0:14Justin Norris: welcome to rev ops FM. Everyone today, we're going to take a deep look at how to run go to market systems at enterprise scale and even more exciting, we'll get a peek behind the scenes at one of the most iconic tech companies around HubSpot. Today's guest is rich Archibald VP engineering for GTM systems and automation at HubSpot. We're among other things. His team manages HubSpot's own HubSpot implementation, FinTech and data engineering. Prior to HubSpot, Rich held engineering and business systems leadership roles
0:45at companies like Intercom and Facebook. And Rich, I just have a ton of questions for you today. So we're going to dive in and I'm really appreciative of you coming on the show. Rich Archbold: Thanks Justin. I'm delighted to be here and as I said offline, thank you so much for all of the content you put out. I am an avid listener of your podcast and just find it so helpful. Justin Norris: Thank you so much. To start with maybe just a little bit about your background. You've written about how you started out, in more of an infrastructure backend engineering type of role.
1:15Transition to business systems. I'd love to understand from an engineer's perspective, what kind of shock to the system did that provide? And what were some of the learnings that you went through as you entered that field? Rich Archbold: so as you said, I started off in backend infrastructure engineering and I think. I always loved solving hard, gnarly scalability and operations challenges and didn't really mind if other people didn't find them glamorous or if there wasn't a shiny front end set of systems you needed to deal with.
1:48So perhaps it was kind of no surprise that I Eventually made the move into business systems and go to market systems. It started out at Intercom, where at the time we had a homegrown billing system that operated as an Integration layer between Salesforce and Zora, and this was maybe 4 or 5 years ago. Now, when, I think we were making so many changes to pricing and packaging at the time, and maybe how we had implemented some of our usage based billing systems,
2:21uh, intercom just really didn't. integrate well with Zora at. we were having a lot of operational challenges, scaling Zora and our CEO at the time said, Rich, you seem to be good at fixing backend systems, any chance you could jump in and take a look at our billing system. And that, really was the beginning jump off point for me, diving into billing systems and realizing that these are like fascinating, very difficult, Computational problems.
2:52there's real magic and innovation in integrating between your pricing and packaging and somebody else's product catalog and billing systems and learning how you need to match up data models and. found it absolutely fascinating. I incredibly difficult and just had a real appreciation for the engineers and leaders who work in this space that there's, there's just really valuable craft. I think, while the engineering work was difficult, the more difficult thing was
3:24stakeholder management and understanding who your internal customers are really being on that point of intersection between oil and water, almost where you have, salespeople and sales that really work to the quarter and have quarterly targets and have, like really hard deadlines they need to hit and really, meaningful performance based compensation that if they can't hit it, it's really materially impacts them and engineers and product managers
3:55on the other side who operate to like a much more gentle or longer term cadence and understanding the friction that can happen when people don't. really understand each other, and, don't really have the right empathy for each other. And that kind of human relationship building and empathy and understanding of. Each side of the coin, I found every bit as interesting and humbling as, the actual engineering work. And, over time, I grew to love the space.
4:26I like the. Breath of different people you get to interact with and learn all different parts of the business. and progressively, then I moved from leading our, teams in, owning, billing to when intercom hired its first CFO. I switched to report to the CFO to build out are first ever centralized business systems function and run go to market applications, which had salesforce marketo and intercom on intercom, our billing team and also
4:56took over our data engineering teams. And again, that was just, So many new things to learn that I'm really kind of motivated by learning, and I love to learn new things. Some people would say they have 20 years of experience. And I go, is that really 20 years? Or is it the same 1 year, 20 times over? And I try and kind of push myself to have a new year of experience every time and getting to report to a CFO and having all of, like, really Senior finance execs in your peer group was just a gift.
5:28Fascinating learning experience. Justin Norris: And you alluded a little bit to the contrast between what you were doing in business systems and folks on the product side, where you've got engineering teams and product management teams. And in that structure, every team's a little bit different, but usually product management is kind of. The interface between, the customer and business needs and then the engineering team who's responsible for how do we bring this to life in a technically scalable way. And then in business systems, you kind of have that stripped away.
6:00You're sort of judge, jury, and execution are all wrapped up into one, so to speak, and that you've got to interface with the business. You've got to define your requirement. You, don't have often that separate product management layer. Was that. Disruptive for you at all? Or did you feel a gap not having those functions surrounding him? Rich Archbold: Yes and no. You're right that it's totally different. And even if you do have PMs, while those PMs share some common skill sets, how they approach their job has to be very different.
6:31That if you're a PM, um, In core products in HubSpot, you are building a product that is a good product market fit for millions of SMBs, and if your product doesn't match 100 percent of the needs of the customer, they still have the freedom to leverage their own engineering team to buy or integrate a different solution. You don't have to do the last mile Customer installation or professional services or customer support.
7:03You've dedicated teams for those. So it's really a, very specialized, very long term, strategic thinking position, but when you operate as a PM of a business systems team, you have one customer, usually an enterprise customer rather than an SMB customer who you have to build 100 percent product market fit for. You need to do all of the professional services, all of the customer support. You need to do absolutely everything yourself.
7:33You find it very hard to say no, because that customer you have a real personal relationship with and that customer generally. Doesn't have the luxury or you actually really don't want them to have the luxury of going out and hiring some other consultant or hiring some other people because that leads to fragmentation, collaboration tax, that's kind of how, you know, you're failing. It's not whether your customer is like telling you, you're doing a great job. It's whether they've gone off and spun up a shadow it, function or. or. not, so I think the 2 roles are, I mean, obviously, there's a lot of
8:05commonalities, where, you aren't doing a Coupa implementation project. What you're actually doing is a procure to pay efficiency project. And so you need to make sure that you understand who is the customer? what is the business problem? What is the ROI? and don't get over focused on. Implementing any 1 particular technology or other. So I think that's where those are some of the, I mean, you need to be really great at communication and have great empathy and great, problem solving skills. But, yeah, I think the 2 roles are quite different in the go to market space.
8:38I think, that's really what we look to have our best BSAs, step in and kind of perform that role where the best BSAs are, part project manager, part systems administrator, part systems integrator, part customer support, part solutions engineer. And, BSAs I've worked with, they really, Look and feel like, a very operationally focused, product Justin Norris: And just in case that acronym isn't transparent for anyone, BSA equals business systems analyst.
9:10Rich Archbold: Yeah, business systems analysts. Like, I've seen multiple archetypes where some business systems analysts will lean more on the, product owner or Product manager, almost like a program manager archetype, and some will lean more towards the systems engineer, solution engineer, integrations engineer, where they're really adept at being able to configure or maximize a particular service or leverage, integration technology, like Workado.
9:42So they're almost more like solutions. engineers, and I think it can be very difficult when you're in a very small environment, you want your to be able to do all of those things. But the bigger you get, it can be sometimes difficult to have, like, you want to have some amount of Swiss army knife, business systems analysts that can really jump in and do anything. But I think once you reach a certain point, that's where you need to think about, do I actually need, A Swiss army knife, business systems analyst, or do I need a dedicated internal go to market product manager.
10:12And I think at HubSpot today, we're, I'd say we're almost like 90, 10 or 80, 20 in favor of each of our individual teams are generally led by a product manager. And we do have some amounts of amazingly fungible, business systems analysts who are really are the. glue in a lot of different projects where they're able to participate in many different ways at many different levels, depending on the project that is at hand. Justin Norris: so maybe digging a little bit deeper than into
10:44your team and its structure, where does it sit within HubSpot? Is it rolling up into like an it function? Is it connected to different functional areas that you support? How does that structure work? Rich Archbold: So, today we roll up through the, Product organization, so the R and D arm as such. So I have a product partner who rolls up through our head of product and I myself roll up through our CTO. But for me, I always think I have many bosses that I, like, I feel equally
11:19accountable to our head of sales, head of marketing, head of support, our CFO. And, while the CTO is my manager, it's great that, like, I certainly report to him from an engineering quality and from an engineering execution perspective, but I need to make sure this is such a highly cross functional space with, so many different, uh, leaders involved that you really have to consider yourself.
11:49You are accountable to the, companies go to market leadership team. Justin Norris: And when you say you have a product, partner, tell us about that relationship. Like, how does that work? Rich Archbold: Yeah, so I have a product partner, Kitty Chu, who is based in Seattle and Kitty and I, so our mission is to help bring HubSpot's go to market strategy to life. And so the, 2 of us will partner with each of the various different go to market leaders to.
12:20Understand what the company's go to market strategy is, how it's going to be evolving each year and then use that to break down. what are the big rock projects? We need to get our teams to execute in each of the different parts of the business. As I said, each of the teams generally have, uh. Product manager, uh, designer and an engineer and all of the product people roll up through my product partner and all of the engineering people roll up through me. Justin Norris: Got it. So it really is structured in a lot of ways, like an externally facing
12:55product team would be with engineering product Rich Archbold: very much so. And we share a lot of the same core values around solving for the customer and making sure that we're, thinking long term and that we're building really high quality into the product services, workflows and data that we generate. Justin Norris: So kind of traveling down the tree, if you like, of, your org structure, beneath you and kitty, how do you structure those different groups or pods? Are they around systems or are they around the go to market teams or
13:28what does that structure look like? Rich Archbold: we try and set up our teams to, try and generate relatively clean relationships with each of our key stakeholders. So there's a marketing focused team or a marketing focused group. There's a sales focused group. There is a customer support and success focused group. There's a partner focused group. And then we have a shared kind of CRM, data, focused, function. As well. And then as I said, we have our FinTech, function, which is kind of like our own homegrown, uh, pricing and packaging, CPQ billing invoicing and,
14:03some of our finance systems, functions. And all of those individual groups or teams at various different layers would have a product management leader, uh, design leader, and an engineering leader. And we have this concept of what we call interlock. So this would be, a set of regularly occurring meetings where you'd have members from, like, the go to market systems function and their relevant counterparts in, rev ops or, the various different in field
14:35leadership functions, like a sales leader or a marketing leader or, uh. support leader, and we, kind of realized that this is like a fast moving space. It's very cross functional. we have a concept of hashtag 1 team that even though we all roll up to different leaders, you have to think of yourself as like 1 cross functional team. It's up to you all to make the decisions and make sure that there's minimal friction. Between Rev Ops and engineering or Rev Ops and product or sales and Rev Ops, we
15:07try and think of ourselves as just one, cross functional team, working together to help bring our go to market strategy to Justin Norris: life. And in a smaller organization, a lot of the systems ownership, you know, usually lives within RebOps or marketing ops, you know, this functional ops teams. Um, How do you think about the roles and responsibilities with your group and then the rev ops function? Like what do they do? They're a stakeholder of yours. I am assuming. but how do those lines of responsibility flow,
15:40Rich Archbold: that's a great question. And what I think of the Product teams that roll up through me as much as possible. We're trying to build systems and build functions that allow RevOps to have the right level of agency and control and configuration. We actually want them to be able to move fast and experiment. And so I think about it as like trying to keep the cycle times. Low and keep the agility high.
16:12So in each of our rev ops teams, we do have some that have hands on ability to make changes or have some work out our skill sets to, do some last mile Automation and, customization, know, I've been, I've seen organizations in the past where, people try and put every single thing inside of that single business systems function or single sales systems function. And I think, I've seen the cycle times be too high, around that.
16:46And I think you really need to strive to give the business the right level of agility and control over the things that are happening day to day. And if there's a small change that. Can be made, you want it made as close as possible to the person requesting it. And I think that's why we have powerful, easy to use platforms like CRM platforms, right? Like that, I, I think that is some of the value of it. So I definitely do like the. Centralized teams to a certain extent, because I think that removes some
17:19siloing and allows you to make bigger, bolder decisions more, quickly. It allows you to stay focused on customer problems rather than parts of a process. But I also think it's just really important that the business feels like they are beholden to you for every single change they want to make, that they will have the right context. you just want to be able to have the shortest possible cycle times for, the categories of change that, that actually makes sense
17:49for. Justin Norris: what you just mentioned to me addresses what I've observed. And also what I think is generally perceived as the major risk and shortcoming of the centralized team is that you essentially recreate the it dynamic where it's like. A big queue of tickets, it's sort of this fortress mentality. They're pushing away work We'll see you in three months and of course that's what leads to the the shadow it cropping up again or the consultants or we just have to do it on the side i've also seen to your point that you know inevitably folks
18:20that are coming up through operations that don't necessarily have like Maybe they develop systems expertise, but it isn't necessarily their background You Eventually when you're thinking about an enterprise architecture vision you go beyond what those folks are able to accomplish at least without Further upskilling and you really do get into a much more technical domain So it seems like you have this sweet spot dynamic setup do you feel that? This is just inevitable that, you know, once you pass, I don't know what it is, 5,
18:51000 employees for, I don't know, whatever the number is that you have to kind of have a centralized business systems function, or do you think that you can just have functional ownership of tools, you know, within ops teams in perpetuity, Rich Archbold: You know, the answer is, I think it depends, if you have a, a relatively simple and stable, business model, and you're not making a lot of changes to your go to market strategy, I think you can keep things simple and keep functional ownership, as long as you can.
19:23But I do think there's certainly a level of change that happens, a tipping point of complexity or tipping point of change where you actually can't make changes in isolation, where if we're going to start selling a new product, we need a new skew in salesforce. We need to be able to meet her. It's usage in a new, different way. In our billing system, we need to be able to account for revenue and in. In a different way inside of our Rev Rec system. It's going to mean different product usage data flowing into
19:55our customer success platform. I'm going to mean that we try and drive cross sell upsell renewal a different way. I think it's really hard to implement those type of transformational projects or really big step function projects without some sort of, central governance or, architecture and, management. So I think, yeah, you just kind of reach a tipping point of complexity where it makes sense to have a more centralized function. But, That doesn't mean that every time you want to add a new field inside of the
20:28CRM, it needs to go to that centralized team either, though, you know, I think that that's where I mean, there's some some types of changes where you really want to make sure the teams closest to the problem have the right level of agency and autonomy over solving them. Justin Norris: a bit of a, like a hub and spoke model, if you want to think about it that Rich Archbold: exactly. Justin Norris: Yeah, no, that makes makes perfect sense. Let's talk a little bit about your stack, which is kind of a fun thing to think about. I presume HubSpot's in there somewhere. what else are you folks using?
20:59Rich Archbold: definitely HubSpot is pretty much the center of everything and we're a massive, marketing hub, sales hub, service hub, and, HubSpot CRM platform. we use a lot of our own CRM UI extension capabilities in order to build the spoke and tailored experiences inside of HubSpot and customize the views we provide to, each of our different reps and our PMs and designers spend a lot of time talking to our internal stakeholders to make sure we're tailoring those correctly.
21:33one of the. More interesting ones that we've launched recently was using our own prospecting workspace to help organize the, prospecting layout and, uh, and tool tips that We provide to our reps and adoption for that. And engagement with that has been, very high from A CPQ. And, and, billing and invoicing perspective, a lot of that is custom to ourselves. And we have, uh, quite an artisanal set of, processes and capabilities.
22:04And that's something that our, FinTech organization takes care of on the back end. We use NetSuite, from a, call recording and coaching perspective. We use Gong and are, and we are big fans of Gong. We think that's a great product as well. from a. Data engineering. So from an analytical data perspective, all of our data flows into snowflake. We're, big fans of dbt and, looker and airflow from, uh, an automation and, data plumbing perspective, Yeah, curious, what else, are you
22:36you mentioned WorkAuto a few times as well. Yeah, so Workado, we're, we're massive, massive fans of Workado. And, use, that as our last mile automation. And, in some cases, integration technology. And, you know, our teams have built some really novel, Automation on top of that. And I've used it to, create some hooks into chat GPT and, yeah, just some really interesting use cases there as well. I come from, the Marketo space, you know, traditionally, although I've always been a HubSpot respecter, but the conventional wisdom, let's say
23:10on the street, at least in Marketo circles, it's sort of like HubSpot. Yeah, it's good. But you know, you might reach a point where you can't scale it and you're going to need to migrate to, and same thing, I guess, with the CRM to Salesforce. Here you are obviously a very large enterprise. Uh, using HubSpot for both CRM and marketing automation. how would you respond? Let's say if someone proposed that point of view to you, like, yeah, it won't necessarily scale with you. Like you've scaled it. what are your thoughts on them? I think that's wrong.
23:41so I've come from a world of, so at Intercom I use Salesforce and, Marketo, and I would say the bane of my life was dealing with, you know, Duplicate leads in, Marketo to, HubSpot, uh, syncs and bottlenecks and integration and having to deal with multi enrichment challenges between, Marketo and Salesforce and trying to implement ring lead as a solution to try and have, centralized enrichment and account management and survivorship. But, when, I left Intercom, I wanted to go to HubSpot because I felt like they
24:14had it all sorted out and I wanted to have one single stack that everything, was there and present with a single consistent data model that I could have marketing and sales, both used together. Didn't have to deal with any integration challenges. It was just all there natively. And now I'm living in the promised land, you know, it works. And so there's a whole degree of challenges that, I didn't have to deal with. that now I don't have to deal with because I have that seamless integration and the partnership between marketing and sales and the handle.
24:44I mean, you're working in the go to market world for a while and you go, what are some of your hard fought lessons learned? Handoffs between teams and tools is like where all of the friction is, going to happen is where a lot of the data quality issues, where a lot of the scalability bottlenecks are going to happen. It's going to make problems twice as hard. I don't have to deal with that. You know, I've got everybody in the one platform with a shared data model. They've got the same source of truth. That does, sound nice. let's think about, innovation on the platform in general.
25:17I assume you've had to push the boundaries in some cases to. You know, solve interesting challenges that you've had. What are just some cool things, I guess, that you're doing on your own platform that might be interesting to folks, as I said, probably the most fun for us at the moment is that prospecting workspace and feels like we have, you know, we're really able to, partner deeply with our reps and understand what is the absolute ideal workflow you want to have. Like we can do lots of rep shadowing and really build that empathy for our
25:48reps and go where do you wish you had? what's your ideal layout here? How would you like things to be prioritized? if you could wave a magic wand, how would this work for you? And the HubSpots, prospecting workspace layout, and then the CRM, UI capabilities give us the ability to really bring that to life. Pretty quickly and easily, which has been, just wonderful to see, I guess. I'm just trying to think of, like, some of the other things that we're that we're doing that maybe just on the, on the prospector workspace, because I'm super interested
26:20in that feature and I've, I've read a bit about it, but let me just tell you, give you my read on what I think it is. You can tell me if I'm right or not, but it kind of solves that perennial challenge of, you know, I've got one lead or contact or person record, whatever we call it within that system's nomenclature. but that person may go through a buying process or a funnel trip Multiple times, and it's very difficult to, report on that, to capture that because the data model is, kind of one dimensional. And so the prospecting workspace sort of introduces this idea of many, I
26:51don't know if you call them leads or what they're called exactly, but many different entities that can relate to that single person that represent those iterative buying journeys. is that what it is? the problem, that we use it to solve is slightly different. And the way I describe the problem is I'm a sales rep and I have hundreds of companies under my name. And, some of these companies are, already customers and I need to, Think about account management and account maintenance and cross sell upsell.
27:24Some of them are cold prospects, and some of them are warm prospects. And I've got hundreds of these inside of my capacity like that are assigned to me and every day. I need to come in and figure out what's the best way for me to spend my time today. there's undoubtedly been. Maybe a lot of signals that have happened. Maybe there's been new enrichment events, or, new activities on our marketing sites that might have contributed to events. Maybe the companies that are actually customers of mine, there's been
27:55some new usage data that has flowed into the system and I'm a rep with. to meet what should I do today? And so what we use our prospecting workspace is a way to, break up the companies segment, the companies that reps own into kind of like cold, warm, maybe hot, existing customers then we can use various different types of, machine learning models in the background. To help prioritize the activities for the rep and make it transparent to the rep why
28:30we have prioritized different things in. Different ways to just make it much easier for the rep when they come in in the morning, not to be hit with, just a search list of here are hundreds of companies that you are responsible for looking after and for them having to think about lots of searching and filtering to actually come in with a beautiful, tailored, high quality, UI that, fits natively inside of HubSpot, where there has been a lot of, you know, thoughtful engineering in the background to make sure that we're showing them the.
29:03Most actionable things for them to do and making it easy for them to understand why Justin Norris: Got it. So it's more about prioritization, sort of task management or activity management and, uh, and directing them. And you've alluded a few times to the UI customization capabilities, and I'll confess, I'm, I'm a little bit ignorant, actually, of HubSpot's capabilities in this area are actually a lot ignorant. what can you tell me? Is it akin to like a Salesforce lightning page layout or like a Rich Archbold: very much. So yeah, very much.
29:34So, so like a lot of the flexibility and customization you have with that, there's lots of kind of like low code ways that you can customize it, but we can go as far as embedding an iframe directly inside of a particular portion of one of those windows. And that can. Be backed by some of our own, fully bespoke and custom built, web applications that our own quoting experience being one of those examples of where we can embed, uh, our own custom CPQ application directly inside of the
30:07HubSpot app. one of the The great things about this and one of our, I guess, overriding philosophies is that we ask our reps every quarter. How do you feel about the amount of tools you have to use? you want to keep reps on the same pane of glass inside of the same app as much as possible. You want to have, Unified data models inside of your CRM. So this is one of those things that we consistently look at is how many different applications are we asking our reps to work in?
30:38How can we cut that down? Is there something else that we can move natively inside of the HubSpot app and that we can have the data, uh, authoritatively stored for directly inside of HubSpot as well? because if it is a rep productivity is improved. The amount of context switching they do is minimized, but also having that data, source of truth inside of the CRM in a nested custom object format allows all of our work out of folks to generate additional, last mile.
31:09Automation and, customization in, um, kind of safe and scalable way then as well. Justin Norris: Let's dive into that one a little bit. I'm also a big fan of work. Um, how do you think about it from a roles and responsibilities point of view within the stack? Like when might you use HubSpot's internal automation capabilities, if at all, versus saying this is where we need to layer in a work order recipe to achieve a goal. Rich Archbold: so I think that work was a great tool for ops teams. To use to automate their workflows, I think when you're doing anything
31:44that HubSpot customer scale, that's actually going to scale with the amount of customers that we have. It's kind of hard to use, uh, work just from a cost perspective. If you're doing something that's going to be running, you know, hundreds of thousands times a day, growing at scale with our customers. And those are the things that, Generally, we would use more scalable software for. I also think you get beyond the level of complexity. Like, Ricardo is great for automating things, but it's really hard to do.
32:15Unit testing or integration testing, there's kind of like a size of complexity that once you actually go beyond it, maybe you want to use some more traditional software engineering processes and best practices for, but I think it's a great tool for ops teams to use to. Automate their processes and yeah, if you're operating at a smaller scale, I think it's fine. I think it's a great integration tool. I generally do believe in that. citizen automator model Work auto use, but I just think there's a point of scale
32:47and complexity where it makes sense, both from a financial perspective and also from a scalability and operational excellence perspective That you will want to convert it into software, but we would often start with, hey, we have this idea for something that we think might be valuable to reps. Let's build it as a prototype in workado and roll it out to a subset of our reps. And if it proves valuable, then we can. think about scaling it up and turning it into a fully fledged software solution.
33:18So it ends up being like a really valuable prototyping and fast paced prototyping tool for us Justin Norris: I had a fascinating conversation with a fellow named Niles Fote, from Trey. so we're Cotto's competitor, but, using Trey as part of their go to market. And he, uh, has developed a fairly sophisticated model where it kind of all their, lifecycle processing is happening in Trey. So they're doing like the event capture from the website. All of that sort of enrichment scoring MQL stuff. And just, I thought made a compelling case around why that was more scalable
33:52and, um, overcame some of the limitations that you get, like with, you know, in Marketo, you're capturing fields onto the person objects, so it can be really hard to manage race conditions because those events can get overwritten before other processing happens. There's like a limit, beyond which you can't go. What are your thoughts about that? Or maybe just tell us a bit more how you run that. call it operational backbone of your system. Are you doing that in HubSpot or have you written sort of your own software for that? How does it work? Rich Archbold: so we have a homegrown bespoke lead pipeline system, given
34:25the, given the scale that we're at. And so we have some custom software that allows us to capture All of those events and send them through a homegrown enrichment territory assignments, and, routing system. And so that really is all homegrown custom software for us. And we have various different levels of operational dashboards, and pay, like, a lot of attention to making sure that. A lead that enters the system actually gets routed to, a rep,
34:59and we pay a lot of attention to when it actually gets to the rep. Did the rep think it was a good lead or did they end up disqualifying it? Because some of the. Enrichment data didn't seem like it was right, or some of the segmentation data, or some of the territory data, or, that's some of the things that we thought made it actually a qualified lead actually weren't a qualified lead. And those are some of the metrics that I paid most attention to, to be honest with you. I looked to see, did the process from capturing the lead.
35:35To qualifying it and getting it in the hands of the rep. What did the rep think? Did the rep think it was good or bad? And why not? And like, are they policy issues or are they software issues? And for the ones that are software issues, do we make sure that we have the right teams thinking about, how to improve our enrichment systems or how to improve our segmentation systems or how to improve our assignment systems. And for ones that are policy issues, are we making sure that we're giving that feedback?
36:06And that the ops teams are getting all that feedback, do they feel like they have the right configuration tools out there? and do they feel they have the right agency to make changes to tweak things that they want, or are those features, We need to build, Justin Norris: that makes perfect sense and you've alluded to data a number of times that's also within your purview So maybe we'll turn to that This is a broad question, but how do you think about data strategy at the scale that you're at? Like, what are the sorts of first level concerns that you're dealing with
36:39when you think about HubSpot's data? Rich Archbold: I think it's the most important thing. and again, if you said to me, what are some of those hard fought lessons learned that you pick up as you go from working in infrastructure to working in, the go to market world. I think as a product person, as an engineer that builds a products for customers to buy, you generally take data quality for granted and you focus on building features. you have full control of all of the databases. You've got to pick the database technology.
37:09if it's my SQL, you can rely on nice, clean, simple, my SQL replication, or even if you need to stream it into some other sort of a search, elastic search database, there's like nice clean patterns for doing that, and you're all about building features. And then when you come into the go to market world, you realize it's the exact opposite You get to buy your features. HubSpot is going to give you features. Salesforce is going to give you features. Marketo is going to give you features. You need to build your data. you're actually, the more systems you have, you end up with a heterogeneous
37:41multi master database replication system with, no good consistency checking. And so you've multiple sources of truth, data replicating, left, right, and center. it's a native integration here, and then it's a workato integration over there, and they're all point in time integrations, and nobody's written a data checking thing. there's no, primary keys anywhere, there's duplicates all over the place, and you're going. Well, you have a new, B2B SaaS software rep going to sell you some
38:12new features to solve your problem, and so people will sell you features. You can buy your features. Features are not the problem. The problem is the data quality. and so that's where I spend, again, and. Inordinate amount of my time, thinking about, how do I make sure people understand the hierarchy of needs of data and that my company object is the single most important thing I need to have good governance and quality over followed closely by my contact object as my, as my 2nd, most important thing and then
38:43flowing down to deals or opportunities and then select contracts or. Subscriptions to invoices and understanding that hierarchy of of data and making sure that we have. Some sort of a governance process on each, like the, Schema management of each of those different core objects that we have some semblance of what does good mean for each of the attributes? Inside of these individual objects.
39:14So, like, how do we uniquely identify a company or an account? And do we have the right semantic and syntactic safeguards in place to make sure that that remains? A unique identifier. how do we think about company size? if we are enriching a company from 7 different sources, how do we think about survivorship rules around each of these things? How do we think about hiding the mess from reps? So that reps only see 1 value rather than the 7 values, but they have the right
39:45level of, lineage and discoverability. So. Even though we're only showing them one they know why it's only one. and so, yeah, I spend a massive amount of my time thinking about and evangelizing, operational data quality inside of our CRM and making sure that each of the different groups and teams are. Thinking about it, that we've got the right schema management strategy, that we've got the right data ingestion strategy, that we've got the right data, processing strategy, that we've got the right safeguards
40:16along the way, that all of this. It's so easy to get caught up in data quality for data quality's sake. And you're like, I do, data quality. No, you're here to help the company bring its go to market strategy to life, which means reps doing their jobs. That's what you're here to do. And one element of it is you need to keep the data, right. So that all of that can happen. And so you can get caught up in a world of like where. Data quality, I'm going to remove some duplicates. And you go, well, you could remove 75 percent of the duplicates, but it
40:48could have zero impact on a reps, work. And so you need to be talking to your reps and talking to your RevOps leaders and go, whereabouts are you feeling your pain points? And where do you think, Is your problem, generating net new business is your problem doing cross sell upsell? Is your problem doing forecast accuracy? is your problem, doing life cycle marketing because you don't have enough product usage data? where's your problems? I'm here to solve your problems. And odds are there's a CRM data quality challenge in there that
41:22we're going to need to figure out. But yeah, just making sure that. Whatever we do is in service of those, business problems and business opportunities rather than, polishing our data quality because you can spend forever. You can spend all of your time Justin Norris: And it's never Rich Archbold: CRM data quality and it's never done. and it's never right. but we're not here to generate good data quality we're here to, delight our customers and. generate leverage for the business and we focus on those two things
41:54first and data quality is a means to an end. Justin Norris: feel like we could do a whole show just on data because, yeah, there's so much in what you just described, both in terms of the current state of things and mess that we generally all inhabit. and accept as normal and the different things that you can do to try to mitigate that just for fun. I want to double click on maybe just one of them to see how you solve for it. Tactically you mentioned about like the data lineage and you might have, you know, seven different, data sources for a particular data point.
42:26you have some means of, Consolidating that, displaying only one to the rep, but in the background you can still track what you got and from where, like how do you solve that specific problem, to take an example. Rich Archbold: probably maybe the most applicable way was, how we were solving it at Intercom and how we were bringing in, ring lead, or I think like zoom info ops OS, what we were able to do was plug in multiple enrichment sources into ring lead and program some pretty complex survivorship rules for these things that then, ultimately updated this, single field in Salesforce.
43:00And we could say, like survivorship rules of where you have linked in data. Um, Always use linkedin data where you don't have linkedin data if you have, let's say zoom info data and some other data provider. And if it's in North America default to, zoom info, but if it is in Europe default to luscia, Data, let's say, like, I don't think there's any, 1 size fits all here. The different data providers will have different strengths in different
43:33regions and different segments. I think that the key thing is, you've got to keep talking to your reps and getting feedback from reps on what's actually working well for them and what's not working for them. I think that, the secret for me, or the key is that you're, rarely ever going to have somebody articulate to you. Hey, your priority zero problem is CRM data quality. You need to go and, I'm going to allow you down tools and not integrate any new software for me or not build any new workflows for me and just
44:06go spend 80 percent of your time cleaning up the data inside of our CRM. that's pretty much never going to happen. And. Okay. If you don't know any better, or if you haven't done up the battle scars, it's very easy to listen to your internal stakeholders. And they'll say, I want this tool. I want this workflow. I want you to launch this product. I want this new enablement source and you end up spending 120 percent of your time, adding new to the system and adding new complexity. Quickly, as you can, and that's the path to getting out of control.
44:40And I think that the real secret to us is having the conviction to know that, this is where you're going to go wrong. You're going to go wrong. If you're not constantly gardening and gatekeeping, uh, some of the data inside of the CRM that you aren't, um. Preaching this to your team that you aren't putting those guardrail metrics in place that allows you to get a sense of. this is kind of how good or bad it is that you're continuously talking to reps and
45:13that you're listening out for these type of keywords, that were, and you're able to tie it back to, I'm hearing data quality, but it's actually a prospecting thing. I can now go sell it to the business and go, Hey, here's how we can improve prospecting, or here's how we can move the numbers you care about. it's by investing In this thing, and I think as engineering leaders, it's up to us to champion that and to have that mindset and to not, unknowingly, walk ourselves into a corner of not necessarily technical debt, but Data debt, I think,
45:45again, if you're building products for external customers, you're all about the technical data and know that like software isn't being properly reused. And we need to take the time to refactor this software so that we're carrying less software. I think the same is true in the go to market world. It's in the data space. It's like, are we taking the time to refactor our, data, eliminate data that's now duplicate or unnecessary. And for me, difference between building product and building go to market is just a role reversal.
46:16One, you build products, the other is that you build data so, that you can buy your features. Justin Norris: And maybe the last question we have time for, but you've mentioned a few times about, this push and pull with your stakeholders, identifying what work you need to do. figuring out your priorities. I think this is perhaps the most challenging and an important job, you know, for any leader in the space. Cause if you make the wrong decisions, all the other work downstream, loses its impact. How do you think about that problem?
46:48how do you manage it? Rich Archbold: I think you're right. I think it is one of the hardest things. so at HubSpot, we have an excellent annual planning process that happens each year, which, that's where we kind of set out a lot of our, uh, Big rock goals, and we try and make sure we're generally hyper aligned with our business stakeholders on what are the biggest strategic needle moving projects that we need to do. And we make sure that we set aside capacity for those things.
47:18But we also know the business is, agile, the world changes around us, and we need to be nimble and responsive to those changes. the first port of call for being able to be nimble and agile to those changes is those, technical BSAs that we have embedded inside of rev ops, who we're continuously trying to build features and capabilities that. they can tweak and respond to. We also have work that enables us to, deliver some amount of agile automation, in that manner as well.
47:50And then we have our. Various different interlock meetings, our sales interlock, our support partner interlock where we have an ongoing intake process where, people could submit new requests. And we review those interlock meetings are happening on a bi weekly basis at the moment. I'd love to say they're frictionless, but they're not, sometimes we will discover dependencies that we didn't realize we had, or dependencies that were maybe bigger than we thought. Um, we need to adapt to them.
48:21it's, definitely not always perfect, but, some of the really interesting things are, What do you do when sales ops wants to do one thing, but your reps want to do something else or maybe the priorities aren't quite aligned where the reps will obviously be the ones to see the paper cuts and the friction with the existing processes. They all like. See and feel them every single day, every single deal they work, whereas the, sales ops or rev ops folks might be 1 step removed from that.
48:54And they're, more hyper focused on the bigger swing, strategy plays or supporting some of the new product launches or supporting maybe a bigger Segmentation change or something like that. And that's actually where it gets really interesting for me, where your first desire is, I want to make those reps happier. I want to fix those paper cuts for those reps, but maybe it's the same group of people that would be working on executing the larger strategic project, those are the really tough ones.
49:24the philosophy I preach is to try and set aside, no more than 70 percent of your capacity for your, Big rock strategic goals and leave 30 percent of your capacity, free to respond throughout the year and to. Address those paper cuts. And then my other tip, is, the more closely you can connect your engineering teams and your systems teams to the reps, the more rep shadowing you do, the more empathy you can build with those reps. The teams will just find a way to go, Oh yeah, we know these people.
49:57We're talking to them a lot. We know they're feeling this pain. We'll find something that we'll find a way to address some element of it. Just so it's not a sharp, but if you say we're primarily here to execute the strategy, but we measure the rep friction and we really care about rep friction as well. and you set aside some capacity for it and you get, The engineers and the PMs and the designers meeting regularly with the reps, everything just gets easier. That's the secret, the closer you get to reps, the more rep shadowing you can do without actually getting in the way, the more empathy you can develop,
50:28for them, the more you can help see around corners and fix things before they break or fix things before you're even asked to. Justin Norris: It's knowing your customer, you know, right? Rich Archbold: 100%, Justin Norris: closer you are to your customer, more empathy you have, the more you'll intuitively be drawn to fix those problems. Rich, this was so, so interesting. Uh, could have probably booked a three hour session. We'd still have more to cover, but really grateful for you joining us today and just being generous, sharing all your insights and what you folks are doing.
51:00I think it's fascinating. So thank you so much. Rich Archbold: Thank you for having me.