Katherine Bayless: How are you? Uttam Kumaran: Good! How’s everything? Katherine Bayless: Good. I just, finished up an interview for the data engineering, candidate, or, interviewing a candidate for the data engineering role. There we go, words. Uttam Kumaran: How’d it go? Katherine Bayless: really good, , I liked him. I, I feel my interview style is, , guerrilla warfare. , people come in and they’re , let me tell you about my strengths and weaknesses, and I’m , , I got, , this thing on my mind, what do you think, ? Uttam Kumaran: , I’m the same… I’m very similar. I’m , I don’t think I’m… I’m not a great interviewer, I’m pretty optimistic about people, and I’m , I see the potential, . Katherine Bayless: ? And then… Uttam Kumaran: I’m usually, , the closer, or the seller, , I’m not the person in the middle, but I also, , I’m , , that’s all nice, but I’m , , I’m thinking… I just got out of this meeting where I’m, , thinking about this, . Katherine Bayless: What do you think? Uttam Kumaran: And… one, for the best people, they… that sells them, because… I’m , , this is how, , life is gonna be if you’re here. And then for the people that don’t… are very prepared, it’s, , really brutal, which is also good, , a lot of our work is unprepared, walking into a situation, … Katherine Bayless: . , and he definitely was in the former camp, where it was , you could tell I caught him a little off guard, but he’s , , , , I have a lot of questions for you. I’m , yep. Uttam Kumaran: Great, great, great. How’s the week going far? Katherine Bayless: Hmm, I will be honest, bit of a rocky, rocky start to the week. , I’ll… I’ll hang in there. . It’s just, , it’s not data-related necessarily, it’s just, , return to office stuff, and, , you’re just , oh my god, , do we really have to keep having these fights? , can we not let it. , choices, and just will, , get out of everybody’s business, but, . Uttam Kumaran: What’s your, what’s the, what’s your TLDR on your perspective? Katherine Bayless: oh, I feel if you don’t trust your people to be getting their job done, that’s your fault, not theirs, ? . , , if you don’t. I’m working because my butt’s not in a chair, then, , that’s a problem I can’t solve. Uttam Kumaran: , it’s hard, , for even… I’ve just worked in remote for a long time, , I was in person in New York, and then, after COVID, stayed remote. if you already have a great team, it would just make it more fun to be in person. But, , I don’t want… I want to be remote. , by virtue of, , eye for an eye in the best sense, everybody can be remote, because I don’t really care. I would there to be some people here in Austin, , we’re trying to hire in Austin, and that would be amazing, but it’s not a requirement at all. . Katherine Bayless: thing, too, is , I , I really do love being in person with people, but I feel if those people are mandated to be in person with me, we don’t have any fun at all, and I don’t that. But , . , a cool team hanging out and doing cool stuff, sounds awesome. be , everybody must be in Monday through Wednesday, or else we’ll kill you, is not fun. Uttam Kumaran: , I wonder… , and it’s just, , it’s also such a drain on people’s lives to, , mute in and… I don’t know, it’s a hard ask. Katherine Bayless: But we can talk about things that are under my sphere of influence, instead of battles I can’t win. Uttam Kumaran: , , , , I wanted to … maybe we can spend some time today, and really my… this week for me is, Sam’s, , planning out the executions for Snowflake setup. I wanted for him to just run the plan by you first before we run a bunch of stuff, and also that, , you see how we think about setting up stuffing, he’s working on that. I went ahead and, , we worked on this Gantt chart, and probably today can just be mainly focused on on that, I’ll… I’ll have it open on my side. Let me just share that, and then, , we just, , have a conversation about How to tackle the next, , few weeks. Let me just take… I’m just gonna also take notes. Katherine Bayless: My brain lives mostly on sticky notes. I had written this one down to mention to you at some point. It’s mostly a tangent, but it was a, fold this one into the back of your brain. when it comes to, , data and, , security and protecting and all the things, , we definitely have PII, , we will be more sensitive about. We also have embargoed data in some cases, , , award winners we just can’t tell people about yet, but staff can know, and, , I know it’s a totally normal use case, I’ve just never had to deal with embargoed versus just, , protect. Uttam Kumaran: , , , I haven’t worked for, , the Oscars, or… ! It’s probably the most fun use case, there’s probably some more… not—fun use cases, but , , , I get it. Especially sitting next to the Pentagon, … Katherine Bayless: But , embargoed versus sensitive. I was , huh. Uttam Kumaran: , great. , we could definitely talk about, , different security postures for that. , great, I’m just gonna… I just have a bunch of stuff open, I’m just gonna leave this there. … , I’ll just give the overall walkthrough, and then I have our Slack. up as that I could read through. But , I was … Overall, just… parts for us is just, , looking through all the sources, understanding, and then starting to make a couple of, , definitions. , what are the core KPIs? , how do we… what is a member versus, , an attendee? , all the different core objects, … that’s what I to work on. The second piece that we did get done last week, I’ll just note this down, is, . getting access to… to Snowflake. And then, it’s … we get into… We’ll talk through today, , what is this phase, which is, , defining and mapping, but in terms of storage, this is, , gonna be what Sam’s working on, which is setting up world-based asset controls, setting up the core warehouses, users, setting up service accounts. This is gonna just set the stage for Anything that touches into Snowflake has a way to get access to things. We have the core… flexible objects to be able to assign roles, to land things, just, , the general setup is there. After that is when… We will start to… talked through ingestion. And , this is where, even today, we can get into, , what are the P0, P1, data sources, we then can use that to go and evaluate, , ingestion tools or building things on our own. Once we make a decision on, , the strategy around ingestion is when we can kick off ingestion. And that , hopefully, that kicks off. Christmas-ish, that if things need to run for a while. they can run during that time frame, and then we have some step that, , we confirm that, , data has landed. . I put this out here because we haven’t listed all the sources that are landing. The next piece we’re going into is as soon as we decide on the ETL tool. Honestly, probably, , during that process, if we have parallel path, we can start to set up dbt Core, GitHub, we can also make some decisions on that, and I can start working with Jay to make that, , , we have access to our… or he’s in the loop on these infra decisions. Dbt Core is free, we can rock with that. And then consider… consider cloud. Again, , we can talk about… I want to talk about your notes on the procurement strategy, too, … Then, , as soon as we start to model things, we’ll move into… building out… building out marts when… when things are down… are landed. once we understand the schema, we can start to say, , , what are the ultimate marts that we’re driving towards? And then… ideally, we parallel a path how does data get… how do those marks get accessed, both writing queries and Snowflake, or if there needs to be… we need to consider, , reverse ETL, and then as as, , a BI tool. And then the BI tool is , , we… We have some amount of making a decision there, setting up the tool, driving towards some reports, and then training folks. Katherine Bayless: In parallel across everything. Uttam Kumaran: Let me just move this guy up… In parallel across everything, we also have, , these, , cross-cutting tasks, where we’re gonna put together, , our metric dictionary, we’re gonna be maintaining a data platform, , visual architecture. as we… as we drive, at any moment, you’ll have, , a visual of, , what is the system. Of course, we’ll have more information noted down on, , any of the contracts, , . the contract, the terms, trying to centralize all of that. Ideally, if we can… if… Since you’re technical, we can centralize all of that in the repo, that’s probably perfect, , versus… Creating more spreadsheets that, ideally, we would just put that in docs there. We would also start to have documentation on, , ETL, , , how often are jobs running, on the data warehouse, and then After we build models, we can start to document models. this is, , the zero to one. what, , I’m very curious about is, . What parts of this are… more important than others. Of course, this is 0 to 1 on, , both the infrastructure, but also I’ve listed that we will model certain things. If you’re more , hey, let’s just spend this phase landing everything. and then… and that’s what we focus on, then we can just prioritize that. If you’re , , let’s just land the two things that we need to drive this question I have, then we can do that. this is where I … We’ll need you as a guide to say, , where to… where to parallel a path, and, , what to drive towards. But, . many thoughts. . Katherine Bayless: , … we can focus more on landing than modeling, but landing in a… Slightly intelligent prioritization. The reason I say that is because I’m . the modeling is going to be where we’re gonna run into, , , the upper limit on people’s existing data literacy, , , pretty quickly. There’s definitely some baseline stuff that will be modelable, but the business rules are gonna be a real challenge to truly surface. , in an odd way, , they’re everywhere, and therefore they’re nowhere. . , we… , , the chatbot thing that I sent you. we had this, , hour-long call with the guy, at one point early on, where he was , I’ll ingest this source, this source, this source, , blah blah blah blah, ? Train the model, etc. And we were , oh… but you need to understand all of the business rules to put that data. together, and he’s , , just send them to me. And we were . Uttam Kumaran: That’s cute. , you’re … Katherine Bayless: , , that AI thing is getting trained off our public website, because we were , what, that’s gonna be the easiest way to get this , is , if we put it on the website, it can go into the AI thing, because otherwise, going back to the embargoed data, we were , if somebody’s paid us, , millions of dollars and told us to keep it a secret, and then the AI. Uttam Kumaran: , you’re just… Katherine Bayless: landing, we can accomplish more, more quickly than we will with modeling. . Is where my head is at. The other thought, which is the more optimistic and less pessimistic flavor of it, is. there’s a lot more citizen data engineers running around than I was realizing. I had this conversation with this guy on Friday, who’s in our market research team. He spent 10 years at, , a market research firm working a multi-million dollar contract for Verizon, and he’s , what are you guys doing here? And I’m . help. And , , there’s totally teams where, , if we could just land it, and hand it over, people… to do with it… , long way of saying, … Uttam Kumaran: , great, , that… Katherine Bayless: We’ll bear more fruit. Uttam Kumaran: , , if we can land the data, and then we identify… , , the… Citizen, DEs. and enable them… , And, potentially, gives… them, home to model. In dbt, , that is a really great way of doing things. Do you have, , the… these folks, , a couple of them top of mind? Because then, ideally, , these are the folks who we would meet, go become friends with, and then they would drive our initial , roadmap. Or at least I could tell you what they’re… I could go find out what they want. Katherine Bayless: , , , Chris Deathloff, is one, and then there’s, there are 3 people who are all named Anna, Anna P, Anna K, and Anna R. . They have… they are not technical in the sense of, , they’re… they’re great with Excel, but that’s as far as they’ve gone. However… they really do understand the data, and they would be able, , to help with the modeling if we gave them a little bit of an understanding of, , , here’s the techie piece of it, now… . And then there’s another, person, Quinn, and then JC. , even Tom, probably. There’s, , there’s really… there’s a lot of latent talent running around. , cool. . , interestingly, then, , , if that’s , . recruiting some assistants in the trenches, knowing that also leadership wants to see things, ? And it’s less likely that I’m gonna find a VP that’s suddenly gonna put hands on a keyboard, which is fair. … The upward trajectory for the data is gonna be, , we need the machinery to help us stop doing things that don’t really make us as much money as they should, or could, or… cost, frankly. it’s the. Uttam Kumaran: , should maybe go into that a little bit deeper at that point. Katherine Bayless: , for example, we have the Investor Partnership Program this year, where we are charging 60,000 in staff time putting this program together. what I don’t want to do is make it sound every idea is a bad idea, because, , the investor thing is definitely something we should explore generally, but, , this particular program probably was not the winner. ? We have another one that’s, , an attendee match thing, similar. There’s these, , show floor tours that bring in a little bit of money, but take a lot of manual effort. We just don’t have great, , decision-making architecture around, , what are we getting out of this program? There’s one question that’s been directly asked by leadership, which was, did we panic too early and move up a bunch of our paid marketing into this phase where CES was still, , free registration for everybody, versus waiting to spend more money on marketing after CES. and cost money, and we don’t have a model to do that. And also, when I said we should focus on profit for attendee, they looked at me I was from Mars. , , some of it is the tools and techniques, and some of it is the ideas and the ways of thinking. Uttam Kumaran: , there’s definitely something around, , what is the service line profitability? , this is very similar in… , in a lot of different places, they just do product-level… you want some product-level insights, what products are we selling the most of? How much are they bringing in? And that alone helped dictate , , what products to lean in more, or what to lean out of. Second is, , . , marketing effectiveness. This you can’t do in isolation until you also look at what are the outcomes for those folks? , are the folks coming in free? Are they end up purchasing ancillary services? Or are you finding that the free… the people that pay end up being the people that pay again and again? And , there’s something around The free versus paid attendee. Katherine Bayless: , it was … it was earlier over the summer, there was a conversation, because if you go to CES, but the next three years, you will get invited for free as an alumni. And if you continue coming, obviously, that’s just perpetual. And there’s talk of switching it to be, , you have to come every 2 years instead of three. And Kyle had done some back-of-the-envelope numbers on that, but, , those were the first time anybody had even asked that question, ? . , . Uttam Kumaran: This is, , do you consider this very similar to, , , just, , a rewards program or, , customer loyalty program, . And also, , talk to me about Kyle, , how should we best work with, Kyle on things? Katherine Bayless: It’s a really great question. , he’s really smart, smarter than he realizes, . , he hooked up, , an ETL process using R last week, and I was , I didn’t know you could do that, I’m gonna be honest. I really did not. I’m a Python girl, what can I say? . , , he’s really smart. He’s also, , very… fast, ? , he is responsive, he’s restless, he’s impatient. , he does my work faster than I can even read the email asking me to do something. , , , delightful. Just delightful combination of intelligence and restlessness. now, , where I’m, , as a person who likes to grow the people she works with, I’m , hmm, I feel I need to get you to slow down and strategize. ? , because, , he is interested in helping, he might spend all of his time helping and not enough time learning, ? And , it might be interesting to figure out how do we really pull him into stuff where he feels he’s not slowing things down because he’s learning, which I would. someone to think, but I do think he might fall into that trap, at least initially. And, , , , , , . he’s gonna race to put out fires if we don’t give him things to do. Uttam Kumaran: , I, this is perfect. , I… whenever we come in, I just had this conversation with another client, is , who are the people that you want to make, , heroes in your org? And what we’ll teach them is to start Doing what you’re doing, which is asking the question. Katherine Bayless: Alright, how do we get him… Uttam Kumaran: To not just take li… take… asks and execute, but to start saying, what do you… what does the business need? How can I present that to Catherine on a weekly basis? answers. , my next point here is , , what are the rituals that you’re running in a data team, and, , how can we, , support that? Because we’re gonna start running rituals, for CTA. , what we found effective in the past, especially if you had these If you have folks that are, . straggling data folks that you want to be… sit next to us as we start to execute these, and they can learn, and we can start start to give them a lot of the stuff that’s, , as core as we can, and they start to level up. , give me a sense of, , what the… what the data team processes look now, and, , , , what your vision for these, , daily, weekly, monthly, , rituals are. Katherine Bayless: , , to be perfectly frank, at the moment, there is essentially nothing, except me occasionally saying, hey Kyle, we should chat again, but this is not my normal behavior, because he was coming over from the market research team and was , … they’re still trying to find the backfill? I can’t remember if they’ve got a finalist yet or not. he’s a one-foot each world, which I do… I don’t know why, but that is where, , everything folds apart for me, and I’m . We need to establish ceremonies. rituals. What I have done in the past is, , just a constant back and forth between, , over-engineering the perfect system, and then being too exhausted to use it, and being too light-touch, and now I need to rally the troops, but I don’t have a way to get them easily. But I usually am, , stand-up, weekly planning. Retros are something… retros slash, I always call it show and tell, demo, retro slash show and tell. , , that’s one that I really want to get better at building the muscle for, and it’s going to be critical here, because, , at my old place, it was a very technical staff, and demos would always die on the vine, because you’re , . ? But here, , people would be , that’s cool, , tell me more. Uttam Kumaran: And … and then give me a sense of, . And this may be a leading question, but, , your capacity… because for us, , we can… , I would love to help you establish these rituals, but I also don’t want to ha- don’t want to create more meetings for you to join, and … give me a sense of, . where you, , for example, if a good solution is , hey, we’re gonna be meeting about CTA on a daily basis, maybe we just have Kyle join that meeting every other day. And maybe just the three of us meet. The other alternative could be, hey, maybe we present to Catherine, , our roadmap of what we’re planning, and Kyle just joins us throughout the week. And then, as we meet these folks. we see who’s also interested in joining us, and, , this is where it would be helpful to know, . In some other… in some… in some other clients, they’re , oh, , but that person’s not in my org, or , I don’t want to waste their time. Other places, they’re , oh, , that person can totally… if we enable them, they’ll service… and they would love to just be talking to other data people. , we’re happy to establish that, but what I don’t want to do is have you in just, , another… 5 or 6 more meetings, where, , I want to… I want to teach them what it’s to have… to, , … execute and deliver for, , a stakeholder, , an end-to-end question, or an analysis, or… and… versus… , part of not giving them a lot of your time is to show them that, , hey, we… Catherine already mentioned the priorities, it’s now up to us to ideate, , maybe there’s… it’s a modeling thing, we need to land the data, , then let’s come back next week, let’s codify our learnings, and then present. , that’s the stuff I’d to … Impart to them, ? Katherine Bayless: , that’s funny, … , , , , , , , , , , . It is… honestly, it’s funny, it is the biggest lesson that I learned in my, , early years transitioning to leadership was, , the more I show up, the less people will walk on their own. And if everybody’s just doing what they think is, , normal. But , I have learned that neglect is the most effective technique. Uttam Kumaran: Calculated, calculated neglect. Katherine Bayless: , , , , ? But, and honestly, it is part of, , what has really been cool to watch with Kyle. , I used the one foot in each world as a bit of an excuse to a certain extent, but, , he’s fucking flying, ? I feel he has to, , sit down next to me and look at a. morning, ? And , , … I… . to everything you said… . Uttam Kumaran: that’s what I’d to help at least do that, and… and we can establish a cadence by which we’re meeting as a team, and I don’t think… we’re still gonna… , this is where I’m still gonna get, , Snowflake and something set up. Katherine Bayless: And… but I want him to really, as we start to land data and do… Uttam Kumaran: go source by source to understand it, , if he can join and be next to us, that would be great. And then I want to get him and anybody else to the point where if they to join those meetings, they can join, or if it’s, , Kyle and us are supporting one of these people because he has the… Background from the market research team, but, , we’re… then our team’s more of, . does Kyle have enough data, and is he trained to run stuff in dbt and in SQL? And then is he poison enough to go present to the org, or tag along with you to meetings and present, and, , take that off of your plate? , that would be great. Katherine Bayless: , . And , , interestingly, , a funny specific side note about Kyle, the market research team does a lot of, , presentations at CES, , that clients , , essentially… Great. ? He hates it. hates it. He’s , I… it’s, , my least favorite part of my job, I cannot stand it. he might be decent at it, I really don’t know. But it’s because certain folks in leadership have, on occasion, popped up to correct him while talking. And I’m , oh my god, talk about instant confidence drain. , I would love. build his confidence as a presenter, but there’s some trauma there at the moment. Perfect. Uttam Kumaran: , , I would … , this is where it’s, , you, you, You bring people to situations where it’s you safely can mess up. ? And we’ll start by just, , working with us, we’ll just show how we would present as an insights team, and then start to have him take ownership of that. that’s great. I’m glad that we at least we have a Kyle, , that we’re thinking about here. Katherine Bayless: I really, , I’m very lucky to have him, for . . The other thing I’ll say is, the organization, generally speaking, because it is , . , , it’s this, , human blockchain of information, ? , working out loud seems to go over pretty here, and , , I don’t know if it makes sense, but, , a… , a Slack channel that’s, , open, not necessarily putting everybody in it. by default, but, , , putting some of these folks that will want to be in it, and then saying, , , anybody’s welcome to drop by, ? , , that working out loud thing would land really here. Uttam Kumaran: , great. , there’s some… there’s some… That’s just , again, , for me to learn a little bit of the culture, but if people are genuinely, … and this is… we have this in our company, and a lot of the companies we join, we try to do this, which is just, . engineers are always gonna be , oh, I wanna… it’s not perfect, and I’m big on, , just ship it in any state and get feedback from anybody you can get that is going to give you feedback. Because you’re just gonna toil, and you’re gonna go down something where you should’ve got feedback 5 steps in. if we can do some type of open data channel, and , , maybe it is something where, … and this is also where I’m interested in how you think about , getting more buy-in from leadership. , one thing that we’ve done often is we’ll be , hey, . we want to make Catherine has a deck of, , all of our accomplishments for this month, that when you go present, you’re , here are some of the wins from the DE team, but we can even, … if there isn’t an established, , cadence of, , a monthly business review, and that’s something that you want to start to do. I don’t know if we’ll be able to necessarily do that. maybe towards the end of next month or early January, but if that’s something that you’re , , that… that keeps… that starts to build, , this proactive cadence in a meeting that you own. ? And it’s less of, , just call Captain when there’s a fire, but you set the tone, you come to that really fair, and it’s just purely about data, it’s not about, , other activities. That could be also good, and then we would just support Wherever needed for that. Katherine Bayless: , , , , let’s see, many, thoughts, that… , , one tiny thing. , culturally now, , , everybody is freaked out, and… Uttam Kumaran: Oh , for January, , , . Katherine Bayless: . , unlikely to get a whole lot of attention, which is good. . But I was wanting to suggest this as a, , , a thing that we would be, , launching as part of, , , , this is the end of the year, and in January, we’re gonna, , start socializing this more. I don’t think I can get away with going as hard as, , OKRs out of the gate, but. I want to be a team that has visible public metrics, . LAs, and I want to report against their delivery against them. I want to… I have a number somewhere, I’ve parked it, of the, . 300 and some manual processes the marketing data team was running before I started. , I want to show the amount of just bullshit we are stopping, ? . I… and I want… I want to have it be some , , KPI, OKR type framework, because I really… eventually, I would to move the needle on the way this organization does its annual goal setting, because now. That’s just a to-do list. It’s just a to-do list. And I’m , I want to start focusing them on outcomes. Uttam Kumaran: , perfect. , and… and this is, , a perfect way to leverage us as, , the backbone, because you’re… what we… what we’re not going to be able to do is… Build the relationships and build the buy-in, but we’ll totally make you’re armed, , with… With those wins, and of course, , help accomplish those, and then set you up to really, , nail those meetings, and get whatever you need, budget, resources, more time with executives, , more decision-making power. , , , great. Perfect, this is great. Katherine Bayless: To see, … , , 100%. Uttam Kumaran: . , and that’s the only way, also, , it’s , the only way you’re gonna get buy-in on going after that tech debt that, , nobody even gets, , what you’re talking about, or why it’s important, is if you’re , in order to do this, I need to clear this out. Katherine Bayless: Then it’s , , , clear that out, whatever. Uttam Kumaran: But if you’re , … the problem with engineering all the time is they all… they can never articulate how the tech debt is, , is slowing down an outcome. Instead, they’re… they just talk about it an annoyance, and then they’re , oh, it’s just annoying, don’t deal with that now. You need to almost… for a lot of our clients, we structure it as, . we’ve gone this far, but the reason we couldn’t go further is because of this, ? … Structuring those stories, especially as, . As you mentioned, , if we start to get hit with, , rules around Access, or procurement, or security, , that’s where you’ll be able to drive, because you’ll be the one with the wins, ? , cool. maybe we can… a couple of things on, , this, , define and map. , one thing that we’re gonna start to work on is a bit of, , a metrics glossary. at this… this will be something that, , we … this will be starting and going on Until we reach that, , , at least the… highest level. KRs, , what it’s defined as, , and there’s… Ideally, some type of cross-functional sign-off on, , this is what a member is, and here’s how it’s defined in the logic. And, , we’re all happy with that, ? And that may take some powwows , but, , that’s what I want to start on. we’re gonna… we’ll kick that off and start to save, and as we work with Kyle and meet everybody, start to keep track of, . when folks mean revenue, , what are they talking about? When folks mean members, what are they talking about attendees? I also know that there’s more than just the CES business, we’ll start to learn a little bit about what the KPIs that people care about. And then, ideally, as we go, once people define a KPI, we’re , tell me the source of truth for measuring that today? Are you pulling a report? Are you looking in a UI? Something that someone calculates, or is there just, , , we don’t even know… we don’t reliably can’t get that, … That’s something we’ll… we’ll do. the rest… Katherine Bayless: . Just on that, , , a couple things that might be helpful. , one is a human, the person that’s going to be the business intelligence analyst, her name is Kai, the Greek letter, C-H-I. She starts Monday, and she comes from, , a very buttoned-up data governance, I just put all the VPs in a room, and we define member, and everybody’s fine. I’m , oh, welcome to the fire, ? . But she’s got a lot of experience in that world, she could definitely help with, , brokering. It’ll help her learn, too, the organization. I also think there is a lot of, . latent, definition, if you will, running around out there. , , for example, the person who runs the CES, , registration site with the… in conjunction with the vendor, , she has this, , 32-tab Excel spreadsheet with every definition and path, and, , , if this, then can’t that, but could this, ? , there is a lot of stuff that exists that we could bring together. as a jumping-off point. There’s still going to be a need to get people in a room and agree on some of the pieces, but I do think we’re not starting as much from scratch as It might look the rest of the chaos, if that makes sense. Uttam Kumaran: , totally, and every… every… every team will have something defined and something measured, but it’s , , do we see that there are 5 different definitions here? , what are the… what’s the dimensionality we need to offer for stakeholders? What are their wants? . Katherine Bayless: , and , , in particular, , one of the things that they talked to me about, , in my interviews, and I’m , we will solve that, but they’re tech debt in the way, ? we don’t charge the correct price for a booth at CES, usually, because we don’t have any integration of that membership data into the sales system, ? , , if somebody calls and is , we’re a member, we’re , I don’t know, maybe you are. We had the same problem with our awards applications. We didn’t have exhibitor and membership status in a way that was connected. Some of that problem is the entity resolution work, that, , these are two different systems, company names are different, they don’t match. But I’m also suspicious, curious, I should say, because suspicious sounds mean, but, , I’m curious if part of the reason that there’s not more trust in that membership data is just because, . there’s things grace periods, and , , , they said they’re gonna send it, , and , membership transactions do tend to be more flexible than, we want 10,000 square feet of exhibit space at CES, in which case, , we’re not gonna call that done until that checks in the bank, ? , some of that, , I don’t even know what you call it, business friction? , , our definition of done and your definition of done are different for important reasons. We need to understand that. We don’t necessarily need to change that, if that makes sense. Uttam Kumaran: , you just… we just need a clear underst… oh, , one, we need a report to show, , , , how many people are outstanding, , we need to somehow reconcile payments collected versus that, and then also, you’ll immediately just be able to see, , who are we undercharging, overcharging? We just did this for another SaaS client of ours, where they’re more… they’re close, they’re… they’re a fast-powing startup, but They just started, , invoicing willy-nilly, and we were , you guys are sitting on, , almost, , half a million dollars in uncollected. Katherine Bayless: Poss, because… Uttam Kumaran: You have mechanism to, , reconcile your seats. with, , what you’re charging people, and you just charge them once, and then they add, , 30 users, and you’ve never… but, , they didn’t know how bad it was, they’re , we’ll get to it. I’m , , that’s, . that’s, , more than half a million in, , ARR that you could just… turn on today. Katherine Bayless: Yay! And you might do that. , , , the guy in finance who manages this, , the piece of the process that’s, , for booth space and payments, he’s got that, , locked down, ? But he knows that it is , done , and why it’s done that way. But, , he sees the potential for that to start eroding, because a salesperson is , I just wanna… I just wanna add this to my, , my numbers for the week, and he’s . ? , that’s not how this is going to work, and helping him defend that line might be good, because we don’t want to have a half million dollars in ARR that just needs to hit the go button, ? Uttam Kumaran: there’s something around, , AR, . Katherine Bayless: , it’s the ship is tight, but I don’t think people understand how much that matters. Uttam Kumaran: It’s interesting. Susan, let’s… my other question is for Kai. she’s… our… , if I can get this , Kai and Kyle are rolling up to you, or what’s, , the… , great. Katherine Bayless: , eventually, I hope it would be, , Kyle would manage Kai and the data engineer, but because everything is new, I’m just keeping it flat that way. It’s not , … sometimes it’s , , too many layers, then you realize. kids, thing, … . And then the guy, , I was just interviewing, the data engineer, , that would also report up to me, but eventually Kyle. Uttam Kumaran: , great. then, , , that, , as soon as we… feel good about, , as soon as we get a couple things set up, , it’s just… it’s just Kyle, we can start talking to him. And then, , , if we can get… , the one thing that is, , you hire… these three… you have these three people, and now you’re , I need one-on-ones with everybody, I need stand-ups that you’re just never gonna get anything else done. Katherine Bayless: , if we can take some of that. Uttam Kumaran: if we could take some of that on and still guide the folks, or, , at least be, , our roadmap. We have a roadmap, and that means that’s also your roadmap, out of these pieces, , what can you take on while we’re set up all the infra? And they… and we help them chase things, but they also come with their fresh eyes, , that could be really helpful. Katherine Bayless: , , . , there’s a lot of potential for, , sunshine and rainbows and unicorns to come out of that collaboration, ? , lots of potential. Especially if we do get a few, , , visiting guests from some of these other. Uttam Kumaran: And, , , these folks are the ones that are gonna, , because they’re new, they have a good opportunity to just, . not worry about what it was. They’re, , naive, … They just start… if they… if we start when they join about, , we have a culture of, , posting questions about data openly, there’s a channel with everybody, you can app people, and there’s, . They don’t have any fear because they don’t know anything, they don’t know any better, that would be great. And then, , and that’s… And now you have three, , You have… Those three… and us asking, , all these questions, it’s gonna be, … it’s not one person’s nosy, it’s we’re all, , hey, we’re completely blocked by, , this person or this process not moving, that’s… it’s very… it’s, , become very obvious, , what’s, what’s not happening, … But as you mentioned, , there’s… even when we hop on the call, , people are gonna be fans of this work, for as many people that may block us, you also have a lot of people that I want to publicly get, , their testimonials and get them to share that it’s been effective, and , , … Katherine Bayless: , , and it’s funny, , a couple times since I’ve been here, people have said things to me along the lines of, , , , oh, , people… people don’t me because I am opinionated or cranky, or whatever, and I’m , I love people that are opinionated and cranky. Uttam Kumaran: , I don’t know what else. It’s nice with people. . Life is… that’s what life is about. Katherine Bayless: ? , . I’m , the crankiness is important data. It’s telemetry. This is going. If I’m blocking something, say it. I might not know. Uttam Kumaran: , maybe let’s go through, , maybe, , the P0 data sources. , I also have open… the Excel that you sent me over. , we can use this. , , you tell me what’s best. , I still think, , our North Star that we signed up for is for, , identity stitching, but… I don’t want… To use, , that to just, , be , , we’re only gonna bring in a couple things, … we can think about P0 as, , , it’s … It’s… it’s the most open, or you can think about it , , it’s already, , we have keys, or you can think about it it has the most data, or you can think about it , we need to shut down that service, why don’t we get the data, we can kill it. But, , those are a couple of the ways I would think about, , , what is P0? if anything… If anything, , shouts out that, I would to arrive on, . At least a couple that are that, that we can start to go after. Katherine Bayless: for P0, the impexium slash remembers, , that data share, that’s , , negative one, ? , the. And for a couple of reasons, partly because it is how I managed to sneak Snowflake in. Also, because the team that has not had, , any real, . they have this incredible, , distribution of efforts on the team in terms of, , spreadsheets, and pulling things together and running their reports, and they, , they’ve got a -oiled machine going, and… . I just love to be , , great, send me all of those spreadsheets, they’re now snowflake dashboards, what would you to do with your time? ? , they’re… Uttam Kumaran: , perfect. Katherine Bayless: incredibly industrious team, and they are spending much time and effort in Excel. I was talking to this woman last week who manages our digital health and our smart cities, , councils, essentially, or boards, we call them. And , these are decision-making bodies for the organization that are volunteer-driven, ? they’re important, but she’s . they’re asking me why they’re on these things, and I can’t tell them what the return on their volunteer commitment is, and I’d to be able to do that, because I also think that one of them, it doesn’t really have much of a return on commitment, and I should convince CTA to maybe, , disband the group, and we’ll start another one if we need it again, thing. That’s fair. He was , I want to kiss you now, ? There’s a lot of brainpower that is just blocked by reporting that takes too long. , Impexio. , then the tricky part is the Salesforce CRM. They would go hand-in-hand with that. The reality is, we are going to onboard 2,000 exhibitors between now and January, and , , there’s just… there’s just time to really get in their way, and , , anything we can do that’s gentle and invisible, great. But , I hesitate to really swat at the beehive when they’re overwhelmed. But Marketing Cloud, however, we can totally incorporate into P0. We’ve… I’ve been working with that team most closely since I’ve started anyway. And there is a specific current interest in, , this swirl between the people we’re inviting to CES, whether or not they are even receiving that invite, how they are actioning it, ? , that thing. , they want to see that pipeline. We have some pretty decent Google Analytics, , once they’re on the website, but getting that, , acquisition end of the puzzle together, , would be, good. if we did, , Impexium Marketing Cloud, and then this one, Merits, line 36, Those would be 3 really high-impact ones for, , useful now thing. The one trick with merits is that they don’t have a single set of REST APIs. They have, , a little bit of this and a little bit of that. I don’t know if we’ll get too terribly far, with, , a true integration, but we have flat files all day, at least we could work with. and model it, it just might be janky, but it’s only four more weeks, ? , I can move my files around in exchange for insights for four weeks, ? Uttam Kumaran: And G… GA is important as . And then, how about, , the… Formstack. Are we using that yet, or ? Katherine Bayless: We , , we use a lot of Formstack. , great. And it’s pretty easy to deal with, , I built a little webhook, for the… , I would say Formstack’s a good low-hanging fruit. And then, EventPoint has really nice APIs, and they’re… , , , . . They have really nice APIs. the only data that’s in there now is, , our speaker data, it’s not necessarily the most high value. However, as you’ll see above, with 3C vents listed. My curiosity is, , could we do all of that Cvent stuff in EventPoint? This is a question I don’t know if anybody even wants me to ask, but it is a question I’m asking. , , I’m curious… if we can get the C event sites integrated, frankly, it would make my life a little bit easier, which is selfish, I know, but… because there’s a lot of, , ancillary CES event registrations happening over there, and when people are , hey, has anybody signed up for that mobility breakfast? I have to remember which C event site that is. Uttam Kumaran: Which, , , , . Katherine Bayless: that EventPoint is a nicer platform, and… If we can get the data out of Cvent enough to be able to figure out, , , this is totally event pointable. Uttam Kumaran: , are there any other, , consolidation opportunities, or are some of these, . the… anything that’s active use… maybe I’ll start with another question. For anything that’s active use, false. are those, , we just need to get the data in and shut it off? Is it, … We’re still paying for it, and … , , give me a sense of that. Katherine Bayless: , truthfully, the meaning behind that data point is Catherine wanted to put it on the list, but not make anybody think it’s active. But the reason I put it on the list is because I don’t know the answer to your questions. Uttam Kumaran: Alright, alright, great. Katherine Bayless: I know we don’t have any more. I’m… unsure if we got data out of it before it got shut down. even if we didn’t extract the raw data from the platform, we do have the historical awards data in the old data warehouse. But, , it’s a question mark. And then the other ones are the surveying platforms, and I just wanted to flag it, because I’m not that that team would have thought about it, ? , oh, let’s make we get all the data out before we close this account. it’s a… Catherine, follow up on question. . Just those three that I had flagged, though. Uttam Kumaran: for Decipher, , . Katherine Bayless: . Alright, , I see that, . , cool. These are great because they don’t hurt anybody, I can ask a bunch of questions, probably. Uttam Kumaran: , I’ll ask Kyle, and then we’ll just be . Katherine Bayless: That’s cool. Uttam Kumaran: I’ll just go down the telephone to find out. Katherine Bayless: , , start with Kyle. As far as the opportunities to combine things, ? many, ugh, it’s , . , , , Causeway, for example, this is, , a volunteer management platform. I have yet to see it. Maybe it’s delightful, but all the team does is complain about how something is broken, and I’m , , , , what are we really using that platform for? Sorry, I should back up. There’s a foundational piece I didn’t mention. There is a very strong desire internally to put more data into Impexium. People don’t. Because it’s a pain in the ass. we can… if we can solve the data coming out, making it look really nice, they might feel more incentivized to put data in, even if it’s difficult. And then we can deal with Pexium and tell them to make their product suck less. But Causeway, , if whatever is in there could be achieved through Impexium, awesome, we could combine. The same with, Quorum, , I know what Quorum is, I used it at my last place, it’s … . , but I’m , I… I don’t know that what they’re doing in Quorum is stuff that couldn’t really be done in Impexium. I also, , it’s a dismal opinion of Quorum, to be totally honest, but that’s just me being snarky. And then even, , some of these things, Qualtrics and Marketing Cloud, , I would never say… obviously, but, , there’s probably parts of those processes that could be elsewhere, , reducing dependency. Formstack, for , we could probably, , find things to combine in there. , all of that to say, lots of options. Uttam Kumaran: every… everything around the… Survey, and then the member… And the exhibitor… , outside of, , the show. Is that… is that a good characterization? Because none of the items here are exclusively related to, , putting on the show, meaning, , the floor data, the on-site stuff… … , anything around security… Katherine Bayless: That’s … that’s probably not a bad column to have included, is , is this a CES-only platform? Uttam Kumaran: I see, , , . , , maybe I that. I just having as many dimensions as we can have, and I’ll just… , I could cut it a couple different ways. Katherine Bayless: , I know, , a lot… a couple of these I know, are… , , , the beacons, and I don’t even know what we use for them. I know they’re just CES. Uttam Kumaran: I assume beacons may… these beacons could probably be QR… , QR codes? Katherine Bayless: Or these are act… Uttam Kumaran: Because I used to work for this company called Flowcode, and I’m pretty Beacons was one of our big competitors. Katherine Bayless: , , that might explain why. , it could… Uttam Kumaran: , it could… it could be these guys? Katherine Bayless: , is there, . Uttam Kumaran: Simple landing page development. , , for example. you could have, , a beacon for every exhibitor, , , you scan their QR code, it goes to, , the exhibitor’s beacon, it’s, , you have, , 5 links, it’s, , those really mobile-friendly pages. Katherine Bayless: that is probably what it is. My brain, when I kept hearing beacons, I was always thinking, . Uttam Kumaran: NFC beacons. That’s also the first thing I thought, but then I used to think… I even worked in this field, then I re-associated, , what it is. We used to, , compete with these guys, … . Katherine Bayless: , that’s funny. , , maybe that is what that is. I do know it’s CES only, but . , and then, DocuSign is CES only, but, , it’s just baked into. There’s, EventPoint, Event Base, ExpoCAD, those are all CES only. Glean In is CES only. the lead retrieval CES only. I do think that is. Uttam Kumaran: But lead retrieval… Katherine Bayless: Could be helpful. Uttam Kumaran: Because it is almost … It is… it also is data about the members, , who’s attending, ? Katherine Bayless: , here’s a funny thing about the lead retrieval data, and this is, , I just, at some point, need to fan out the answer to this, but, , everybody has this weird idea that we can’t use it, because they’re , , it’s not our data, it belongs to the exhibitors, and I’m . Uttam Kumaran: What does that mean? Oh, I see what , this is, , when, you go up to an exhibitor, and they’re, . oh, I love your… I love your demo, , , let’s grab a fee booking, let’s grab a demo next week, and then… Katherine Bayless: , can I scan your badge? Uttam Kumaran: , , , oh, , I could put you in our mini CRM for follow-up. Katherine Bayless: , and I’m , I’m pretty we can use that data, guys. I have not yet… Uttam Kumaran: , give me a sense of, , you would be using… because you… you would technically… oh, you would look at which exhibitor got, , the most… badges, Or, , somehow slice it to show, , . Katherine Bayless: , I just… I data. Uttam Kumaran: , , , , I know, , I would… all this. We’ll figure out the use case, but trying to think about, , if I had to go talk to them about what… , , … , I feel this is… This is also something that, say, the balance of, . There’s some stuff that’s essential, but there’s also some stuff that’s, , you guys have, , one of the most unique Lore, exhibition data, ? , there’s probably a lot more that can help you inform things . Book traffic and form things , oh, this vendor clearly got the most scans, maybe we should have put them… we should just, . Put them on stage tomorrow, because they’re just blowing up, and, , everybody wants to, … maybe this… and this is something that, . maybe the people who are running a show on a daily basis want to look at, , how many leads were exchanges. , I have a general, . Katherine Bayless: That would be cool. Uttam Kumaran: This is, , that’s where it goes into, , you guys uniquely are spitting out a lot of data that doesn’t exist. And , it’s , what products Can you guys develop? around… Around that. Katherine Bayless: , , where my brain was, , foot traffic plus lead retrieval gives you , , a, … , in the other order, , lead retrieval would be, , your numerator, foot traffic is your denominator. If we were able to knit that together in the, , what we call Eureka Park, where all of our startups are, which, asterisk. going back to definitions, we define startups incorrectly, and it is going to be a business problem. I don’t know how long it will take to fix that, but, , it’s this weird set of rules that don’t make any sense for real startups. And when we say startups. it’s a grab bag of tiny companies, . But if we fix that, I’m , foot traffic and lead retrieval in Eureka Park, where the startups are, , could we get a sense of predicting, , how they might do, ? , if this is what your, , your ratio was that year, and you come back next year, and, , you met with these investors, can we. Uttam Kumaran: , but your salespeople should also have, , here’s how many scans you got last year. , in… for example. I worked at, my first job was at WeWork, and one of the things I worked on was I worked on, , the badge scanning data. That was, , the , , mystery… data set that I brought to, , life and made them a lot of money, because they were … they were , we have this badge scan data from our badge scan vendor, but, , nobody’s ever modeled it. I’m , you have every keycard scan to every WeWork door. that exists, and that we had, , a lot of WeWorks, I was . you can do much for that, and eventually, we built a huge model around keycard scams, but not only, of course, you could do a lot of human stuff, but we wanted to show that, , hey, your people, , for example, we would sell both floors, and then we would… companies would have, , hot desks. And we would say, , you have, , 5 hot desks, but you see there’s, , 10 people coming and swapping in and out of that. You should open an office there, … Katherine Bayless: . , and that’s what we would do. For example. Uttam Kumaran: you could say, hey guys, you guys got, , 100 different lead scans, but you’re in the corner, you guys should move up to the floor, and, , here’s what you can expect. And you can also, for your salespeople, you could say, on average, people on this side of the floor, which is X dollars, get X percent more scans than the. people, ? there… and that… that is the story for the salespeople to tell, ? Katherine Bayless: . I, . , to me, , seems there’s totally things that could be… Uttam Kumaran: , I’m also interested, because the nice thing about this stuff is, , if it’s… if sales drives it, and if it’s, , your data is helping them close… net… close net new business faster. Or give them a story. All doors will open to follow the money a little bit. Katherine Bayless: , if they’re , we need this data, because. Uttam Kumaran: We want to put together, , a… end of CES dashboard with, , all the metrics, and, … or when I go call them for next year and getting a bigger booth, I want to have, , a… customer health dashboard, ? I want to see how they did last year. They’re gonna eat that up, because probably they’re going in and winging it, ? Katherine Bayless: , and , truthfully, in a funny way, , there might be a lot of opportunity to start that work on the membership side, because the membership team winds up being the, , , I don’t know, , the kid brother, , running around behind, , hey, us too, us too, ? Uttam Kumaran: , I see what you mean. I’m not hearing you. Katherine Bayless: there, it might help them tell their story and broker some trust with the sales team, because , . Uttam Kumaran: , I see what you mean. Katherine Bayless: , the sales. Uttam Kumaran: Sue me, because… . Katherine Bayless: They’re lovely people, but part of the challenge is, , for a long time, we didn’t have to do sales. We just answered phones and took checks, ? But now, the world is different, and we have to do more selling, selling, and I don’t. that reality. Uttam Kumaran: But also, , without a data story, , I’ve… and this is where I don’t know… Katherine Bayless: This is really helping, because I have some friends that are in the event world. Uttam Kumaran: all of them struggle with this type of giving data back and telling the story of the ROI, which was… it’s good and bad. It’s good because you guys can not do that and continue to be the premier place to advertise, because you tell that story. It’s also tough because it’s just not used to that, … But this is where, , you want to go to your customers, say, you had a great ROI, you should continue spending with us, versus calling them, and then both of y’all are, , debating whether it was worth spending money on a booth, when it totally could have been, but, . Instead of just, , showing the numbers and being , great, cool checks in the mail. You have to, , dance or use other sales tactics to figure it out. Katherine Bayless: . , I told you the Samsung story from, , the last. , , , . , it’s brutal. , let’s see, map your show, merits… Nix Halo, more agency. And then… Salesforce CRM, , obviously. Session scanners… Uttam Kumaran: Oh, this is great too, . Katherine Bayless: , that data we have, and nobody. to play with. And, also, nobody knows how to import into a database, because it’s too big. And I was , , . Uttam Kumaran: Oh, . , it’ll solve that problem. That’s great. This is great, too. This is great for… to show, , what sessions are doing , and… Katherine Bayless: , then Titan, turn out now… and Zoom info. And that’s the extent of the, . Uttam Kumaran: Oh, you guys do use ZoomInfo for enrichment? Katherine Bayless: I don’t know. one of the CES one of the other CES tools has it, , baked in, and it may be… it might be… Oh. I’m not . But , I’m not… I don’t think we have a… Uttam Kumaran: You guys already have… , . , I can find out, because, . It’s decent. Katherine Bayless: , , and the other thing, too, is, , the… , Impexium as a product has concept of, , baked-in enrichment, ? , , at least in HubSpot or Salesforce. They’ll, , , get this. Uttam Kumaran: Goodbye through their credits, . Katherine Bayless: , but , if we could bring any enrichment into Impexium, that would make their lives better. Uttam Kumaran: And then talk to me about, the… Turnout now? Is this… , I’m also trying to think about, , it’s , we can get you, , a… a win… , during CES, that’s, , there’s some live data coming in to show people. , is this important to see, , what people are asking for? That’s something they already produced? It’s , I’m talking about more the… the chatbot. There’s probably some stuff we can do with the attendee matching, but for the chatbot, for example, if you can see that, , people are asking a certain type of question, is that, , anything relevant. Katherine Bayless: … for the attendee matching thing, I’ll say, apparently, I don’t really understand all of it, but it’s a very tiny program, and I don’t know if it’s tiny because nobody wants it, or if it’s tiny because it’s restricted to certain tiers or something that, but it’s a teeny tiny . Uttam Kumaran: Oh, really? . I wonder if you can show that, , people who match attendees, , stay longer, or go to more stuff. Katherine Bayless: now, we can’t get more than 30 people to sign up for it, … Uttam Kumaran: , alright, then there’s not much… not very significant. Katherine Bayless: But the event co-pilot, the little AI chatbot. , let’s see, I know I tell a lot of stories and talk a lot, but… Uttam Kumaran: , that’s fine. Katherine Bayless: Last year, our board was , this is embarrassing that we don’t have AI. Oh, is it? then Panasonic was , we will help. And , yours truly got on a call with a woman who leads Panasonic’s AI division, because they wanted to help us build AI at CES. Uttam Kumaran: And she asked what… Katherine Bayless: tools we used, and I said, , we don’t own any of the source data, and our mobile app is in ColdFusion, and I started 4 weeks ago, and they think this is a conversation that should happen. We had a good chat, and then Panasonic politely bowed out, and I said, we’re gonna hire somebody, because whatever AI we throw at CES in 7 months is gonna fail, and I am not sending my career down with that ship. . , I feel bad. The guy is a little bit of a… , he’s an interesting character, but, , this thing’s gonna get eaten alive. It is gonna get eaten alive, because it’s terrible. Uttam Kumaran: , , I saw the… I saw the thing. I wish we… I wish I had, … Katherine Bayless: It’s worth the call. Uttam Kumaran: Should’ve started a little earlier. , this is tough. … Katherine Bayless: Now, that being said, I do think, in a weird way, if we… because he has shown us what the analytics will look . we get, , a daily digest, but there is the ability to consume in real time. I just need to read his email and see what the details were. But, . We could maybe do a, … siphoning the data out, surfacing what’s going versus not going , , at least we’re learning what we need to fix. Uttam Kumaran: And who’s this… and you said it’s guys, this is a one-guy company? Katherine Bayless: , the CEO and co-founder is the guy that’s joining our technical calls and telling us what he’s building, I’m assuming it’s not too many more people than him, but it’s definitely one guy, but it’s not many. Uttam Kumaran: , at minimum, to just be able to say, . , but… and then, is Casey on our… on our side… , is he owning the, , the fact that it works? Is it more about whatever that guy needs? Katherine Bayless: , Casey, she is just stuck with this, … , great. we had a woman named Emily Kaiser who worked here until, , about September, before she moved on, because she doesn’t live in the area, we did the RTO thing, going back to what I explained about at the beginning of the call, ? Emily had been managing the… she was intending to manage the AI chatbot, she’s been managing. Casey is gonna be on maternity leave, probably? But she’s been holding together the website, the mobile app, and the AI chatbot, and probably 10 other things, and somebody needs to buy her an all-expenses ClubMed vacation when she’s done, because I don’t know. , I know I’m driving her insane, because I’m, , day late, dollar short to everything she needs. And I can feel it in her eye contact, and it’s very fair. I’m not objecting at all. But , , , she’s way underwater, but somehow holding it down. Uttam Kumaran: at least, , maybe we can just get the data from the chatbot, and at least there’s, , one other level of, , is this thing working? And you’re not, , relying on the vendor to tell you that it is or not, or, , anecdotally, , it’s lagging or whatever, … And if it’s just one dude, then, , it’s one email away from being , , we need this data, thanks, and I can start that. Again, , even… that’s… that seems … Low effort, because it’s, , one guy, one couple emails to get access. And in the spirit of not making too many things P0, because then they’re not P0, I feel this is probably… Katherine Bayless: half decent list. Uttam Kumaran: I know you mentioned P negative-1, I’ll think about that. But a lot of these are great, and then , , , we at least… I wanna… I wanna see the… the GA data, I want to see the membership And then, the only, , bonus win is if we can get some of the live… if we can get some of the on-the-floor CES data. piped in, it could be a really good win for you, if we, , something can come out of that, or, . for example, I assume, , if a good example is, , Kyle, you have access to all this live CES data, we’re testing it the week before, now you can start sending reports live, , that could be a good win, but I’m… if that’s, … we’ll just see. If that’s, , a… if I can… we can also get that as a win, then it’d be a win, but there also is some good consolidation opportunities to start just asking questions about some of these sources, . finally get an answer on some of these exist. Can we, as I start learning about Impexium, I can start giving, , my opinion on, , what we can consolidate in there? Things that. Katherine Bayless: , . And … , also along the consolidation lines, another person that I… I do really want to bring pretty closely into the project, , is Jay. , he’s a… he really… he’s a very compelling cast member in my mind, because he’s been here 20 years, he has every to be burnt out and all the things, and yet, , he’s still, , slacking me, , in the night, , hey, what do you think about blah blah blah? And I’m , you care. much, ? I’m , I wish you had a staging environment and a bigger team, but… I’m smart, and he cares, and , . I think the consolidation and the procurement and, , some of these things would be… , he has the, , the seat at the table for, ? , and it would probably help him establish a different relationship between IT and the organization, because he’s victim to the thing that, , happens to anybody who stays along. He’s , they still see him as the help desk guy from… ? , reset my password, why doesn’t Outlook, , whatever, whatever. And it’s , , I want to put him in that strategic spot where it is , , I’m advising on, , consolidating these platforms and staffing up in this area because I see the technical needs of the organization. He does, just nobody thinks about him that way. I want to make him look good. , great, … Uttam Kumaran: , , that’s perfect, and I’m thinking that at least, , , we should make… a core data team channel, that could be the one we have now. it would be great to start, , a data IT channel. as we start getting some wins, we can start a larger, , show-and-tell channel. The reason I don’t want to bring… this needs to be, , the core… crew, because… The, every channel will become a, , this looks wrong situation. I just wanna, , I just wanna not have every channel be that. And, , unless you’ve… unless you’ve, , seen that before, which I’ve seen now many times, you… you don’t, , learn your lesson, try to, . have a core crew that can lean on each other, and it’s , hey, this is wrong, , can I call you real quick? Katherine Bayless: To fix this, and then we have, . Uttam Kumaran: these, , external-facing ones, that’s a good way to start. Katherine Bayless: if we can build the muscle of failing out loud and getting more comfortable, that’s great, but we are totally in the stage now… , and I don’t want to get started. Please don’t just put my dirty laundry out there, ? Uttam Kumaran: , , , I agree, and I don’t… I want… I want these… the folks that we support, I want them to, . Be able to, , Fail first with us, get the sense, and then start to, . Works for some of the external folks. Katherine Bayless: . And then also have a team to fall back on when eventually they, . Uttam Kumaran: They get to a meeting, they don’t have the thing, or something breaks, . Versus, , oh, one… how do I… I ship some dbt models, and, , one knows about it, and, , build a real, , data team, ? … , the last piece… , I read through your message. this is really helpful, I will… I’ll keep this in mind. the lovely thing is, , if you… if… one perspective I have, given your situation, is to try to consolidate as much stuff into Snowflake as possible. I also just talked to Snowflake today, and they’re offering some really great, . some really great, , discounts that I don’t… I don’t think they were previously advertising. For example. I talked to the guy today, he’s , , we’re offering, … he offers, , marketing partner discounts, where … 25% of your Snowflake contract can go to approved partners. And I was , when did this release? He’s , just recently. I’m , dude, that’s, , great, because now, if I, , can find a tool that fits their partnership on the ETL side, and you can adapt some of that money, that could be… , a good win. Katherine Bayless: the Snowflake rep that I was talking to mentioned that too, and it was, , one of those that I was , I filed this away for, , that’s a later problem, or opportunity, but . And he said, technically, it’s 49%, because it just can’t be 50, or else it’s technically money laundering, ? . , we’re saying 25, but he’s , it can be more, and what. , but but I … Uttam Kumaran: And he said… he said you… even if you were to take that amount out, you would still get the discount tier from the top level. , my guy told me the same spiel that your guy did then. Katherine Bayless: , you were better poised to understand and think about it than I was thinking. Cool. , , … I did run it by our finance team, who are a little, , maybe on the older school side. Really lovely, but, , I was , you said we have money to burn, end of year, ? Use it or lose it, , can I buy a bunch of Snowflake credits? And they were , , but… Would we use them this year? … but I do think in the new year. Uttam Kumaran: What did, , how did… , what did they… are they… are they concerned about, , what is… how do you win them over on decisions currently? , , , how do they… . Katherine Bayless: , I… I… I… I don’t… I don’t know yet, to be honest, ? , I… I have not yet figured out, because that… that whole, , pay for, deliver for, that is a brick wall that I am just fascinated by. I’m … . I just… it doesn’t have to be that way, ? , I also know that they are one of the teams that refuses to move off of their map network drive. , sometimes payroll gets, , held up because the drive goes down, which is just weird to me. I’m , what? , , , truthfully, there’s a part of me that’s , , can I get all the pieces on the chessboard lined up just , where, , Jay, who has really strong relationships with them… . …could be the, . Uttam Kumaran: It’s fighting for that. Katherine Bayless: , because it’s , that whole accounting and finance team needs an overhaul of, , just technology writ large, but, , they’re not gonna let me in away, but they know. They might. . Uttam Kumaran: I just want to know whether they’re looking for… if it’s, , we’re investing in the software for growth, or is it, , we’re getting a good deal? Katherine Bayless: Definitely. Uttam Kumaran: , we fight for the… we’ll fight for discounts and the deals , but sometimes… it’s, , for… for use, it’s , . But some people are really, , I want to know that we got, , the highest discount. Possible, and we, , pressured, and … Katherine Bayless: , they just want to spend as little as possible. , . , it is interesting. Uttam Kumaran: I can ask Jay, too. I’ll ask Jay when I… when I meet with him about, , what he thinks. But it’s nice to know that, , we want to procure stuff through AWS Marketplace, that’s great. that’s, , will be my… , that’s gonna be , , my principled approach, apart from bringing in other tools, and… for those other tools, I’m gonna tell them, you get on the marketplace, because we can’t do this deal, , … But that’s… Katherine Bayless: They can do that, , . Uttam Kumaran: , and if I know Fivetrane you can procure, I should look at some of the other ones, but, , that’s helpful. the next, , the next biggest decision for us to… probably make is, , ETL, what I’m gonna start to do is… is look through for these tools, , what platform can do some or all of them. Also, , , a lot of these ETL platforms, they offer free trials. Some will build connectors, if it is, , hey, we’re gonna roll with two until, , one of our preferred ones builds them all out. , we can do some interesting things. And , , Fivetran is a really generous trial program, and I can call them and get more. And there’s opportunities. The other thing is, , you can do a lot of consolidation within Snowflake, , , dbt, you can run entirely within Snowflake, I’m gonna look into that as . If it’s all flat files into S3, we can use Snowpipe, we don’t even have to go through… Katherine Bayless: . ETL for a lot of stuff. , in general, , having a, , a pretty nice, , flat file ingester into Snowflake… Uttam Kumaran: Totally, it’s the best. Katherine Bayless: , , and , part of me is , even for some of these systems that might have a connector, there’s gonna be much legacy… and, , , I go back and forth, and , on the one hand, if I support it, they’ll continue doing it. On the other hand, , there’s much legacy stuff that people are gonna be , . I don’t know how to, , find it in all of that, but I have my spreadsheet, and I want to use it, ? Maybe there’s a grace period where I allow spreadsheets. Uttam Kumaran: , that’ll be up to you. That’d be how much. Katherine Bayless: Can’t remember. Uttam Kumaran: stick you want to do on the BI side, but… , , I… this is a good place to start. … , as much as we can consolidate into Snowflake, would be really helpful here. And then, , I kinda… I’ll keep your… your notes, in my… I don’t think anything… Shouldn’t necessarily be above that. Price point. But the ETL decision is gonna be the one to make soon, but really, I’m gonna go through and map out all the ways we can get data out, and then start to identify, , some vendors, and , I would say, . We’re good at just getting… if we can even… a lot of these folks They’re just, … they’re just software vendors, you can really put pressure on them and get really good pricing. Especially because you guys are great logo, and , , we’re good at, , really Getting good deals for these folks. Katherine Bayless: , . , and we do allow use of our logo, … Uttam Kumaran: It’s great. , cool. I feel I’ve took… taken… . Katherine Bayless: I was gonna say, the Impexium thing, I keep forgetting, it did come with… Power Automate, by the way, , … Power Automate, a key… I’ve never used Power Automate, idea, , what this all entails, and they emailed us the key in plain text, I sent back, please rotate this, and send me again in a secure way. But but , I don’t know if there’s any, , leverageable potential on that one. Uttam Kumaran: , the other… the other thing I was gonna mention is, , , typically we would implement… , if most of your stuff is modeling, then we can get away with just dbt, but… , we would typically implement, , a Dagster or, , an Airflow, but the other thing is, you… Snowflake has now a pretty good task-based orchestrator that can execute Python workloads directly in Snowflake. this is a fun… I this gig, because it’s, , we just try to consolidate as much into Snowflake as possible, and , it’s , , it’s a great… now that Stuff has a lot of those tools, I feel we should be just fine with that, for the most part. Katherine Bayless: , that’s a good… It is … it’s a good way to think about it, too, from, , the educating and growing people’s, , comfort with some of this stuff, is, , we can probably get really far in Snowflake, but also start telling the story of, , look, we’re pushing it to its limits. If we really want to get further, we, . Uttam Kumaran: . This tool to it, ? . . But I wouldn’t, , I would say 2 years ago, or even 3 years ago, , it would not make sense to do a lot of this in Snowflake. They were just really immature in some of these other things, but they’ve built the platform in a way where you can do everything in there. And then we can decide, , , to split things out pretty easily, but… Given that the tax year is not the budget, but more of, , the approval side. That’s gonna kill all momentum. Snowflake will… for our IT and InfoSec folks, that’s great, because it’s just one. Katherine Bayless: It is something that I do. Uttam Kumaran: And, , Snowflake has, , the AI features, too. … , maybe I think about that. It’s such a… , I feel some places, , it’s low budget, but, , you can get stuff signed. Here, it’s a little bit opposite. . , I want to, , make that we’re not, , bogged down in, , getting a big vendor for 6 months. And then we just , , die, ? There’s just, , momentum. I don’t want to do that. I just want to be , we got Snowflake signed off. Great, I’m gonna… we’ve become the poster child for, , leveraging all Snowflake’s features. Katherine Bayless: , , when I… the guy I interviewed earlier, he was, also, , elbow-deep in AWS Glue, and I was , how do you it? And he’s , it sucks, but I know how to do everything I need, and. Perfect. , . Uttam Kumaran: , it’s good for security, but, , you could tell, they just, , don’t… they don’t need to… Build out that product, because it’s, , comes bundled. , it’s… You buy everything at one shop, they’re not gonna innovate on, , every single thing. Katherine Bayless: There is maybe some room and interest in the new year for, , a big picture observability tool, Datadog or something that. , Jay’s very interested in, , now, he’s , my kingdom is tiny, but I’d it to be bigger, and if it’s gonna be bigger, I need to be able to watch it, … Uttam Kumaran: , let me talk to him about that. That’s also a lot of stuff that I’ve done in the past. , we’re evaluating all the major data observability logging platforms. , Datadog is , . , , king there now, but there are a lot of options. I can… I’m gonna… I’ll put that on my notes to talk to him about. Katherine Bayless: . , we used Datadog at my last place, and I will say that I managed to harness, , 3%, , of the functionality. Uttam Kumaran: Extremely confusing, and now it’s massive. There’s, , a couple of other firms that are really good, , LogRock is good. Katherine Bayless: There’s a couple. Uttam Kumaran: other companies. I’ll ask. him, because ideally, , I want to make to use it for… Our stuff on the reporting side, and then if he can get stuff out of that, too. That would be great, … Katherine Bayless: . In the meantime, I did at least set up, , , CloudTrail and AWS Config and some of those things, that we’re at least monitoring all the stuff that’s happening in there, but… Uttam Kumaran: , perfect. This is great. I have many… much stuff answered. This is amazing. Katherine Bayless: It’s funny, , I feel every time I talk to you, I’m , , I can do this, I can do this. Uttam Kumaran: , you could totally do this! , , I just want to make that, , we’re not… we’re not , oh, it would be great to run, , this ritual, and then, , you run it. It’s , I’m . I will take on… , we can take that stuff on and have people run with us, and then eventually, , I’m gonna start having them try to run some of those. . Because you want us in the… in, , whatever the worst situation is. Or, , whatever the hardest, or the thing where you’re , just go figure this out. That’s where you want us? Did you want to save your people from… that, because that… that… we won’t… , I… I won’t get emotional, is what I’m saying. , , I just… I come in, and I’m , I’m just a consultant, , I’m asking these questions, , I’m working for you. But, , the internal people, I want to make they , , learn, and they do that safely. Katherine Bayless: , , , . . , it’s awesome, really awesome. , , I’ve already… I’ve mentioned you to a couple other friends, I don’t know if you’ll ever get any phone calls necessarily, but . Uttam Kumaran: Oh, I appreciate that. Katherine Bayless: former boss of mine, who’s now at the American Association of Physicists in Medicine, ? , , , Physicists in Medicine. and he’s looking for a director of IT membership and marketing for $120,000 a year, and I was , good luck, dear. But he’s , girl, it’s crazy over here. Physicists invented the internet, they think they’re technical, they built everything in-house, and they are not… the people who should have done that, and I was , you should call this guy out. Uttam Kumaran: I did… I did at physics, , but none of my work is physics, , it’s , if I’m lucky, we get to do some multiplication around here, , maybe some division. only once sometimes in my career were they doing, , derivatives that for SQL, but . , I’m happy to… happy to talk and help anyone out, , that’s awesome, I appreciate that. Katherine Bayless: , . I’m just really excited. The collaboration is much appreciated, and I feel … , courage to fight one more day. Uttam Kumaran: , definitely. , and we’ll… we’re good at, , trying to codify these things and documents and keep them fairly up-to-date, hopefully, as you go to meetings, , you have… you’ll have this Gantt chart, you’ll have some project plans, and then as we start to roll, and we have more check-ins, we’ll have decks that you can repurpose, … Just be easy for you to share wins, , that, … Katherine Bayless: , , , the OKRs will be really helpful, too, because, , I’ll be leading the charge with saying, , here’s a way that I am measuring my performance, and sharing it openly, and be , what? Uttam Kumaran: , I know, it’s rare. Katherine Bayless: . . For slide decks, do you use Gamma? Uttam Kumaran: We started using Gamma, I… I was using it, and then the new version released, and I haven’t used it. I’m trying to get my team to build using on the API, because we have, we have, , we have a pretty good deck format, and we have, , our brand book, and I’m , guys, the new API just came out, , can we start to automate some of these? that’s, , gonna be someone’s, . Christmas project to, , think about how to do, because we do a lot of deck work, and we’re… we’re pretty good, and, , but I would it to be all prompt-based, we’re not spending time, , doing Google Slides things, but we… I… we’ve used… we use… some people in the company use Canva, . Katherine Bayless: , , , I’ve… Uttam Kumaran: Pretty good. Katherine Bayless: , , , obviously in my world, I don’t have to make slides that often, it’s more if I’m giving a talk, I do it. I increasingly… I’m getting to the point where I’m , hmm, it’s very good at design and layout stuff, but I’m , I wish it would also, , comment on your narrative arc, or , … Uttam Kumaran: , usually I’m exporting it to PDF and putting it in the GPT. And asking for, , the copy. What I found is , . Doing an ugly version first, nailing the content, and then… Shove it into there, versus people start there. Because it’ll just, , hallucinate a bunch of stuff. And I usually make an ugly version. I almost treat it I would… some of our decks, where I’m , I make a really lucky version and have someone on our marketing team help me just, , make it look way better. But, , I couldn’t… but, , they… they’re good at that. They can’t… they wouldn’t have been able to take what’s in my brain and start it there, … I’m almost trying to treat it more , , if it was, , a marketing intern, I would have to give it, , the first ugly version or something, that it… it doesn’t, , try other stuff, ? Katherine Bayless: , , , . , I feel in sticky note, and, , one line at a time. . I’m , , whatever this is can’t be a prompt, whether it’s to a human or an AI. I’ve gotta, . Uttam Kumaran: Are you… do you use, , Whisper Flow at all? Or anything , voice-to-text? Katherine Bayless: Hmm, , what is this? Uttam Kumaran: consider Whisper. I’ll send it to you, you should consider using it. , a lot of times, , when you’re… when you’re working with AI, you’re, , doing a lot of typing, and , of course, the AI’s typically Small changes to your prompt can improve the… , the outcomes a lot, and usually my… my rule of thumb is that, , it’s only as good as your context, you need to just give it a lot more. And typically, of course, people are typing, and typing is not as effective as if you were just to speak. Typically, you can speak and talk, and talk, and talk, and you get 2 minutes of prompt, versus, … I don’t know if I could have written that, , you’re… it doesn’t flow that way. I’m not a writer, it doesn’t flow directly to the hand. , I’ve been using it to do speech-to-text for almost, , a year now, and it’s amazing, because I just sit here, click the button, talk. Press… it types it out, and it’s not , when you use a voice mode where it talks… I still want the output to come back in text. But instead, I’m able to give it way more input. Very similarly, , if I have to type out a long answer to somebody these days, I don’t type it at all. I’m usually just, , fucking it out. It’s very, very helpful. Katherine Bayless: , , I use voice a lot, but, , I’ve… it sounds silly, but I’ve never thought about a tool? , and even I’m, , really. Uttam Kumaran: , but you used to have Dragon, ? Dragon used to be, , the big speech-to-text. But it used to be for, , writers, and, , it was janky. Now. Voice understanding is very, very good, and real-time. Katherine Bayless: , what I’ll do a lot of times is, , I’ll do… , I opening eyes, because it’s processing the raw audio versus transcribing and writing. I’ll talk to ChatGPT voice mode, and then I’ll say, , , now write a prompt that I can take to Claude. Uttam Kumaran: , , , to each their own, , you should be… you should just be doing this all in cursor. You could easily switch, ? Katherine Bayless: , I … Jay was interested in, , hearing from you, , probably, and anybody else. He’s , he’s… because he’s , all the cool kids are using cursor, but I can’t figure out, , should I really. Uttam Kumaran: Oh, I would… I would love to show him. He’ll be, , in there all day. , it’s just… I was just talking to a friend yesterday, and I was , I want… everybody should… every role in my company should be in Kirscher, because You could write, you could do project planning, you could do mermaid diagrams, you can ship PRs, you can call external things. You can build apps. I’m , , maybe every… maybe I should just force everybody to do their job in cursor, because for, , for you guys, , we’ll create a repo, we’ll show… I’ll put in the project plans, I’ll put in our meetings, I’ll put in these notes. And then it makes it way easier to iterate, , when I’m writing a dbt model, I do want it to have context on, , what we’re here to do, that it makes the trade-offs or focuses on the thing. And I’m , . , , I want everybody to be working in there, ? Even the… even our PMs, , I want them to start there. Because, . Katherine Bayless: , that context ops is everything. , , I… To a certain extent, that’s ultimately what Jay and Mai’s team will become someday in the future, ? It’s, , as your context ops team, and some of the people… , this is knowledge engineering is the… . , ? . , , some of the projects that I built out in Claude when I started were , , I would… because I’m a… I do do handwritten notes, because it helps me focus. And I had it, , trained on my handwriting I could take a picture of my notes, and it would transcribe them, and then create these little, , atomic knowledge cards, and then I would put those into the project where I was building on AWS, and , , I could ask it, , , to use the instruction set for the, , console expert with the context of all these meetings. And, , , I blew people’s minds when I would be , hey, what was that thing I needed to ask Emily about? , that thing. It makes a difference, , . Huge difference. Uttam Kumaran: reference. , now I’ll show them our, , cursor setup. , maybe I’ll set up your repo for cursor , and then I’ll send it to them, . Katherine Bayless: , , , , , that would be good, , because I can keep you going for it, and I would love to, , move in that direction. Uttam Kumaran: , perfect. , I appreciate the time. I’m gonna do my best to get some of this organized. Maybe… are you off all week, or are you working until Thursday, or… Katherine Bayless: I am, , I’m around. , from noon to 2 on Thursday, I will be forced to do family things, but otherwise, please save me from the holidays, . Uttam Kumaran: , if I can get something organized and start to get us to think about one of these directions, that’d be great, and then ideally may… , I know a lot of folks may be off or not really responsive, maybe we wait to hit people up for access if I need to for next week, but at least I can start to… Oh, , . , . , cool. Katherine Bayless: , I was gonna say, , , a lot of people are out, but I’m here, and it’s a great time to, , soak. Uttam Kumaran: Oh, great. Katherine Bayless: time and attention, but , everything that’s in that list, I can grant… or whatever. , , , , just. And then you… They did send us that, data share in the other EIA. , whatever makes sense for getting the data out and into somewhere else. Or even just looking at it in there, , but… But , definitely don’t be shy, and , I have meeting tomorrow or Wednesday, I am. Uttam Kumaran: Oh, great, , then I’m gonna try to jam on some stuff, probably tomorrow, and then see what I can get out. Katherine Bayless: Awesome. He tries around tomorrow and Wednesday, he’s out Thursday, Friday. Uttam Kumaran: , maybe I’ll see if, , maybe the three of us can hop on, or… . just say hi again, and, , start thinking about overall plan, and I can brief them on, , , what we’re… where we’re going, … Katherine Bayless: , , , that’d be good. Uttam Kumaran: Perfect. , appreciate it. Thank you much. Katherine Bayless: Thank you. my brain is about to, , flatline, say. Uttam Kumaran: , this was a lot. This was a lot to go through, … This is good. But very helpful. Just once, hopefully, on a lot of these. Katherine Bayless: Don’t worry, don’t be shy. But they say it takes an adult, , 6 times before they remember a fact or something that? Uttam Kumaran: Oh, , maybe, probably. , that’s why we need the AR, we could get that down to one… one or two. . Katherine Bayless: But… Cool. , thank you, thank you. , alright. Uttam Kumaran: Thank you. Glad to see you. Katherine Bayless: Thank you.