The below transcript has been auto-generated for your convenience. Please reference the source video/audio for direct quotes or to clarify any errors.

Speaker 1: (00:00)
Thanks everyone for joining us today. Uh, we're gonna get started, um, as Jeff just asked. Uh, also, we're also gonna feel questions if they're directly relevant to what we're talking about, like on on a current slide, we'll address them. Otherwise, we've got some time at the end that we're gonna, uh, we'll go back through answering additional questions, uh, any comments in the chat as well, but questions, there's an, uh, q and a section. Just drop them in there and we'll, uh, we'll knock 'em out and let's get going. Cool. So, again, thanks for joining us. Uh, I'm Ryan, uh, Avi, our revenue and solutions architecture teams at App State. We're an elite HubSpot partner, and I've brought two various smart gentlemen with me. Uh, do you wanna go ahead and introduce yourselves?

Speaker 2: (00:45)
Yeah, sure. I appreciate being called a smart gentleman. I'm Nick Cardone. I work on go-to-market strategy for Operations Hub, and super excited to be here talking about data management and data quality today.

Speaker 3: (00:56)
And also never have been called a smart gentleman, but I'm also happy to be here. Hey, I'm Jeff. I'm a product lead, uh, at HubSpot, part of our CRM group, and I also led a bunch of the, uh, data quality and data management work we've done in the last year. So very excited to talk about with you. And here's some questions that y'all have.

Speaker 1: (01:16)
Cool. So we've got a, uh, packed next, uh, 45 minutes or so, uh, to go through this. Um, gonna go through data modeling, um, setting that up in HubSpot, uh, thinking through setting up custom objects and, and reviewing data model. Um, some ways that you can keep your data clean inside of HubSpot. Uh, going through automation capabilities, uh, monitoring a data, and then my favorite topic, uh, data sets and, and reporting. We'll go through all that. All right, so I'll turn it over to Nick Kirk later.

Speaker 2: (01:54)
Yeah, for sure. So, just like to kind of intro what we're gonna be talking about today and give you a peek behind the, like, the product philosophy curtain when Jeff and his team and the rest of the folks here that worked on this project we're thinking about data quality and data management. Uh, in c R m, they eventually came up with kind of like a two-pronged approach, and I'll let Jeff chime in if he has any extra thoughts here. But we thought about it in, in two ways. So number one is preventing bad data from getting into the system in the first place. So if I'm like protecting something valuable, I've got like an art collection, I'm gonna put up cameras and maybe like some alarms or something like that. But I'm definitely also maybe gonna have like some guards patrolling around inside because we know that like, bad data's gonna get into the system regardless of what we do. So we also want to give folks the ability, um, to monitor and fix data issues over time. So that's kind of our two-pronged approach. Prevention and then treatment, um, if and, and where we have to

Speaker 1: (02:44)
Thumbs up. All right, cool. So we think about, uh, just data modeling in HubSpot. Uh, definitely wanna prevent any informa any bad information from getting in. Um, I know a process that we always go through is just how does your, uh, business actually work? How do you sell? How are you marketing? Uh, what types of information you need to track for customer service, uh, and understand the, the major buckets of of information. So HubSpot's got a really great, uh, tool to allow you to go and visualize not just standard objects, uh, but any custom objects that you've got in your portal. So within your instance, if you've got things like subscriptions, maybe you wanna track subscription history, uh, maybe you're syncing information over from, uh, from an e r p. An example, I'll, I'll touch on a couple of times. Um, tracking sales or purchase orders, a as an example. And you wanna be able to actually manage that directly in HubSpot. You've got a nice, clean way to go and visualize how everything is connected, allows you to kind of tell the story a little bit better, uh, and actually see what holes you might have, especially when you're looking at reporting.

Speaker 1: (03:48)
So, e r d, any, any relationship diagrams are, are always a lot of fun. Uh, I know I, I enjoy them. It's probably the nerdiest statement I've ever made, but it's something that, uh, that I think we, we spend a lot of time on, and it's really the foundation to everything in your crm, right? So how, how things are, are set up, how they talk to each other. Uh, if you are, um, if you're managing any type of like, like I mentioned, subscription or payment, uh, any custom objects that you're bringing in, being able to understand that you've got, you know, uh, a subscription now, but you wanna track changes, updates, uh, major, um, additions to it, you'll probably need an additional object. So being able to just have a visual way to map that, tell the story, and then also understanding connections. Very often we go in, into a report may look, uh, for something specific and not understand how to be able to pull that. So being able to pull this up and say, great, I might need another object in here to track, uh, historical, or I need a bridge between these two objects to get the reporting I want.

Speaker 1: (04:54)
So cleanliness, uh, there's probably a good saying in here. Clean data, I don't know. I was gonna say an apple of day keeps the doctor whatever. Clean data definitely prevents a lot of headaches. Uh, so HubSpot's got some, some really great ways to, uh, keep, to get it clean, uh, prevent bad data from even coming in, and then of course, keep it clean on the, on the back end. So I mentioned the example of, uh, syncing information over to an external system. Uh, frequently we're bringing in, um, things like IDs or dollar amounts. So you've got property, uh, validations that can help you, uh, format that data properly, prevent special characters from coming in, um, make sure that the data that's, that's getting sent over as an example is accurate. So you don't want, um, as an example, dollars that go over with, you know, three decimal places.

Speaker 1: (05:45)
Something you wanna prevent with the validation. Uh, maybe you have some workflows, custom code action that's relying on, uh, a string of text. Also wanna make sure that it doesn't include special characters, all things that you can, uh, limit with validations. So something else, maybe like an order number, text property. So you're passing order numbers over, uh, or payment amounts as I mentioned. All things that you can set up. The common example, uh, that, um, that I actually use for, for validations are, uh, length of inputs. Um, so first name, uh, order number, if, if, uh, just to go back to that one, it's 10 digits every time in, in your system. You wanna make sure we've got a minute, a mac set up. People can't as gonna say fat finger, uh, that they can't, they can't mistype, uh, certain things across. Easiest way to break an integration, especially if it's custom to a system, uh, is to put too many numbers. Too few numbers, or you throw an A in there when it's looking for all, for all numbers. Um, so another thing that we'll talk about is unique, uh, value enforcement, but you can also combine both of them, um, which we'll talk about in in a second. Um, Nick or Jeff, anything on you guys wanna add on validations?

Speaker 2: (07:09)
No, I think I was gonna tie them to unique value enforcement specifically around like string length, but I think you're about to touch on that. So

Speaker 1: (07:17)
Go ahead.

Speaker 2: (07:19)
Oh, me takeover? Yeah, so, oh no, I'll,

Speaker 1: (07:21)
I'll talk, but

Speaker 2: (07:22)
Yeah. Yeah, I mean, I just, that way it's perfect. Yeah. So we think about unique value enforcement for certain types of properties. We can require that, uh, two records do not share the same value. So a cool example, and like a common example we've been seeing is like vin. So via, uh, vehicle identification number, if you have like a car's custom object, you could enforce both unique value as well as the length of a strain to make sure that I'm both collecting the right amount of characters and that the VIN for one vehicle in my c r M and my car's custom object does not match, uh, another VIN for another record. So almost like kind of de-duping kind of data quality, um, but a cool use case for property validations as well as unique value enforcement.

Speaker 1: (08:00)
Yeah, and if you think about custom objects too. So if you have a person and a person has multiple cars, or if you're, uh, let's say you're an I S P, right? And you've got a person, uh, and they've got, uh, multiple locations with different modems that you need to, to pings use case we actually just worked for recently. Um, so they've got, you know, a modem at home and that has, uh, a unique ID that you need to store in, in HubSpot, they've got another one at work. I need to be able to ping various different routers. You gotta make sure you got the right ones. So you don't want, you wanna make sure that you've got some level of unique value to be able to go and identify those. Um, just any and anything that's syncing between systems, you're storing external IDs in there. If it's unique in the other system, it's gotta be unique in HubSpot as well. I think we talked about, uh, a a lot of the different, uh, various use cases, but things that you can combine, uh, definitely combine together. Um, and we pretty much covered that. So a question here, oh, we'll actually come back to that. Okay. I'm gonna turn it over to Nick, uh, to talk a little bit about automation and we'll save some of the questions per the end.

Speaker 2: (09:11)
Awesome. Let's do it. Um, I am very excited for this because it starts to get into ops sub territory, which I am very, very excited about and passionate about. So, um, we talked a bit about setup. We talked about enforcing, um, quality data and, and data rules on the way in. Um, now we're gonna kind of move over into that second prong, that second pillar that I mentioned before. Uh, and that's about monitoring and cleaning data once it's already in the c R m. And the first thing we're gonna talk about is data management, data quality automation. And we've had this feature, so for the format, data action and workflows, uh, for over a year now, launched with Operations Sub Pro in April of 2021. And we've always been able to do some pretty cool stuff with it. But recently we have added a formula mode to this format data action that you can see in the screenshot that opens up some pretty cool new use cases.

Speaker 2: (10:00)
So, to set the scene here, imagine, uh, a use case where I've got a contact or multiple contacts filling out a web form, uh, online, and I want to clean up some of their data, uh, as they input it. Cause I don't know what they put in, I don't know if they capitalize their first name. I might want to standardize a date that they gave me, uh, to be month, day, year instead of day, month, year, those kinds of things. I could do that if I were to turn off this custom mode toggle in the screenshot. I've got an awesome dropdown of actions I can apply to that property that's associated with that record and then clean it up as I go. So right at the source, right as it gets into my C R m, I'm cleaning it and it's ready to be used maybe in like a thank you for submitting your form email.

Speaker 2: (10:38)
I make sure that the n in Nick is capitalized for a nice, like first touchpoint with that, with that customer. But what's really cool now is, if you can see here, I've got this custom mode toggle turned on. So on like the top right portion of this screenshot, uh, we have like an expression builder now, and I can do some cool stuff with this. I can still do all the, the baseline formatting, like I could, you know, use the capitalized function to capitalize the first name. Um, I could use the date standardization function to standardize date. Um, but what I can do now is actually start to get into like data transformation and data creation territory, where before we were doing formatting primarily, but now this formula builder allows us to build more complex expressions that can actually create new properties and do a bunch of cool stuff. A couple cool use cases here, um, that we've been seeing from customers and even internally recently. Uh, I'll actually, I'll call out too. So you can actually see in this screenshot we've got a plus time function being used. So we're using an existing, uh, property. I think it's contract start date. It's very small on my screen and I'm looking to my right here. Contract start date. Yeah. Can you actually,

Speaker 1: (11:42)
Can you zoom that? Can you zoom that in?

Speaker 2: (11:44)
I can try. This is gonna get messy cause I'm not good at it. No, you'll have to trust me on this one because I don't know how to zoom very quickly here. We'll

Speaker 1: (11:51)
Circulate this around

Speaker 2: (11:52)
Also. Yeah. Afterwards. Oh, that's not gonna work. Okay. It says, and you can trust me, I'm not lying to you here on a recorded webinar. Uh, it's contract start date and I'm adding a length of time to that date. So imagine HubSpot, right? To use ourselves as an example. We typically have like one year contracts with our customers, but we like to start a renewal conversation a little earlier just to make sure we're on track, vectors are aligned and we're in step of the customer before that contract anniversary hits. Um, I could add like 10 months to that contract start date, and then use that new renewal conversation start date to power automation, maybe a task for a, uh, success manager or an account manager. Uh, I could use it to build a list. I could use it for email outreach. I can create new data points using this format, data action, um, that help me build, um, just cooler or more streamlined processes downstream.

Speaker 2: (12:42)
Uh, second example, I'll call out here, it's not visible in the screenshot, but we have a concatenate function. And one of our solutions engineers internally actually had a cool video that he, uh, shared a few weeks ago where he used a bunch of different address properties and concatenated them into a string with some fixed str, uh, text properties and was actually able to make a hyperlink, uh, for Google Maps out of an address. So if you have like a service team working in the field, or you just wanna see where a customer's located and have like a easily accessible maps link, you can do that with a concatenate function. So again, we're not just cleaning data, uh, we're creating new kinds of data and opening up new use cases, um, with this tool. I'm gonna pause if you guys have anything to add or if there's anything in the chat that I, uh, that's topical that I may have missed.

Speaker 2: (13:30)
Cool. Um, pushing on, so we talked about formatting data that's automated. Um, we've got this spicy new tool though that I love and it's called data quality automation recommendations. And we are actually now using, uh, AI to detect issues with data formatting and data quality in the C R m and, uh, like one click fix it. So I'm actually just gonna show this screenshot on the next page. Um, we've got this tool that's been around a while, it's been kind of hidden, this fixed formatting issues tool. This is available in ops sub starter, and it's accessible from the contact and company index pages in the CR r m. So if you were to like actions, uh, fixed formatting issues, you'd be taken here. But what we've just released a couple months ago in Operation Sub Pro is this automation button in the top right of this screenshot.

Speaker 2: (14:15)
And if I click on that, and again, I apologize, this will be hard to see because the screenshot's a little small. HubSpot is checking default properties for issues with capitalization, potentially combined fields where there shouldn't be combined fields, and then issues with spacing and punctuation. So it would tell me, Hey Nick, you know, we scanned your C R m, there's 127 contacts with a lower case first name, or there's spacing issues in 300 of your email addresses. Or maybe I've got a bunch of first and last names that have been combined together. Do you want to click this button and fix that now? And I'll say, sure, I'll click that. So it's gonna fix all those issues for me. And then moving forward, you can see up here, um, it says two automations on that fix is gonna keep running in the background to make sure, uh, that those issues don't pop up again.

Speaker 2: (15:00)
So once we've found an issue, we'll kind of keep this background sort of shadow workflow running in the, in the c r m for you to make sure that issue doesn't happen again. So think of this as like a way that we're kind of pushing HubSpot, c r m, uh, to be a self-healing c r m if you think about it that way. So again, acknowledging the fact that we're human. Uh, it's a big c r M platform. Bad data's always gonna find a way in. We're giving folks the tools they need, um, to proactively address data quality issues and automatically address data quality issues as well.

Speaker 1: (15:29)
Yeah, I think, I think it's a nice balance between having too much validation upfront, uh, anyone who's been an end user in a system having to go through, uh, I know in, in other systems, not in HubSpot, uh, but having to, having to guess, uh, what you need to do or just having a hundred different rules if the follow to just enter a small piece of information, uh, can be a pig deterrent, uh, to spending time in a crm. So for users to be able to have some validation upfront of like the really critical piece of info. So especially if something's like sinking over, uh, or pushing through an external system immediately. Um, but then after that, the the little sort of nuanced things that aren't gonna destroy anything, they'll just maybe make some bad data you have to clean up later. Having something that can go in and clean that up is, is a nice, nice addition. So it's, I mean, I know feedback that we've gotten.

Speaker 3: (16:16)
Yeah, well, I think one thing I would add here, Nick, is the, the AI models, which is, it's mostly machine learning that's running behind this stuff. The reason why we went to automation was because we had an over 99% acceptance rate on our suggestions. So that told us that we were, uh, our models were, were developed to the point that this could really save people some time. Um, I would share that it's even good with, uh, names that you might think, uh, are really difficult to have blanket rules around, like, uh, I don't know, McCormick is jumping to mind. There's a capital C in there. If I choose, uh, make sure that my names are capitalized, is that gonna totally wreck havoc on those kinds of names? And actually the model recognizes them properly. So the way to think about all this is we are trying to figure out one of those things that are taking our, our CR admins and data folks a ton of manual time to like, review and fix.

Speaker 3: (17:13)
And it takes them away from the stuff that they would love to do that would actually move the business forward, but they're stuck making sure that their data is, is usable. There's also a question in the chat, um, from Rihanna about, hopefully I'm saying your name right, about how to I, if you could fix some of the issues, but not all. And I think, Nick, if you want to expand on this, fine, but we, we offer both, um, like observability for the issue. So we'll tell you if there's an issue and you can choose whether to accept it or, or reject our, our, um, our proposed fix, uh, for each of them. And then we, we also have as, as Nick is kind of showing in this screenshot, if you can make it out, uh, that you could turn on any of these rules or not. Um, so that if there's, so that if for example, you, you feel really strongly or you feel like HubSpot's handling capitalizing last name, you can turn that one on and we'll change them all for you. It's also running on the workflow system, so you have full historical logs of all the changes that we've made, but you don't have to turn on all of these rules, uh, and, and have the HubSpot just run away with your data. Hopefully that's an answer to your question.

Speaker 2: (18:29)
Yeah, it's a good call out. If I were to move this inset to the left side, you'd actually see an accept or reject button, so you could go through one by one, um, and accept or reject those suggestions. Cool. Um, let's talked about automation. So the format, data action automation recommendations, um, now we're gonna move over to monitoring. So our idea with operations sub at like a high level, even zooming out from, from just data and data quality and modeling, is to take folks that are operations professionals that are admins that maybe just are responsible for those kind of tasks but aren't officially operations folks maybe yet. Um, we want to take them from being reactive firefighters to making them proactive, maybe more of like a fire inspector or a fire marshal, uh, than a firefighter. And that's the idea kind of behind monitoring with data quality.

Speaker 2: (19:20)
And I'm really, really, really excited to introduce, um, and talk about the data quality command center. And this is another new feature that we, uh, we announced at Inbound a few months ago, uh, in operations Sub Pro, but think of this as a home base for data quality and observability. So it's gonna help you monitor your data and its health on an ongoing basis. And kind of going back to that firefighter versus fire marshal, let's identify problems before they have any negative repercussions. Um, so here's a screenshot. So if I were to drop in under reports data management and then data quality, um, I'd be able to see a couple different things. And I've actually screenshotted two of the three tiles available, the information tiles available in the command center. Um, on the left we have properties on the right, we have records. And the first thing you'll notice is that you get a good sense of trends in this data over time.

Speaker 2: (20:08)
So on the right we just talked about formatting issues, and that'll be this orange bar here on the portal. So if I were going in and fixing these things regularly, if I were kind of preventing formatting issues using validation rules or the format data action, I'd ideally see this dropping down over time, which I do. I could even maybe detect a spike in this if all of a sudden, you know, I've been doing really well, like this blue line with duplicates and I see a spike, you know, some couple weeks ago I could go back and say, Hey, like, what caused that? Did we, you know, install a new integration? Did we change a process, build a new web form? And I could kind of get a sense of what's going on. Um, on the left we have property information too. And what this is gonna help you do is track duplicates, empty properties and unused properties.

Speaker 2: (20:47)
And if I were to click in, there's actually a, a link a level deeper than this, I'd be able to see fill rate for properties. I'd be able to see a list of all my properties by object, I'd be able to see where they're being filled from, uh, so different fill sources. And then I'd even be able to see what, uh, downstream tools, those properties are being used in, like data sets, lists, workflows, that kind of thing. So we give you like a home base here just to keep tabs, um, on how your data model is being used, um, and just get a sense again, uh, of any issues that might be popping up before they become big fire drills, uh, for you all to take care of. Uh, on the right. We also have links here into the formatting issues tool that I talked about before.

Speaker 2: (21:26)
And we have an a link back into the duplicate manager, which we haven't talked about, but that's another tool that exists, um, that helps both with individual duplicate detection and merging as well as bulk, uh, duplicate merging, uh, as well. Last thing on this screen that is unfortunately below the fold and would not come up in my screenshot, uh, is a data sync health tile. So data sync, uh, that's a kind of a collection of integrations built by HubSpot that exist in operations hub free and starter, uh, each individual app, if you were just to have, say, operations starter, it would show you a bunch of information about the health of that sync. So any potential issues, you know, duplicates, just stuff that might be happening that in, in that integration that we don't want to have happening. Uh, but the tile in the command center is actually gonna show you a summary of all the different data sync integrations that you have in the health of your integrations, um, at a high level. So again, just kind of collecting all this data that may have existed in five or 10 different areas for properties, for records, uh, for integrations, and bringing it all into one command center, uh, for folks to use.

Speaker 3: (22:27)
Yeah, that's ex that's exactly right, Nick. Like bringing it all together was the point here. The, my dream is that you're coming in with your cup of coffee, you sit down in the morning, you're taking a look at this screen, and it has all the data points you need to know, do I need to prioritize some kind of work across the C R m because we have bad data coming in, or we have some kind of integration that's not working or something. Or can I get onto the strategic work, the new insights that like the sales leader is asking for, or that report for the board meeting that my c e o is asking for. Um, and, and to, like, before we launched this, that visibility would've taken, you know, 15 different clicks across a bunch of different apps or even exporting the data in the case of properties.

Speaker 3: (23:15)
So it was like really hard to get that insight. So there's a, um, a lot of stuff we want, we want to do with, uh, this particular tool. One thing that I would call out that we are planning to do, so I feel comfortable saying it, uh, today, you would have to actually come in to see if there was an issue. Uh, a much better system that you're probably thinking of at home is what if it could just send me an email if there's some kind of thing I need to go take action on. So we're, we're gonna do that as well. You need the observability, but it would be great if you had proactive monitoring and we're building those, those systems right now.

Speaker 1: (23:51)
Sure. Very reassuring for anyone here who's, who's an admin or has some level of admin oversight, uh, the last thing you wanna be doing is spending all of your time caught up in which I, it's ultimately critical. A lot of people are probably responsible for it. Um, the quality of data, uh, cause if it's not good reports, as an example, may not be that great, uh, may not be accurate. Um, but just one more thing to, to make their jobs easier so that you can actually go and do the more proactive things that, uh, that people actually wanna see. This is the kind of thing you don't get credit for, uh, for, for doing. And if you don't do it well, it's visible and if you do well, it's not, it's invisible. Uh, so I think this obviously takes, it takes you a long way to, to make sure everything is looking great.

Speaker 2: (24:35)
Yeah, I love that. And we wanna, we wanna like give, we wanna treat operations and admins and, and data leaders like first class citizens. I feel, I don't know how you guys feel, but like sales gets their time in the sun. Marketing gets their time in the sun service, but like operations is just like the Batman of the cr r m like in the shadows, like, I don't know, jumping around. They're

Speaker 1: (24:53)
Bad, but you don't get credit when things are good.

Speaker 2: (24:55)
Yeah. So like we, we wanna give them a home base, we wanna make sure that like you all are recognized. Like this group kind of is recognized as like a, a critical part of, uh, of the, uh, of the CRM as well and of the team as well. Cool. Um, last bit of features that, or last set of features here that we're gonna run through. We're gonna talk about reporting and data sets. We talked about setting up a data model, keeping data clean, the importance of all that. But like, ultimately we're trying to do stuff with the data, right? So we want to build automation, we want to have lists, we want to create good, uh, content for messaging. We want to just have this stuff to do things with it, right? And like one of those big things is reporting. So we can think about now we have good clean data in the CR rm, we've got a great data model that we've set up.

Speaker 2: (25:37)
Let's take it to the next level and do some really cool reporting, uh, using data sets and operations hub enterprise. And I'm gonna introduce kind of two key ideas and two key, uh, benefits of data sets now. And those are reporting enablement, and I'll explain what that means in a second, as well as customization, which I don't think needs that much of an explanation. But reporting enablement is this idea that, um, some of your teams, so if you're an operations person, some of your internal teams, your customers, right, they are gonna want to go a level deeper than your standard dashboard, right? You, you might have a sales leader, you might have a a, a frontline, like a a go-to-market team that just needs to, uh, build a custom report, but maybe they don't need to see other teams' data or they don't want to be bothered with, um, going through the c r M to find all the properties and objects they need.

Speaker 2: (26:23)
We can make that whole custom reporting process way, way easier for them with data sets and operations of enterprise. And I'll elaborate on that, uh, on how we do that in aac. Um, customization is also super important too. So if we think about taking the idea of data transformation that we've been talking about, um, with the format data action with data cleanup and data quality and bringing that to reports. Um, with data sets, we're unlocking customization within reports that allows you to transform, um, and just generally customize, for lack of a better word, the data that you're gonna be using in those reports. And if you see in the screenshot here, I'm actually building a dataset and I will sort of frame data sets as like a meal in a box, almost like a HelloFresh type situation for reporting in HubSpot. So instead of me going to the store, picking out my own ingredients, deciding what to cook, all that kind of stuff, someone in this data set is picking out my ingredients.

Speaker 2: (27:16)
So my objects in my properties, they're customizing it a little bit, maybe with, you know, different names for properties. We're using the formula builder and functions to create new blends of things, new, new data points. And then it's all gonna be packaged up for me in a nice little link inside of the customer report builder. So I know that, hey, if I'm on sales leadership and I want to check out my monthly reports on deals or whatever it might be, I'm gonna click on sales leadership dataset and I might just see these four properties here, my report builder instead of the potentially hundreds of properties that I would have in the entire c r m. So that's sort of that idea of reporting enablement. So as I creating this data set, drag properties from the left to the middle, I'm sort of choosing what my users are gonna see in that next step when they go to build a custom report.

Speaker 2: (27:58)
Uh, you can also see this, uh, like this formula builder over here. It's like another expression builder similar to what we have in the format data action. But I can do a bunch of cool stuff here. And we are basically allowing you to create calculations that are going to be used in custom reports. And the super simple example I've coded out here is just multiplying deal amount by 12. So maybe I track revenue in my deal pipeline in monthly revenue, but I'd like to build a report that looks at annual revenue. That's fine. Let me just multiply a deal amount, my monthly deal amount by 12, and now I've got annual revenue. And I could take it a step further, even some other use cases for this. I could use our function library to maybe do custom, uh, date diff calculations between any two dates on any two objects in the C R M I could use conditional logic within the formula builder may be to calculate a variable commission or to categorize deals based on size and products or categorize accounts based on different risk factors and report on the health of my customers at a high level.

Speaker 2: (28:56)
Um, and I can just generally, I've kind of been alluding to it, but I can do cross object calculations in here. So I have contacts and deals, I have also two other sources in this report. I could blend properties and data from all four of those objects if I wanted to into one calculation, which is not possible anywhere else in the platform. So I'm very, very excited to have that in data sets. That's a ton of power that we unlock, uh, with this formula builder and function library inside data sets.

Speaker 1: (29:20)
Yeah, I think Jeff just added in there. Definitely the, the first thing I think of when think of data sets is the cross objects. So being able to do some level of, um, of calculation of sifting through data. So if I've got, you know, the deal amount, you know, as an example, if you're someone that recognizes some revenue at closed one, and then there's some level of like onboarding that can recognize additional revenue, uh, if you're, you know, a services business has has something like similar, you can actually add those two together, store them, you know, any, anywhere you want. Um, uh, but being able to report on that, I think is the big, the big piece. So if I wanna look at, uh, a cohort and aggregate of that information, I want saved filters. I think, Nick, you touched on that also. So I'm continually, if we've got a way that we look at our pipeline, uh, based on a particular user that only hits a certain stage, and that's something that we're gonna do repetitively, we can kind of package that up. Um, but the, the calculations piece, and I don't know Jeff if you have any other cool examples, but probably the most powerful unlock that you're gonna have a hub spot right now.

Speaker 3: (30:23)
Yeah, I mean, I would think of, uh, this might be a tough vi vi visualization without, uh, drawing on a slide, but the way I would think about data sets is you have your object layer and you have the properties that belong to each rec, like each object and the records on those objects and data sets sit across the top and it can pull data from any of those objects in the properties inside of it to create new insights. One of the more interesting ones I've seen recently is a customer that is comparing the, uh, company size data that's stored on the contact level versus the company sized data that's stored at the company level to identify places where their reps might be reaching out to the wrong, uh, prospects based on that data. Making a table like that would've been really difficult, uh, a ton of manual work before now using data sets, you can compare those values and create your own visualization that calls out. These are the, you know, dozen, uh, records we need to look more closely at because something's wrong with the data, which can lead to all kinds of learnings and just a much more efficient sales team. So it's just giving people the power to like, interact with their data and just have more impact.

Speaker 1: (31:36)
Yeah, I always boil down to like doing math at, at, at one additional level. So comparing how you did in this segment versus this segment, it's not something that you can do, uh, without, it's, you can't do it without this.

Speaker 2: (31:51)
Yeah, I totally agree. And one thing I'm excited about, we can, I actually have permission to talk about this new feature that we're gonna release here shortly, but we will soon allow you to actually map these calculations back to an actual C R M property. So up until now, for about the last year, they have lived inside the report builder, right inside a data set. They haven't really been accessible or operational or able to be operationalized in the platform. Uh, but shortly we'll be, uh, moving into a, a limited beta for the feature that's going to allow you to actually copy these calculations back into a property. So very excited about that to connect data sets and the power here, um, to the rest of the platform. So stay tuned on that. Awesome. I think that's all we have in the way of content. So I guess now we can, I'll probably stop the screen share and we can open it up to any questions that I may have missed and just general questions that folks have now.

Speaker 1: (32:42)
All right. Let's see. Anyone else has any other, any other questions that they wanna drop into the q a section? I don't see any other open. Give everybody a minute and just as a, as a follow, if anyone has any, uh, specific questions, um, whether it's about just setting anything, any of these up, uh, something that we can feel, but also I think Nick and Jeff would love any feedback suggestions. Um, everything's awesome. I'd love to, to also see this other feature or how would I do x, y, Z, uh, I'm sure they'll welcome any emails, phone calls, visits to their house, any of those types of things.

Speaker 3: (33:27)
I, well, I'm also wondering, I mean, we launched this stuff in September and for some folks we ha or for some of these features we had of beta prior to that, who all that's on the call has already used some of these versus this being like pretty net new and you're going off of those awesome screenshots that Nick was sharing.

Speaker 2: (33:57)
I think we actually do have a question in Q

Speaker 1: (34:00)
For the, for the beta, but yeah, we do, yes. What platform unlock data enterprise in the chat as well, but, uh, love to get access to this better. And we have another question. Uh, UB can be a relatively relatively abstract pitch to non-op professionals. What would you say the best way to internally pitch the value to those that don't understand ops principles would say the biggest wins in the hub as a whole? Um,

Speaker 3: (34:35)
Yeah, this is like, uh, they're like reading our Slack chats.

Speaker 2: (34:39)
I know this, this is,

Speaker 1: (34:41)
It's, you guys probably have something well prepared for

Speaker 2: (34:43)
This. I think we do. So I think like the way I would start is to pitch the time saved. So we've talked about a lot of features today that probably save you hours and hours and hours. So you either pitch like, Hey, maybe here's what I make times the hours that , I have spent doing this stuff in the past. Ops hub pays for itself times, whatever. Um, you could start there or you could say, Hey, if I'm freed up and I'm not exporting to Excel to do data management or, you know, I'm, I'm able to run reports a little bit better, here's what else I can do. Here's the other value I could add for our organization. Uh, I think that's a good starting point. And then just tying whatever you're gonna be able to do with Operations Hub to another team. So saying like, Hey, I'm gonna build sales reps this property and it's gonna save them like, you know, a minute or something like that every time they create a contact or, Hey, I'm gonna build a cool report for marketing and sales and here's what they'll be able to do with that report.

Speaker 2: (35:30)
Um, I've seen a lot of success when we have prospects that are able to pitch successfully internally. They kind of tie it to, to other teams and what they need.

Speaker 1: (35:40)
Definitely. I I think also for, for any org that's, that's heavily invested into, into HubSpot, the just ability to unlock that. So we've already invested, you know, time, effort, and this is supporting our whole business or part of our business. Um, think of it as an, uh, an enablement. So if there's pushback on whatever cost or just need, you know, the, the amount of thing information you could supply to other teams, if you're talking to, you know, C-suite executive leadership, uh, and say, this is information that I can provide to you right now, I'm spitting it out to an Excel sheet, just as Nick was saying, with time, like spitting this into Excel, I'm spending 15 hours, uh, a week putting stuff together for, for our board meetings. Uh, this is something I would be able to do in five seconds. Uh, so that huge backlog of items, uh, of things that, uh, everyone's been asking me for, I could probably get to those. And is, you know, uh, small, small addition of ops, uh, of operations of enterprise, uh, is it worth it for that? I mean, to me it's a no-brainer.

Speaker 3: (36:48)
Cool. Went into this work thinking that, uh, sales leaders were not gonna get the story around data quality and how important it is. And I have been blown away by how many of them are trying to sell me on data quality. So I think we, I think we have assumptions around how people feel and there is even pushback. I mean, anytime you wanna spend more money or get like new things, there's pushback, but ultimately they care very deeply about the data and the data quality, uh, uh, the quality of that data that their prospects are using. For example, they're spending a ton of money on ZoomInfo. They want to know, is that stuff even worth it and can they hold, uh, their, um, development reps, you know, accountable for those prospects that are coming across.

Speaker 1: (37:37)
For sure.

Speaker 4: (37:42)
Cool.

Speaker 1: (37:42)
Any other

Speaker 3: (37:45)
Follow a couple questions in the chat?

Speaker 1: (37:48)
Yeah,

Speaker 3: (37:50)
I'd love to have specific cases Tuesday about a rolling date range.

Speaker 3: (37:58)
Oh, this is a great question from, uh, Colin. Um, would love to have specific case to, uh, show a rolling date range of a particular value. So like close rate for reps over the last 30 days. This is something that I think I, if one of our, um, reporting product leaders was on the call, they would say it's something we don't do particularly well, but we are focused on closing that gap. So I I I think data sets can get you dataset, you might be able to slice the data in a particular way to get you closer to that. I think our reporting tools are gonna take another step forward this year or in 2023, um, to make that significantly easier column.

Speaker 2: (38:40)
Yeah, I can, I can add on there. I think one thing I'll call out, you might be able to get a little closer. We do have a function, uh, now function in data sets that we'll call in today's date into the expression. So you may be able to use that to kind of hack together like a past 30 days or a rolling date range in some cases. But, um, things like close rate and just any type of like aggregate calculation like that, like a percentage, um, again, like show me win rate, show me percentage of contacts by Citi. Um, I don't want to commit us to too much or give away too much too soon, but we are gonna have a much, much, much easier, uh, solution to do those kinds of like bulk aggregate rate calculations in reporting in the future.

Speaker 1: (39:21)
Cool. Now good time one, we got one more, one more question. Stop. Enterprise typically require development resources to set up or done with hubs handle customers on which, which pieces you're talking about. I think most of it does not require development. Correct me if I'm wrong. I think the only thing well doesn't distinguish enterprise in particular, um, may wanna work with somebody, uh, if you use the Snowflake, uh, the Snowflake integration. Uh, and the other thing is just on with pro operations of pro and above, uh, just custom code actions. Um, I don't think anything else actually needs, uh, development at all. If you're familiar, um, with building formulas. If not, there's some really awesome, uh, contextual help in that, uh, in that builder in the data sets builder. So, um, should, should be able to get pretty much will value that. And developers, the ways to create buckets for the incoming data to live. Think about someone who works for Southwest Airlines, someone else who enters Southwest Airlines, Inc. This an automation to change the name. Is there an automation to change the name each time or is there an option to accept both

Speaker 2: (40:46)
Not outta the box with what we talked about earlier, like the the data formatting tools. That could be the case for where maybe a custom code action would come in if, you know, I'm getting hey, like Southwest Inc. Southwest Incorporated, I, I know I get a couple like typical responses. You could code out a custom code action that just recognizes those and then cleans up or, or standardizes the value the way that you want it.

Speaker 3: (41:10)
This might be just me over-indexing on the example, but that's gonna also get flagged by our duplicates tool. Um, if that's, if for example, that's using the same domain or there's other information that overlaps, um, we're gonna catch that on the way in and either say this is a duplicate, um, so we're just gonna automatically put the data on the record or it'll come up in a dupes check, a duplicates check as being potentially the same record. And then Rena, you'll be able to choose whether to merge them or not.

Speaker 1: (41:43)
I have a question on that, on those, those lines that I'm sure, um, everyone is dying to ask or wants to know about. Uh, is that looking at the relative similarity of the two phrases or is it the frequency that it happens and, and, and similarity of the words? So I'm thinking of a situation where maybe there's not, I don't know that if it comes in that the, they're structured the same way every time. They're different every time. Um, there's just an ink at the end. Uh, I'm blanking on a different example where like, I think a custom code action would work great cuz it's predictable. You, you know, what the difference is gonna be. But is that something that would be picked up or It depends on how closely related they're

Speaker 3: (42:30)
Yeah, exactly. And we have customers on both sides of that line where they're getting stuff flagged that they don't want to be flagged. Um, and then they're also on the other side where they are trying to figure out why these two didn't get flagged. So we're always trying to hit the right, yeah, the right spot. The last thing we want to do is, is have people in there looking at potential duplicates and finding that most of them aren't. That would be a total waste of time. Um, in this case it would be. Uh, so we have two kinds of duplicate checks. Uh, one kind is like the CRM based duplicates check where we're actually looking at the data on the rest of the properties for that particular record and saying these two, you know, for example, the domain name on these are the same, they're duplicates of each other.

Speaker 3: (43:16)
And the CRM version of it is very straightforward. It just calls it like, it see, sees it. There's another version that is the one that we provide in opsa, which is, uh, more AI based. So it's looking at all the properties on the record and looking like you were saying, Ryan, for things that look pretty close. They don't always to like the human eye look perfectly the same, but the model is picking up certain things that it thinks are are close. So in this, for this example, it probably would've caught those, um, because just the, the only difference is the word ink, which maybe the model has been trained to recognize is similar. But, uh, if the domain name for that record was the same, you know, someone's filling out a form or something like that and they used their domain name, it would've been flagged by the CRM on the way in.

Speaker 1: (44:06)
Got it. And I think outside of that, if there's some level of predictable pattern, um, that you wanted to run through, you could always use it, use a custom code action to flag something or change data, um, that's gonna fallback for any. I think a lot of the things that we talk about, uh, not, not data sets in particular, but on validations, uh, as well, if there's something that doesn't exist, you can run a custom code action to flag, uh, or send a notification or, um, any of those types of things. It's like a fallback, especially if you don't wanna put too much validation up front.

Speaker 3: (44:39)
Um, yeah, for sure. I mean, our dream would be to just give you one great way to solve every problem that our data admins have. But I think where we're at right now is we need to give you a number of different ways, one of which is gonna fit your particular problem, and then over time we'll like make all the tools smarter so we can hopefully, uh, make it a little bit easier or just Nick will do more, uh, guides, uh, that you, that you all can reference on the best way to solve, you know, a particular problem.

Speaker 2: (45:10)
Uh, looks like two more questions. Short answer to Gray's question. Yes, I think a good example of that we have, uh, the ability to do enterprise or like sophisticated lead routing and that, uh, can actually reference a hub DB table. So say postal code by rep. We could kind of do like a v look up on a db uh, db table like that to assign leads. So that's actually a great call out and cool use case. Um, last question. If multiple country properties contain a variety of country values, country code abbreviation spelled out, would this allow for pulling the latest updated field and populating with a consistent type? I'm not sure I know the answer to that. Jeff, do you have any input there? Sorry,

Speaker 3: (45:51)
I'm still trying to read the question.

Speaker 2: (45:53)
Yeah, me too. I think that it's,

Speaker 3: (45:57)
I, I think I see,

Speaker 2: (45:59)
So maybe that's similar to the Southwest question, honestly. Like if I'm getting like NC for North Carolina, North Carolina spelled out, maybe if someone's calling it NorCal for some strange reason, could I like identify all those in standardized? Maybe Daniel, is that the question? If you're,

Speaker 3: (46:14)
Yeah, so, so if that is the question then yeah, we're working on that right now in the data quality tool. That'll be the next kind of validation that comes out. Or, or just check. It's not a validation, but a check there is, uh, is specifically around locations. So we'll be able to do some pretty cool stuff like tell you if the city that, uh, the city and state that they entered in a form match, for example, which can be huge for, uh, anybody who does sales by geo. So this, this kind of question I think fits in that, in that space. Um, Daniel pretty well.

Speaker 2: (46:55)
Cool.

Speaker 1: (46:57)
All right, well thanks everyone for your time. Uh, thanks for coming. Uh, thanks Jeff and Nick, uh, for your time as well. Good to have you. And um, what what'll send this out, follow up everybody as well so we can watch the recording along with contacts. Uh, so if you have any additional questions, you can route that to the right person and yeah, hope everyone a great day. Thanks guys.

Speaker 2: (47:18)
Awesome. Thanks y'all.