Meeting Title: Omni x Nautica Training Session Date: 2026-05-01 Meeting participants: Greg Stoutenburg, Rebecca Bruggman, Uttam Kumaran, Nandika Jhunjhunwala, Caitlyn Vaughn
WEBVTT
1 00:00:30.660 ⇒ 00:00:33.959 Rebecca Bruggman: Hello! Did I get booted, or did everyone get booted?
2 00:00:34.910 ⇒ 00:00:36.609 Greg Stoutenburg: Sorry, that was me.
3 00:00:36.610 ⇒ 00:00:39.129 Rebecca Bruggman: I was wondering, I was like, oh no!
4 00:00:39.510 ⇒ 00:00:40.459 Rebecca Bruggman: I guess they’re not, like.
5 00:00:40.460 ⇒ 00:00:40.830 Greg Stoutenburg: Yup.
6 00:00:41.760 ⇒ 00:00:47.739 Greg Stoutenburg: Classic, no, I’ll never not do it. I shouldn’t be… I shouldn’t be trusted to do this.
7 00:00:47.740 ⇒ 00:00:50.350 Rebecca Bruggman: Oh, good.
8 00:00:52.160 ⇒ 00:00:53.890 Greg Stoutenburg: Sorry.
9 00:00:54.510 ⇒ 00:00:56.030 Greg Stoutenburg: Sorry.
10 00:00:56.740 ⇒ 00:00:59.750 Rebecca Bruggman: Oh, good. We’ll give everyone a minute to jump back in.
11 00:01:01.760 ⇒ 00:01:03.910 Greg Stoutenburg: Technical difficulties.
12 00:01:04.819 ⇒ 00:01:05.929 Nandika Jhunjhunwala: All good.
13 00:01:05.930 ⇒ 00:01:15.299 Greg Stoutenburg: The difficulty is that I should… I should have to have a third and fourth check before I can end a meeting. Alright, let me try that again.
14 00:01:16.140 ⇒ 00:01:17.440 Greg Stoutenburg: Competently.
15 00:01:18.040 ⇒ 00:01:22.649 Greg Stoutenburg: Go ahead and keep going, I’m just gonna, I’m just gonna make, Utam the meeting owner.
16 00:01:22.650 ⇒ 00:01:23.310 Rebecca Bruggman: Okay.
17 00:01:23.310 ⇒ 00:01:23.740 Greg Stoutenburg: seed.
18 00:01:24.240 ⇒ 00:01:24.980 Rebecca Bruggman: Okay.
19 00:01:24.980 ⇒ 00:01:25.900 Uttam Kumaran: Thanks, Greg.
20 00:01:26.190 ⇒ 00:01:29.789 Rebecca Bruggman: Oh, good. Okay, let me reshare. Is that coming through for everyone?
21 00:01:30.210 ⇒ 00:01:31.130 Uttam Kumaran: Yes. Yes.
22 00:01:31.130 ⇒ 00:01:48.669 Rebecca Bruggman: So I was just noting that, especially with, like, a lot of these fields where they have, like, similar names, but you know they have different meanings, that’s where, like, having the topic curation and also having, like, the AI context curation really matters, versus just sort of, like.
23 00:01:48.700 ⇒ 00:02:09.039 Rebecca Bruggman: unlocking everything to Blobby, because then it can just lead to not, like, not the specific answers you actually want to be given by Blobby to unlock that self-serve. So, that’s why we do have the default recommendation to, have it be, searching through topics, which you can kind of see within the chat of, like, auto-select a topic.
24 00:02:09.039 ⇒ 00:02:18.229 Rebecca Bruggman: You do have the option to open up for folks who have the right permissions levels to basically query all views and fields. You just run the risk of, like.
25 00:02:18.230 ⇒ 00:02:24.080 Rebecca Bruggman: Kind of confusing answers, or it not actually going to, like, getting the right answers that you want, if that makes sense.
26 00:02:26.090 ⇒ 00:02:28.359 Caitlyn Vaughn: So it’s better to pick a topic?
27 00:02:29.240 ⇒ 00:02:53.050 Rebecca Bruggman: Yes, because then you can do the… you’re… it’s better to have the setting basically be that the AI always builds off of topics, because that’s where you’re getting that, like, semantic layer goodness of basically saying, we have all this raw data, but again, like, some of these tables feel very similar, and it’s, like, hard to sort of differentiate between the ones to use. But if you have the topics where,
28 00:02:53.050 ⇒ 00:02:54.459 Rebecca Bruggman: I’ll just go in.
29 00:02:55.010 ⇒ 00:02:59.190 Rebecca Bruggman: Gonna show… 1… Da-da-da-da…
30 00:03:02.070 ⇒ 00:03:06.930 Rebecca Bruggman: Or, like, this one, you have, like, a ton of AI contacts in here, which is super helpful.
31 00:03:06.930 ⇒ 00:03:07.450 Caitlyn Vaughn: Because.
32 00:03:07.450 ⇒ 00:03:26.859 Rebecca Bruggman: Then, when Blobby is, like, looking through all the topics of, okay, this person has asked me this question, what topic do I probably need to pull from? Like, all of this additional information in the semantic layer really helps guide where the, like, answers to that question are going to be pulled from, if that makes sense.
33 00:03:27.070 ⇒ 00:03:27.750 Caitlyn Vaughn: Hmm.
34 00:03:31.010 ⇒ 00:03:43.409 Rebecca Bruggman: So yeah, that’s where, like, the upfront investment of having, like, the topics that are answering, like, the key questions, like, are super helpful, because again, then, like, if we had this open to,
35 00:03:43.740 ⇒ 00:03:52.970 Rebecca Bruggman: you know, all your tables and looking for opportunities, it would probably be doing a lot of what we’re doing right now, being like, which table do I go to?
36 00:03:52.970 ⇒ 00:03:53.530 Caitlyn Vaughn: Yeah.
37 00:03:53.530 ⇒ 00:03:57.970 Rebecca Bruggman: And, like, the real answer is, like, the data doesn’t actually exist how we need it to be.
38 00:03:57.970 ⇒ 00:04:00.560 Caitlyn Vaughn: Okay, interesting.
39 00:04:00.980 ⇒ 00:04:03.859 Caitlyn Vaughn: So, could you walk us through building a topic?
40 00:04:03.860 ⇒ 00:04:06.289 Rebecca Bruggman: Yeah, yeah, of course. Cool.
41 00:04:06.610 ⇒ 00:04:15.039 Rebecca Bruggman: Okay, so let’s do… we’re gonna do DIM Customer360. We’re in All Views and Fields. Let’s click the triple dot, Modeling, make a topic.
42 00:04:15.690 ⇒ 00:04:18.929 Rebecca Bruggman: And from here, we’ll open up this nice…
43 00:04:18.930 ⇒ 00:04:37.269 Rebecca Bruggman: like, topic builder. This is why I personally like to do it in the UI. Otherwise you’re, like, building in YAML, indentation gets funky, and a lot of this just gets, like, done for you using this UI, which is my personal preference. Cool, what do we want to name this? Like, Opportunities Topic? Does that feel right for now?
44 00:04:37.270 ⇒ 00:04:38.030 Caitlyn Vaughn: Yeah.
45 00:04:38.520 ⇒ 00:04:49.599 Rebecca Bruggman: Okay, cool. The ops topic. And then what’s nice is you can put it into a topic group, so that’s that kind of foldering system that you’re seeing that already exists.
46 00:04:49.600 ⇒ 00:04:50.580 Caitlyn Vaughn: Should we…
47 00:04:50.580 ⇒ 00:04:56.050 Rebecca Bruggman: Should we put this in a current folder, or is there a new folder we could put it in that might make sense?
48 00:04:59.380 ⇒ 00:05:01.209 Uttam Kumaran: Nautica, I don’t know if you have thoughts.
49 00:05:02.280 ⇒ 00:05:07.129 Nandika Jhunjhunwala: I think maybe under go-to-market sales?
50 00:05:07.500 ⇒ 00:05:08.789 Rebecca Bruggman: Cool, it’s put there.
51 00:05:09.030 ⇒ 00:05:09.360 Nandika Jhunjhunwala: Okay.
52 00:05:09.360 ⇒ 00:05:14.960 Rebecca Bruggman: And then we can say, topic for,
53 00:05:15.760 ⇒ 00:05:22.140 Rebecca Bruggman: Answering… and you can update all of this, I’m just putting something in so you can please be answering questions.
54 00:05:22.280 ⇒ 00:05:29.370 Rebecca Bruggman: Related to opportunities for… What am I gonna do? Oh, I think I might have spoil it right. Okay.
55 00:05:29.520 ⇒ 00:05:32.750 Rebecca Bruggman: And then this will show up.
56 00:05:33.070 ⇒ 00:05:35.660 Rebecca Bruggman: to end users as well as to Blobby.
57 00:05:35.880 ⇒ 00:05:36.320 Nandika Jhunjhunwala: Got it.
58 00:05:36.320 ⇒ 00:05:49.999 Rebecca Bruggman: when they’re selecting it. So that can be the nice thing. This is just going back to that, like, good UI UX experience, like, almost thinking of yourself as, like, the product manager of the Omnidata product at your company, of just, like, making sure self-service is as straightforward as possible.
59 00:05:50.290 ⇒ 00:06:01.669 Rebecca Bruggman: I’ll leave this blank for now, but you can also add some of the eye contacts. Actually, we’ll click the… the little… oh, my… I always joke that look for sparkles, that’s where the AI always lives.
60 00:06:01.670 ⇒ 00:06:02.140 Nandika Jhunjhunwala: Sure.
61 00:06:02.140 ⇒ 00:06:02.730 Rebecca Bruggman: If you’re a notch.
62 00:06:02.730 ⇒ 00:06:03.380 Nandika Jhunjhunwala: Yeah.
63 00:06:03.380 ⇒ 00:06:14.949 Rebecca Bruggman: You can add your own AI context, or this is basically saying I’m gonna… based on what we have in here, I’m going to generate some AI context. This is actually a little bit big.
64 00:06:15.360 ⇒ 00:06:34.060 Rebecca Bruggman: That you can just pop in here if you want to do that as well. So, optional things there. We’ll go into joins. Now, if this already has joins as defined within the relationships file, which is basically, like, the master file that, shows, like, how all tables relate to each
65 00:06:35.060 ⇒ 00:06:35.830 Rebecca Bruggman: in,
66 00:06:35.830 ⇒ 00:06:42.369 Nandika Jhunjhunwala: What would be sick to do, is that outside of this, or where does that live? Sorry, to, like, sidetrack.
67 00:06:42.370 ⇒ 00:06:46.250 Rebecca Bruggman: No, no, you’re all good. Great question. So relationships are under the shared model, they’re right here.
68 00:06:46.250 ⇒ 00:06:47.050 Nandika Jhunjhunwala: Okay.
69 00:06:48.000 ⇒ 00:07:04.949 Rebecca Bruggman: I’m not seeing any, often these will get auto-generated if you have, like, primary keys already defined. Like, basically, if there’s something that Omni can infer to know that, like, the table’s already related to each other, you can also just generate them on the fly, but,
70 00:07:04.950 ⇒ 00:07:11.669 Rebecca Bruggman: Utam, maybe I’ll ask, do you know why, like, there might not already be, sort of, relationships defined within here, or, like…
71 00:07:11.670 ⇒ 00:07:21.770 Uttam Kumaran: I don’t know if we actually… I can ask the team. I’m not sure if we actually use the relationships much. I think we’re just… we just had these, like, flat, wide, kind of, like, marts tables.
72 00:07:21.770 ⇒ 00:07:22.180 Rebecca Bruggman: Okay.
73 00:07:22.180 ⇒ 00:07:31.270 Uttam Kumaran: Typically. But again, I think right now, like, it’s even helpful feedback for me to say, okay, if we want to move to, sort of, like.
74 00:07:31.820 ⇒ 00:07:40.930 Uttam Kumaran: sort of thinner, maybe just, like, join in all these views together through the relationships, and then let Blobby kind of infer. Like, that’s one thing that we can do.
75 00:07:41.910 ⇒ 00:07:47.790 Nandika Jhunjhunwala: that would be super sick. For example, all the account, like, tables of account IDs, like.
76 00:07:47.820 ⇒ 00:08:01.949 Nandika Jhunjhunwala: just defining those relationships here would, like, I think really simplify the cross-functional queries leadership might want to do, and maybe that could also help us consolidate some topics here.
77 00:08:02.180 ⇒ 00:08:07.549 Nandika Jhunjhunwala: Like, if we have, like, ARR broken down by, like, monthly and components and things like that, maybe…
78 00:08:07.730 ⇒ 00:08:09.759 Nandika Jhunjhunwala: And Rebecca, like, tell me if I’m, like.
79 00:08:09.760 ⇒ 00:08:10.480 Rebecca Bruggman: No, no, no, that’s great.
80 00:08:10.480 ⇒ 00:08:11.460 Nandika Jhunjhunwala: Incorrectly, but…
81 00:08:11.460 ⇒ 00:08:12.180 Rebecca Bruggman: No, you’re good.
82 00:08:12.180 ⇒ 00:08:16.490 Nandika Jhunjhunwala: If we do define those relationships, is there, like, more scope for us to have, like.
83 00:08:16.940 ⇒ 00:08:25.720 Nandika Jhunjhunwala: broadly defined topics that leadership can go into one of those topics and get all that data that they want, but, like, instead of going into, like.
84 00:08:25.850 ⇒ 00:08:27.940 Nandika Jhunjhunwala: Two or three separate topics.
85 00:08:27.940 ⇒ 00:08:29.859 Rebecca Bruggman: That’s bang on, yeah. So…
86 00:08:29.860 ⇒ 00:08:30.260 Nandika Jhunjhunwala: Okay.
87 00:08:30.260 ⇒ 00:08:39.539 Rebecca Bruggman: It kind of depends, the caveat I’ll just say is it, like, sort of depends how you, like, approach things with DBT, because, like, I’ve definitely gone on the phone with customers before where I’m like.
88 00:08:39.549 ⇒ 00:08:39.979 Nandika Jhunjhunwala: Hmm.
89 00:08:39.980 ⇒ 00:08:43.369 Rebecca Bruggman: what can we join? And they’re like, literally nothing, because everything.
90 00:08:43.370 ⇒ 00:08:44.030 Nandika Jhunjhunwala: Nothing.
91 00:08:44.039 ⇒ 00:08:51.629 Rebecca Bruggman: It’s a broad white table. So, you know, if that’s sort of the approach that you’re taking, where, like, everything is just sort of, like, pre-modeled in dbt, then.
92 00:08:51.630 ⇒ 00:08:52.240 Nandika Jhunjhunwala: Potentially.
93 00:08:52.240 ⇒ 00:08:55.709 Rebecca Bruggman: like, that’s where you’re doing things. Again, just sort of depends on your approach.
94 00:08:55.710 ⇒ 00:09:16.239 Rebecca Bruggman: But if you want to have more, sort of, like, distinct tables where you have more flexibility in terms of not having to, sort of, like, pre-aggregate stuff so much… Yes. You can do joins within Omni, and that allows you to not have to do so much, like, big wide tables with pre-aggregation, just gives you a little more flexibility, which can be nice.
95 00:09:16.670 ⇒ 00:09:17.330 Nandika Jhunjhunwala: Yeah.
96 00:09:17.730 ⇒ 00:09:19.850 Nandika Jhunjhunwala: Yeah, thank you, that sounds great.
97 00:09:19.850 ⇒ 00:09:20.340 Rebecca Bruggman: Cool.
98 00:09:20.340 ⇒ 00:09:20.970 Nandika Jhunjhunwala: Yes.
99 00:09:21.430 ⇒ 00:09:25.969 Rebecca Bruggman: So we can actually, if we come back here, we can actually add a join right here. Again.
100 00:09:26.520 ⇒ 00:09:35.749 Rebecca Bruggman: where I love the UI components, it just does a lot for you right out of the box. So, I’m gonna new. From Dim Customer, what’s another table we’d want to join to from here?
101 00:09:36.720 ⇒ 00:09:42.560 Nandika Jhunjhunwala: We can do Salesforce for sake of example, or any customer table.
102 00:09:42.560 ⇒ 00:09:43.099 Rebecca Bruggman: This one?
103 00:09:43.100 ⇒ 00:09:44.220 Nandika Jhunjhunwala: Yeah, yeah.
104 00:09:44.220 ⇒ 00:09:51.190 Rebecca Bruggman: Perfect. Alright, we’ll do account ID, account ID, we can even infer relationship, I love this too. Does one-to-one seem correct?
105 00:09:51.190 ⇒ 00:09:52.040 Nandika Jhunjhunwala: One go on.
106 00:09:52.170 ⇒ 00:09:54.049 Rebecca Bruggman: Beautiful. Okay, so then we’ll add.
107 00:09:54.680 ⇒ 00:10:03.259 Rebecca Bruggman: So now we’ve brought this in. So you’ll see automatically, this comes in as another table within this topic.
108 00:10:03.730 ⇒ 00:10:15.280 Rebecca Bruggman: And so you’ll also know my name for this is the little pancake stack. But, this is sort of, like, collecting any changes we’re making at the workbook level that we might want to promote.
109 00:10:15.280 ⇒ 00:10:26.349 Rebecca Bruggman: up to the shared level. So here we have our topic, but things we can also promote up are things like the relationship that we just built here to put it into the relationship file, if we so choose.
110 00:10:26.830 ⇒ 00:10:27.430 Rebecca Bruggman: Mmm.
111 00:10:27.430 ⇒ 00:10:27.770 Nandika Jhunjhunwala: Hmm.
112 00:10:27.770 ⇒ 00:10:41.089 Rebecca Bruggman: So you can just continue to do that from here, like, if you want to add to the relationships file here, you can do it, like, write in notes, sort of, let’s see… Let’s see… join…
113 00:10:41.150 ⇒ 00:10:58.640 Rebecca Bruggman: I don’t know why it’s not doing the auto-completing. You can just, like, type in here sort of the, like, join paths within here as well, but I just find that, like, doing it in the UI and then, like, promoting it up, and then it just auto-creates the YAML for you is, like, personally how I prefer to do things. You don’t have to worry about, like, indentation, remembering the syntax, stuff like that.
114 00:10:59.440 ⇒ 00:11:02.130 Nandika Jhunjhunwala: And, sorry, just to confirm, when you say promoting it.
115 00:11:02.400 ⇒ 00:11:06.819 Nandika Jhunjhunwala: If you shared folder, that means that topic becomes publicly available.
116 00:11:06.960 ⇒ 00:11:07.510 Rebecca Bruggman: Correct.
117 00:11:07.510 ⇒ 00:11:08.490 Nandika Jhunjhunwala: Or… okay.
118 00:11:08.860 ⇒ 00:11:23.220 Rebecca Bruggman: Yep. So right now, where we’re building is just at the workbook level, and so it would just be contained and be able to be used within this workbook, but if you wanted it to be, like, available to anyone who opened up a new workbook, then you’d have to put it into the shared model.
119 00:11:23.220 ⇒ 00:11:29.599 Nandika Jhunjhunwala: Got it. And… one workbook has a SQL query for one chart or insight.
120 00:11:30.210 ⇒ 00:11:31.079 Rebecca Bruggman: So you can have…
121 00:11:31.080 ⇒ 00:11:31.949 Nandika Jhunjhunwala: can it be?
122 00:11:32.060 ⇒ 00:11:32.849 Nandika Jhunjhunwala: It can be both.
123 00:11:32.850 ⇒ 00:11:34.210 Rebecca Bruggman: things. So you can do…
124 00:11:34.210 ⇒ 00:11:34.690 Nandika Jhunjhunwala: Hmm.
125 00:11:34.690 ⇒ 00:11:40.850 Rebecca Bruggman: Like, new tabs along here. So, like, right now we’re building a topic, but I could just sort of, like, open up a bunch of new tabs.
126 00:11:40.850 ⇒ 00:12:01.869 Rebecca Bruggman: And anytime you open a new tab, you can pick a new topic to build on, you can pick SQL, you can go to Browse All Views, so you can pull in a bunch of different topics into a single dashboard. And that’s why I go back to, like, topics answering questions, because you can pull in a single or multiple topics into a single dashboard. It doesn’t have to be, like.
127 00:12:01.980 ⇒ 00:12:10.269 Rebecca Bruggman: one topic answers every single question for every single file and query, you can have that sort of, like, flexibility, which can be nice.
128 00:12:12.800 ⇒ 00:12:13.550 Rebecca Bruggman: Cool.
129 00:12:13.550 ⇒ 00:12:15.200 Caitlyn Vaughn: Return back, sorry.
130 00:12:15.660 ⇒ 00:12:17.440 Rebecca Bruggman: Doing good. Welcome back.
131 00:12:17.590 ⇒ 00:12:29.270 Rebecca Bruggman: Cool. Alright, so just to catch you up, we started building a topic, we added a join at the workbook level, to add more, tables into this topic.
132 00:12:29.270 ⇒ 00:12:38.340 Rebecca Bruggman: You can continue adding them, so, like, as many as you think make sense, we can do that, but we’ll go through the other steps, just for the sake of time, but easy to do that here.
133 00:12:38.740 ⇒ 00:12:44.619 Rebecca Bruggman: One thing you can also do that’s really helpful is, if there are certain fields that you know are confusing, like.
134 00:12:44.620 ⇒ 00:12:45.000 Nandika Jhunjhunwala: you don’t.
135 00:12:45.000 ⇒ 00:12:58.890 Rebecca Bruggman: a user to see them within this topic, you don’t want AI to have access to them, you can just click them off, and just say, I don’t want this to come in. Like, maybe I only want, maybe I don’t want some of these, like, hyperlinked ones. You can add them back in.
136 00:12:59.100 ⇒ 00:13:00.500 Rebecca Bruggman: So I’ll remove these.
137 00:13:01.130 ⇒ 00:13:03.849 Rebecca Bruggman: You can just unclick them, and then if we come over to the YAML…
138 00:13:04.540 ⇒ 00:13:14.509 Rebecca Bruggman: you’ll see that automatically removes those fields that I just clicked off for you, which is really nice. Yeah. So it just makes it really easy to, like, curate the topic.
139 00:13:14.530 ⇒ 00:13:25.360 Rebecca Bruggman: To say, like, this is the exact information that I want within here, and here’s removing stuff that I know is, like, confusing, or, like, will just, is old, or duplicative, or anything like that.
140 00:13:26.720 ⇒ 00:13:30.160 Caitlyn Vaughn: Okay, cool. This is… this is great. So, I guess…
141 00:13:30.160 ⇒ 00:13:31.060 Nandika Jhunjhunwala: Savi…
142 00:13:32.130 ⇒ 00:13:32.690 Caitlyn Vaughn: Yeah, go ahead.
143 00:13:32.690 ⇒ 00:13:33.320 Nandika Jhunjhunwala: Okay.
144 00:13:33.470 ⇒ 00:13:37.720 Nandika Jhunjhunwala: And then Blobby will not preference the hidden fields when querying the data.
145 00:13:39.160 ⇒ 00:13:39.970 Rebecca Bruggman: Exactly.
146 00:13:39.990 ⇒ 00:13:40.520 Nandika Jhunjhunwala: Okay.
147 00:13:41.250 ⇒ 00:13:46.109 Caitlyn Vaughn: So, have we been doing the modeling with dbt and not in Omni?
148 00:13:48.020 ⇒ 00:13:51.520 Uttam Kumaran: Yeah, I mean, all of our core modeling, which is, like.
149 00:13:51.620 ⇒ 00:13:57.080 Uttam Kumaran: past, case whens, all of that is in dbt, and then in Omni, it’s mainly…
150 00:13:57.310 ⇒ 00:14:07.730 Uttam Kumaran: creating the views, and then sort of, like, core relationships, or basically just, like, right now, it’s just solving for the core dashboards. So that’s just a…
151 00:14:07.990 ⇒ 00:14:10.350 Uttam Kumaran: That’s just sort of an architecture choice.
152 00:14:10.690 ⇒ 00:14:14.699 Uttam Kumaran: I think, you know, on the OmniDocs, I know you guys recommend
153 00:14:14.810 ⇒ 00:14:24.859 Uttam Kumaran: one or the other, and you have to push down and everything. So, for us, that’s our… that’s our typical way. We… we do all the core, you know, modeling in
154 00:14:25.100 ⇒ 00:14:34.699 Uttam Kumaran: in dbt, and then we make it available in Omni. Part of this is also so that when we… we have reverse ETL use cases, we can also send that data
155 00:14:34.920 ⇒ 00:14:45.059 Uttam Kumaran: you know, to other systems pretty easily. But I don’t know, Becca, if that’s, like… I know it’s kind of, like, a broader debate sort of topic all the time, so wondering what your thoughts are.
156 00:14:45.790 ⇒ 00:14:59.409 Rebecca Bruggman: You know, I think to… I think that’s a totally fair approach, just to say that. Like, I think doing a lot of the, like, heavy modeling, especially for things that are set, like, you know that you want them a certain way, like, having it at DBT completely makes sense for… especially from a performance perspective.
157 00:14:59.420 ⇒ 00:15:20.120 Rebecca Bruggman: I think the difference that I’ll just name is being able to add a lot more of, like, the semantics of, like, what does this data actually mean? How do I use it? Not just, like, what is, like, in the data, like, what is the model of the data, what is the view, is the kind of stuff you want to put in Omni. And I am seeing some of that within, like, if we come back to the topics of, like…
158 00:15:20.400 ⇒ 00:15:33.820 Rebecca Bruggman: the sort of, like, AI context within here, but I think there’s probably, like, even more that can be done around, like, curating what fields are showing up, adding more AI context to, individual fields, thinking about,
159 00:15:34.050 ⇒ 00:15:53.400 Rebecca Bruggman: topic curation, and how to kind of scope a topic so that you’re answering all the core questions that you need, knowing that you can pull, like, multiple topics into a single dashboard, and they don’t need to be that sort of one-to-one. So I think those are the kind of components in Omni that are sort of separate from dbt and, like, very naturally live in Omni that are good to invest in.
160 00:15:55.080 ⇒ 00:15:59.519 Caitlyn Vaughn: You can pull multiple topics into one… you said, dashboard.
161 00:15:59.520 ⇒ 00:16:18.430 Rebecca Bruggman: Heck yeah! Alright, let me show ya. So if you do another… oh, let me save changes, just because I made a little… a couple of changes on this topic. If you add a new… so we had our Thunder Salmon, which is our brand color, that’s a little Easter egg. If we add another tab, we have our little cherry tab.
162 00:16:18.430 ⇒ 00:16:30.140 Rebecca Bruggman: you can pull in anything you want. You can pull in a different topic, you can go into SQL mode, you can go back into browse all views. So, within a single dashboard, you can pull in, a variety of topics, or kind of anything that you’re wanting to do.
163 00:16:30.390 ⇒ 00:16:37.940 Caitlyn Vaughn: Hmm… I guess that makes sense. What I was trying to do the other day was join, like.
164 00:16:38.370 ⇒ 00:16:44.230 Caitlyn Vaughn: Let’s say, customer name to revenue, or whatever, which are in two separate topics, let’s say.
165 00:16:44.800 ⇒ 00:16:49.770 Caitlyn Vaughn: And put it into a single dashboard, which wouldn’t work. Or, sorry, into a single chart, which wouldn’t work.
166 00:16:49.980 ⇒ 00:16:56.359 Rebecca Bruggman: That’s where, like, making sure you have all of that within, like, a single, like, queryable dataset,
167 00:16:56.360 ⇒ 00:16:56.820 Caitlyn Vaughn: Yeah.
168 00:16:56.820 ⇒ 00:17:00.090 Rebecca Bruggman: Because that’s in, like, a single query versus, like, two different types.
169 00:17:00.860 ⇒ 00:17:01.979 Rebecca Bruggman: If that makes sense.
170 00:17:01.980 ⇒ 00:17:07.200 Caitlyn Vaughn: Yeah, it does make sense. Okay, so that’s just a matter of, like, setting up our data the right way.
171 00:17:07.200 ⇒ 00:17:08.999 Rebecca Bruggman: Yep, exactly. That’s bang on.
172 00:17:09.000 ⇒ 00:17:09.710 Caitlyn Vaughn: Okay.
173 00:17:10.599 ⇒ 00:17:12.220 Caitlyn Vaughn: Okay, yeah, that makes sense.
174 00:17:12.470 ⇒ 00:17:13.079 Rebecca Bruggman: Sweet.
175 00:17:13.369 ⇒ 00:17:30.819 Rebecca Bruggman: Coming back to the topic… okay, so we have our topic that we’ve been making, we added a join, we removed a couple fields just for fun, and then this Curate tab I really like, because it allows you to add labels, descriptions, AI content.
176 00:17:30.820 ⇒ 00:17:31.340 Nandika Jhunjhunwala: That’s kind of.
177 00:17:31.340 ⇒ 00:17:32.259 Rebecca Bruggman: Like, in bulk?
178 00:17:32.260 ⇒ 00:17:33.240 Caitlyn Vaughn: Hmm…
179 00:17:33.240 ⇒ 00:17:49.070 Rebecca Bruggman: Which is really nice, versus having to, like, go to each of the individual fields, either within the workbook or within, like, the view files or anything like that. It just makes it a lot easier to kind of, like, do it in bulk. And, you can also scope it down to just this topic or more broadly.
180 00:17:49.070 ⇒ 00:17:54.149 Rebecca Bruggman: So it just sort of depends on, like, where you want it to exist, like, what sort of, like, level of information.
181 00:17:54.700 ⇒ 00:18:00.200 Rebecca Bruggman: And then, I’m gonna skip the diagram, just because I often find it more confusing than helpful.
182 00:18:00.200 ⇒ 00:18:16.480 Rebecca Bruggman: And then this is sort of what gets generated. So, we did all this work in the UI. This is what, if we were to push this up to the shared model, so coming back over here, this is basically what is getting generated for us based on all the actions we just took.
183 00:18:16.480 ⇒ 00:18:19.709 Rebecca Bruggman: So, how to interpret this is, this is the base model we picked.
184 00:18:20.340 ⇒ 00:18:24.090 Rebecca Bruggman: This is the label we set for it, so, like, what is the name of this topic?
185 00:18:24.330 ⇒ 00:18:30.780 Rebecca Bruggman: This is the folder exist under? This is the description we gave it. These are the fields we removed.
186 00:18:30.940 ⇒ 00:18:41.360 Rebecca Bruggman: That we didn’t want to have included, and then this was the other table that we joined in. Now, this just says joins, and then, to this,
187 00:18:41.480 ⇒ 00:18:46.610 Rebecca Bruggman: Table, because within the relationships file, that’s where we’re defining how the two tables look.
188 00:18:46.610 ⇒ 00:18:47.110 Nandika Jhunjhunwala: No.
189 00:18:47.110 ⇒ 00:19:00.050 Rebecca Bruggman: how they relate to each other, and so within the topic, we’re just telling Omni, like, hey, you already know how these two tables relate to each other, I’m just telling you I want these two tables within this topic, but you already know how they relate to each other.
190 00:19:02.400 ⇒ 00:19:05.299 Rebecca Bruggman: Cool. Any questions on this, on these pieces?
191 00:19:05.910 ⇒ 00:19:13.090 Nandika Jhunjhunwala: Yeah, when you’re writing out the descriptions for each of the fields, does that get surfaced
192 00:19:13.260 ⇒ 00:19:23.190 Nandika Jhunjhunwala: In the UI elsewhere, when, like, someone’s querying the data or looking at the data, or does that only show up, like, when we’re, like, creating that topic?
193 00:19:23.820 ⇒ 00:19:29.539 Rebecca Bruggman: It will show up, both for, so you see all these descriptions.
194 00:19:30.190 ⇒ 00:19:33.199 Rebecca Bruggman: I basically use them for, like, hovering over the field.
195 00:19:33.200 ⇒ 00:19:33.800 Nandika Jhunjhunwala: Got it.
196 00:19:33.800 ⇒ 00:19:46.690 Rebecca Bruggman: So if, like, folks want to say, like, oh, what does this actually, like, mean, or, like, something like that, it can be helpful to kind of just have a little bit more of that guidance. And the other benefit is this also goes into the AI context window.
197 00:19:46.690 ⇒ 00:19:59.230 Rebecca Bruggman: So, like, anything where you’re adding additional, like, I’ll say, like, business-level context, whether it’s description, AI context, label, anything, that all is beneficial to the AI context window.
198 00:20:01.130 ⇒ 00:20:02.020 Nandika Jhunjhunwala: Makes sense.
199 00:20:02.210 ⇒ 00:20:02.750 Rebecca Bruggman: Sweet.
200 00:20:02.750 ⇒ 00:20:06.299 Nandika Jhunjhunwala: The other question, I have is around…
201 00:20:06.420 ⇒ 00:20:09.790 Nandika Jhunjhunwala: what’s, like, best practices to configure a topic? I think…
202 00:20:09.930 ⇒ 00:20:29.160 Nandika Jhunjhunwala: maybe this goes back to, like, our discussion about modeling in the dbt layer versus, like, semantics and Omni. But we have a couple topics that don’t have any, like, joins, or any sort of, like, configuration. Like, some of the topics are just plain pulling from one base view, and that’s about it.
203 00:20:29.560 ⇒ 00:20:38.810 Nandika Jhunjhunwala: And, like, I think Caitlin and I saw, like, some room for consolidation there, like, with joining other tables and so on, but would love to get your take on, like.
204 00:20:39.110 ⇒ 00:20:49.600 Nandika Jhunjhunwala: what’s the best way to configure a topic? Like, how many joins… maybe there’s… there isn’t one answer, but, like, how many joins should there be on a base view? And stuff like that.
205 00:20:50.770 ⇒ 00:20:54.659 Rebecca Bruggman: Yeah, I mean, it… very candidly, it’s an… it depends, because…
206 00:20:54.660 ⇒ 00:20:55.150 Nandika Jhunjhunwala: Yeah.
207 00:20:55.150 ⇒ 00:21:09.459 Rebecca Bruggman: it kind of goes back to, like, are you doing those, like, big wide tables in dbt, where you’re basically doing all the pre-aggregation, and then you are just basically having, like, a single table that comes to a topic, or are you making them,
208 00:21:09.460 ⇒ 00:21:19.910 Rebecca Bruggman: more queryable and more flexible by making them more, like, standalone, and then you would be doing those joins and having them come to a topic. I’ve seen both.
209 00:21:19.910 ⇒ 00:21:35.119 Rebecca Bruggman: you know, in the time that I’ve been, you know, like, running a lot of, like, customer implementation, so it really just sort of depends on, like, how you want to approach. I will say, if you… the benefit of having, you know, not those sort of just, like, big, wide, pre-aggregated tables is that you just
210 00:21:35.120 ⇒ 00:21:41.889 Rebecca Bruggman: to have more flexibility. And you can still do things in Omni, like aggregate awareness, where you can get some of that aggregation.
211 00:21:41.890 ⇒ 00:21:42.390 Nandika Jhunjhunwala: Like, weird.
212 00:21:42.480 ⇒ 00:21:57.930 Rebecca Bruggman: If, like, you’re looking to do that, but you just have a little bit more flexibility. But I will say there isn’t, like, a set number of joins that you should have, and I think there’s the other side of things where you want to be mindful. This is why I kind of come back to the, like, what questions are you answering with this topic?
213 00:21:57.950 ⇒ 00:22:07.509 Rebecca Bruggman: Because you also don’t want to have this sort of, like, mega topic to answer every question, because then you almost have, like, too much data all in one place.
214 00:22:08.030 ⇒ 00:22:28.169 Rebecca Bruggman: Especially when you can, you know, sort of pull different topics into a single dashboard. And just having that little bit more curation for, like, AI to know, like, for opportunities, I go to this one. For, pipeline, I go to this one. For how, like, my, customer health is doing, I go to this one, versus, like, having that all crammed into one thing.
215 00:22:30.100 ⇒ 00:22:36.020 Caitlyn Vaughn: Yeah, I think that makes sense. I think what we’re doing right now is the pre-aggregation tables.
216 00:22:36.970 ⇒ 00:22:46.460 Caitlyn Vaughn: Which is causing us to not have as much flexibility in, like, if we want to change, instead of having, like, measures, you know, we can’t, like, change…
217 00:22:46.580 ⇒ 00:22:56.669 Caitlyn Vaughn: time, or measurements, or things like that, which is kind of tough, and then it’s also kind of hard to see what is inside of everything. Yep. Yep.
218 00:22:57.350 ⇒ 00:23:00.430 Caitlyn Vaughn: So maybe we’ll think about, like, unpacking those.
219 00:23:00.880 ⇒ 00:23:03.530 Rebecca Bruggman: Okay, cool. Yeah, I think it’s great food, food for thought.
220 00:23:04.570 ⇒ 00:23:09.180 Uttam Kumaran: Yeah, I think bringing a lot of this from dbt into the OmniLayer as measures.
221 00:23:09.210 ⇒ 00:23:26.219 Uttam Kumaran: we have a lot of the intermediate models, like, cleaned up for opportunities for all the core objects, and you just, like, rotate them into Omni. Kind of, like, almost… you can kind of just replicate it as relations between the int models, and almost, like, create the prod marts in Omni.
222 00:23:26.430 ⇒ 00:23:35.740 Uttam Kumaran: Yeah, I mean, we have a lot of stuff already modeled ready, it’s just we’ve been focusing just on creating the marts needed for the dashboards and iterating on that. So I think, like.
223 00:23:35.920 ⇒ 00:23:41.310 Uttam Kumaran: cleanly just shifting Marts, or some of it, into Omni could be…
224 00:23:41.570 ⇒ 00:23:43.880 Uttam Kumaran: you know, helpful, because then you can quickly go into Omni.
225 00:23:44.330 ⇒ 00:23:45.169 Uttam Kumaran: And then edit.
226 00:23:45.400 ⇒ 00:23:47.599 Uttam Kumaran: You know, a specific metric definition.
227 00:23:48.710 ⇒ 00:23:51.080 Uttam Kumaran: Versus going into dbt and doing that.
228 00:23:51.400 ⇒ 00:24:08.239 Caitlyn Vaughn: Yeah. Yeah, and I imagine, like, it’s just Nadica and I obviously taking this over, and, like, we’re both half working on this, and then half working on, you know, lots of other things, so if we could do this in Omni, it would be a lot easier for us versus, like, going back into dbt and having to, like, remodel everything.
229 00:24:08.560 ⇒ 00:24:16.480 Rebecca Bruggman: Exactly. Yeah. Because that is… that is the thing, like, dbt is, like, its own… its own being, so I’m very cognizant of that. Yeah.
230 00:24:16.480 ⇒ 00:24:29.500 Rebecca Bruggman: that’s the other thing of, like, having everything living in DBT. It’s like, if you’re super comfortable with it, then it’s great, but if it’s something where, it just takes some, like, additional, like,
231 00:24:29.500 ⇒ 00:24:41.100 Rebecca Bruggman: building and, like, maintenance and overhead to, like, if you’re doing everything in dbt. And there’s just a lot more, like, UI flexibility within Omni. So, yeah, it just depends on how y’all want to maintain it.
232 00:24:41.470 ⇒ 00:24:47.249 Caitlyn Vaughn: Okay, cool. Yeah, that makes sense. I think we are aligned, and that’s very reassuring.
233 00:24:47.250 ⇒ 00:24:57.290 Rebecca Bruggman: Okay, great, love that. Sweet. Any final questions? Was this helpful? Anything that’s sort of, like, still lingering for y’all?
234 00:24:57.670 ⇒ 00:25:10.689 Nandika Jhunjhunwala: Yeah, that was so helpful, thank you so much. The other question, I guess I just don’t know how to do this, but, like, how do you define measures in Omni? If there’s a quick walkthrough, if you have time, that would be great, otherwise all good.
235 00:25:10.940 ⇒ 00:25:16.800 Rebecca Bruggman: There you go. Cool, let me do… okay, so we have… I’ll just say, like, using query…
236 00:25:16.840 ⇒ 00:25:34.510 Rebecca Bruggman: Cool. Just to show, once we build the topic, you can then use it in query. It’ll stay here, I’ll save this, and then y’all can come into the branch if you want to utilize it. We’ll come into our little cherry query, let’s go into… Actually, we can just go into the topic we just built.
237 00:25:35.260 ⇒ 00:25:37.470 Rebecca Bruggman: Let’s see…
238 00:25:38.390 ⇒ 00:25:45.019 Rebecca Bruggman: If we wanted to do days to renewal, so right now we just have this count, this is the green, or the measure, so they’re just aggregations, that’s how you can think about it.
239 00:25:45.020 ⇒ 00:25:45.550 Nandika Jhunjhunwala: Yep.
240 00:25:46.010 ⇒ 00:26:02.040 Rebecca Bruggman: So if we wanted to come in, like, days to renewal, we could click on the triple dots, modeling, oh, sorry, aggregates, and then there’s all days that are just right out of the box for you. So if you wanted to say, I want the average of days to renewal.
241 00:26:03.110 ⇒ 00:26:04.989 Rebecca Bruggman: They’ll just, like, come straight in there for you.
242 00:26:05.200 ⇒ 00:26:12.790 Rebecca Bruggman: Now if you’re doing this, then you’ll see it’s about… it’s 96, which is… and then you can use that from there, so if you wanted to say… let’s see…
243 00:26:13.250 ⇒ 00:26:17.019 Rebecca Bruggman: Average days to renewal, and then maybe, like, industry?
244 00:26:19.630 ⇒ 00:26:20.440 Nandika Jhunjhunwala: Hmm.
245 00:26:21.640 ⇒ 00:26:23.040 Rebecca Bruggman: So you could do something like that.
246 00:26:24.880 ⇒ 00:26:46.960 Rebecca Bruggman: And then I will say, again, we’re doing all this in the workbook, because the UI goodness I really, enjoy. You’ll see the pancake stack is continuing to increment, so you’ll see, like, this topic is here, this new relationship that we built is here. This stays to renewal average is here. I’m going to add this to the branch, which will not push it to your shared model, but it will push it to the shared model
247 00:26:46.960 ⇒ 00:26:52.840 Rebecca Bruggman: within the context of the branch only. So if I came… I’ll just sort of walk you through this, so I’ll say, add to branch.
248 00:26:53.450 ⇒ 00:26:56.839 Rebecca Bruggman: This takes it out of the workbook, into the branch.
249 00:26:56.840 ⇒ 00:26:57.350 Nandika Jhunjhunwala: And then.
250 00:26:57.350 ⇒ 00:27:03.689 Rebecca Bruggman: If I go back into… Our shared model, again, within my branch, I’ll just refresh.
251 00:27:05.350 ⇒ 00:27:08.150 Rebecca Bruggman: We’ll show staged.
252 00:27:08.900 ⇒ 00:27:19.959 Rebecca Bruggman: This will show that topic within shared. Again, this is only within the branch, this is not actually pushed to your shared model. And then within here, this will also show that measure we just made.
253 00:27:22.780 ⇒ 00:27:26.680 Rebecca Bruggman: But if I exit this… All that goes away.
254 00:27:29.470 ⇒ 00:27:37.019 Rebecca Bruggman: So it’s a nice way to sort of, like, containerize things, and then also, if you’re, like, making, like, for example, making modifications to, like, a current topic.
255 00:27:37.020 ⇒ 00:27:51.379 Rebecca Bruggman: doing it in a branch can be really helpful, because then you can be like, okay, for this topic, I know these dashboards are built on it, let me make sure I’m not, like, blowing anything… blowing anything, making these changes. I’ll put it in the branch, we’ll kind of go around and check, and then I can merge it from there.
256 00:27:51.580 ⇒ 00:27:58.989 Caitlyn Vaughn: Okay, awesome. Yeah, this is actually really helpful, and the measures are really easy. It’s just, like, it’s out of the box, basically.
257 00:27:58.990 ⇒ 00:28:01.280 Rebecca Bruggman: Yeah, it’s so easy within the UI, which is great.
258 00:28:01.280 ⇒ 00:28:02.310 Caitlyn Vaughn: Amazing.
259 00:28:02.860 ⇒ 00:28:03.490 Rebecca Bruggman: Sweet!
260 00:28:03.880 ⇒ 00:28:14.420 Rebecca Bruggman: All right, y’all, well, thank you for making the time on a Friday. I really appreciate it. And, you know, happy building and querying and building out your topics. I’m looking forward to seeing what y’all, work toward.
261 00:28:14.750 ⇒ 00:28:15.510 Caitlyn Vaughn: Thank you so much.
262 00:28:15.510 ⇒ 00:28:17.779 Nandika Jhunjhunwala: Thank you so much. Yeah, it was so great.
263 00:28:17.780 ⇒ 00:28:19.260 Uttam Kumaran: Rebecca, appreciate it.
264 00:28:19.260 ⇒ 00:28:21.169 Rebecca Bruggman: You’re so welcome. Alright, happy Friday. Bye, y’all.
265 00:28:21.170 ⇒ 00:28:21.630 Caitlyn Vaughn: Bye.
266 00:28:21.630 ⇒ 00:28:22.199 Nandika Jhunjhunwala: Have you read it.