worst. 576 00:49:23,034 --> 00:49:23,235 [Cindi Howson]: Wow. 577 00:49:23,355 --> 00:49:25,218 [Cindi Howson]: So I think Tim summed it up. 578 00:49:25,278 --> 00:49:30,186 [Cindi Howson]: Wait, are you telling me two-third of our time spent is like a waste of time? 579 00:49:30,166 --> 00:49:33,862 [Cindi Howson]: Why am I writing books and why am I hosting a podcast? 580 00:49:33,882 --> 00:49:35,649 [Cindi Howson]: I'm just gonna get on with building stuff. 581 00:49:36,392 --> 00:49:36,774 [Cindi Howson]: Okay. 582 00:49:36,794 --> 00:49:40,530 [Michael Helbling]: I don't like the data that we've uncovered here. 583 00:49:42,046 --> 00:49:50,622 [Tim Wilson]: I mean, I get a lot of value out of hosting the podcast because we get to have excuses to say, hey, why don't you come on and explain semantic layers to us? 584 00:49:50,662 --> 00:49:56,613 [Michael Helbling]: So yeah, that is actually doing a podcast is one of the ways I learn new things. 585 00:49:56,714 --> 00:49:58,798 [Michael Helbling]: So that's something you could add to the mix. 586 00:49:59,599 --> 00:49:59,880 [Michael Helbling]: Yeah. 587 00:50:00,921 --> 00:50:02,805 [Michael Helbling]: So when is your next book coming out? 588 00:50:03,308 --> 00:50:03,849 [Cindi Howson]: I don't know. 589 00:50:03,949 --> 00:50:07,956 [Cindi Howson]: Can I take a break from the podcast or stop something? 590 00:50:08,417 --> 00:50:09,018 [Cindi Howson]: I don't know. 591 00:50:09,218 --> 00:50:09,679 [Cindi Howson]: I don't know. 592 00:50:09,779 --> 00:50:11,582 [Michael Helbling]: This is what I was trying to figure out. 593 00:50:12,203 --> 00:50:13,184 [Cindi Howson]: What should I do next? 594 00:50:13,285 --> 00:50:13,525 [Cindi Howson]: Yeah. 595 00:50:13,765 --> 00:50:14,226 [Cindi Howson]: Yeah. 596 00:50:14,246 --> 00:50:14,867 [Michael Helbling]: Fair point. 597 00:50:15,969 --> 00:50:16,390 [Michael Helbling]: All right. 598 00:50:17,011 --> 00:50:18,073 [Michael Helbling]: Tim, what about you? 599 00:50:18,113 --> 00:50:19,495 [Michael Helbling]: What's your last call? 600 00:50:19,947 --> 00:50:21,830 [Tim Wilson]: Well, I guess follow on. 601 00:50:22,271 --> 00:50:28,300 [Tim Wilson]: There is a sub-stack that I discovered a couple of months ago from somewhere that is We Have the Data. 602 00:50:28,821 --> 00:50:29,883 [Tim Wilson]: It's kind of silly. 603 00:50:30,043 --> 00:50:35,111 [Tim Wilson]: It's kind of data visualization candy, but it's WeHaveTheData.net. 604 00:50:35,692 --> 00:50:39,238 [Tim Wilson]: I think it's a couple of times a week, and it's just kind of a 605 00:50:39,218 --> 00:50:42,744 [Tim Wilson]: It's like NOMLAC news, but data visualizations instead. 606 00:50:42,785 --> 00:50:45,570 [Tim Wilson]: So they're pretty lengthy. 607 00:50:45,610 --> 00:50:53,163 [Tim Wilson]: They're a collection of often kind of trivial data visualizations, but it's kind of a fun scroll in my inbox. 608 00:50:54,666 --> 00:50:55,367 [Michael Helbling]: Outstanding. 609 00:50:56,229 --> 00:50:57,852 [Michael Helbling]: All right, Moe, what about you? 610 00:50:59,063 --> 00:51:02,347 [Moe Kiss]: I want to do a plug for Cindi's podcast. 611 00:51:02,728 --> 00:51:08,536 [Moe Kiss]: I was lucky enough to be a guest back in October and it's called The Data Chief. 612 00:51:08,856 --> 00:51:18,209 [Moe Kiss]: And as you can tell, I ended up hanging out after the show and picking Cindi's brain for like another 30, 40 minutes about all of these topics, which is why she's here today. 613 00:51:18,249 --> 00:51:22,715 [Moe Kiss]: And she just has such a range of like really incredible guests. 614 00:51:22,795 --> 00:51:24,677 [Moe Kiss]: It's a really different format to our show. 615 00:51:24,718 --> 00:51:28,783 [Moe Kiss]: So really encourage you to go check out The Data Chief podcast. 616 00:51:29,540 --> 00:51:35,991 [Michael Helbling]: I'm standing in, and yeah, we'll put a link to that in our show notes as well, so people can find it. 617 00:51:36,372 --> 00:51:37,874 [Tim Wilson]: You're supposed to hand her at the beginning. 618 00:51:37,995 --> 00:51:39,477 [Michael Helbling]: It's fine. 619 00:51:39,818 --> 00:51:41,240 [Michael Helbling]: We'll hand her all over the place. 620 00:51:41,641 --> 00:51:42,563 [Michael Helbling]: What's your last call? 621 00:51:43,164 --> 00:51:46,249 [Michael Helbling]: Well, I'm so glad you asked him. 622 00:51:46,229 --> 00:51:50,234 [Michael Helbling]: So a good friend of mine, Mary Gates, actually made me aware of this. 623 00:51:50,534 --> 00:51:58,344 [Michael Helbling]: So Informs, which I'm sure we're all familiar with, they have an initiative called Pro Bono Analytics. 624 00:51:58,384 --> 00:52:14,565 [Michael Helbling]: So I'm a big fan of any analytics initiatives that I've been able to be part of them over the years that help nonprofits and allow people to give of their skills and data and analytics to nonprofits and mentorship and things like that. 625 00:52:14,545 --> 00:52:18,470 [Michael Helbling]: Pro Bono Analytics is an initiative run by Informs. 626 00:52:19,291 --> 00:52:21,333 [Michael Helbling]: And so I just wanted to give that a shout out. 627 00:52:21,353 --> 00:52:26,079 [Michael Helbling]: I was not familiar with this before, but it looks like a very cool organization. 628 00:52:26,119 --> 00:52:33,849 [Michael Helbling]: And so if you're a nonprofit and you're listening, that might be an amazing place to partner with them to get help with data initiatives. 629 00:52:34,049 --> 00:52:39,896 [Michael Helbling]: And if you're a professional in working in data and you want to find a way to give back, that might be an amazing way to do that. 630 00:52:40,037 --> 00:52:42,800 [Michael Helbling]: So we'll put a link to that in the show as well. 631 00:52:43,261 --> 00:52:43,641 [Michael Helbling]: OK. 632 00:52:43,621 --> 00:52:48,892 [Michael Helbling]: As you've been listening about on this topic of semantic layers, I'm sure you have thoughts. 633 00:52:48,932 --> 00:52:49,914 [Michael Helbling]: I'm sure you have questions. 634 00:52:49,994 --> 00:52:51,277 [Michael Helbling]: We would love to hear from you. 635 00:52:51,818 --> 00:52:53,481 [Michael Helbling]: Go ahead and reach out to us. 636 00:52:53,682 --> 00:52:55,305 [Michael Helbling]: And there's three main ways you can do that. 637 00:52:55,385 --> 00:53:04,043 [Michael Helbling]: You can do that through LinkedIn or the measure slack chat group, or you can email us at contact at analyticshour.io. 638 00:53:04,023 --> 00:53:07,148 [Michael Helbling]: And yeah, we'd love to hear from you. 639 00:53:07,989 --> 00:53:17,525 [Michael Helbling]: Cindi, once again, this has been a very information-rich and awesome episode, and primarily because your deep knowledge and expertise in this field. 640 00:53:17,565 --> 00:53:20,109 [Michael Helbling]: So thank you again so much for joining. 641 00:53:21,431 --> 00:53:22,693 [Cindi Howson]: Thank you for having me. 642 00:53:22,713 --> 00:53:28,663 [Cindi Howson]: I feel like we should do this over a cup of coffee or a glass of wine at some point. 643 00:53:28,711 --> 00:53:30,775 [Michael Helbling]: Yes, I wholeheartedly agree. 644 00:53:30,795 --> 00:53:38,973 [Michael Helbling]: That's how this whole podcast started was because we're all drinking at an analytics conference and said, we should put this on the radio. 645 00:53:39,795 --> 00:53:42,641 [Michael Helbling]: That's a great idea. 646 00:53:42,681 --> 00:53:43,202 [Michael Helbling]: That's right. 647 00:53:43,262 --> 00:53:46,028 [Michael Helbling]: Another drunken, great ideas. 648 00:53:46,008 --> 00:53:46,509 [Michael Helbling]: All right. 649 00:53:46,870 --> 00:53:52,419 [Michael Helbling]: Also, if you are somebody who puts and is not directed at you, Cindi, this is back to the audience. 650 00:53:53,642 --> 00:53:58,450 [Michael Helbling]: If you're someone who puts stickers on your laptops or whatever, we do have stickers and we'd love to send you one. 651 00:53:58,851 --> 00:54:01,916 [Michael Helbling]: You can actually request one on our website so you can go and do that. 652 00:54:02,718 --> 00:54:04,000 [Michael Helbling]: And then 653 00:54:03,980 --> 00:54:17,241 [Michael Helbling]: Obviously, no show would be complete without saying a huge thank you to all of you listeners who go out and share ratings and reviews with us and tell us how you're enjoying the show. 654 00:54:17,321 --> 00:54:18,603 [Michael Helbling]: So please continue to do that. 655 00:54:18,663 --> 00:54:20,005 [Michael Helbling]: We look forward to that feedback. 656 00:54:20,145 --> 00:54:21,427 [Michael Helbling]: We appreciate it very much. 657 00:54:22,389 --> 00:54:22,790 [Michael Helbling]: All right. 658 00:54:23,451 --> 00:54:28,198 [Michael Helbling]: As we wrap up, I know that no matter if you're trying to 659 00:54:29,495 --> 00:54:34,762 [Michael Helbling]: Build one ring to rule them all type of semantic layers, or if you're spreading it out across verticals. 660 00:54:36,184 --> 00:54:39,028 [Michael Helbling]: I know both of my co-hosts, Tim and Moe, would agree with me. 661 00:54:39,889 --> 00:54:40,990 [Michael Helbling]: You should keep analyzing. 662 00:54:41,771 --> 00:54:42,552 [Announcer]: Thanks for listening. 663 00:54:43,053 --> 00:54:55,510 [Announcer]: Let's keep the conversation going with your comments, suggestions, and questions on Twitter at @analyticshour, on the web at analyticshour.io, our LinkedIn group, and the Measure Chat Slack group. 664 00:54:56,190 --> 00:54:58,894 [Announcer]: Music for the podcast by Josh Grohurst. 665 00:54:59,617 --> 00:55:04,081 [Announcer]: Those smart guys wanted to fit in, so they made up a term called analytics. 666 00:55:04,101 --> 00:55:05,062 [Announcer]: Analytics don't work. 667 00:55:06,363 --> 00:55:09,066 [Charles Barkley]: Do the analytics say go for it, no matter who's going for it? 668 00:55:09,406 --> 00:55:12,249 [Charles Barkley]: So if you and I were on the field, the analytics say go for it. 669 00:55:12,569 --> 00:55:18,534 [Charles Barkley]: It's the stupidest, laziest, lamest thing I've ever heard for reasoning in competition. 670 00:55:19,075 --> 00:55:21,017 [Michael Helbling]: We'll just do our best with it. 671 00:55:21,457 --> 00:55:22,818 [Michael Helbling]: It's why we have an audio engineer. 672 00:55:24,280 --> 00:55:26,341 [Michael Helbling]: Hi, Tony. 673 00:55:26,362 --> 00:55:28,003 [Michael Helbling]: Hi, Tony. 674 00:55:37,045 --> 00:55:42,651 [Tim Wilson]: Rock flag and semantic layers are 30 years old.