#11: How to turn cold outbound into an always-on GTM conversation, with Keith Jensen

(upbeat music) - Welcome back to Revenue Architects. I'm John Lloyd, and I'm joined today by Keith Jensen, president at Codra AI, a platform that acts as a strategic AI implementation partner to help businesses drive realistic impact. Keith, welcome to the show. - Appreciate you being here. It'd be fun. - Prior to your current role, you've held multiple positions in marketing leadership, whether that's as a CMO or a fractional CMO, you're breaking our long running streak of having rev-ops leaders and only rev-ops leaders on this podcast. So thank you for helping us mix it up a bit here. - From sales teams too. I mean, that's helpful, right? - That's right. It's all in the family. So in today's conversation, we're really gonna get into a lot about engagement trends, the state of cold, outbound, and the importance of the TAM underlying every marketing campaign. So to sort of set the stage here for everybody, reply rates to marketing and sales communications have been falling for years. That might not be a surprise, but call it context overload, call it short attention spans, call it whatever you want. Many teams assume that the answer to this problem is using AI to send more communications and faster. So instead of doing that, Keith, what you and your team have done is really rebuild what sits underneath the communication layer. So I guess my first question to you is, in a world of falling reply rates, and faster AI-driven communications. At a high level, what's your approach to cold outreach? - Well, I think the way that people have to reframe it is that while everyone, people say, oh, cold email's dead or cold outreach is dead, and the fact that reply rates have been falling and obviously it's harder to do. Like if you put links in an email, it's more likely to go to spam. If you have anything in there like a pixel that's gonna help you understand whether the person opened it or not as a signal, that's more likely to go to spam. So you have less ability to kind of get the signals up front, meaning like did they open it, did they reply, right? There's less of that happening. But if you think about it a little bit differently, more like you have the opportunity to have an always on conversation with your total addressable market. That's how you should think about it. And when you break it down, the messages that you send out, you can always get more creative with them. You can do things like, hey, we're doing a white paper in your space, would you like to contribute to it? You look like a great person for a thought leadership. You know, you can reach out, you can, hey, we're doing an event in your area or we're going to this event in your area, whether you're gonna be there or not, would you like to come for a coffee? You know, I've got a couple other people going on Tuesday at noon, you can meet some people in the industry. There's always ways to get a better response rate, even if it isn't direct to say like, can I book 15 minutes? So I think one part of this, like the framing like that. But the other part to think about is, how do you get those other signals that are further down the line that somebody might be interested? Just because they didn't reply or you don't know whether they open doesn't mean they didn't find some interest in it and go to your website. So there are tools now that can help you resolve who comes to your website without them filling out a form and you can match that back to your tam. So if you have somebody that you've been sending emails to and then they happen to hit the site, you got a pretty good idea that they came there because of that email, right? Now if you want to score them up from there and you want to add them into a seller, Cadence, you could do that or you maybe want to drop them into a different type of messaging. Cadence, you can do that then. So there's other ways to kind of gather those signals and figure out whether it's actually, you know, working is kind of a tricky word, but whether it's actually doing the job that it was intended to do, which is raise awareness for your brand to, you know, that company. - Yeah, and I think that resonates and I think signal recency is a huge pillar of that too, right? It's like your signals will expire. They will deprecate along with your other data. How do you approach, you know, keeping things fresh on the foundational layer? - So I think having that, like most, the most robust and up-to-date TAM possible is probably one of the most important things. People change jobs, things are, you know, things change, you get their, you know, you have a wrong information from one you had if you didn't have a phone number and all of a sudden now the phone number is available on one of these tools. And so I think keeping that TAM fresh is really important. We've used lots of different ways to do that from API waterfalls to, you know, AI to kind of help susstrew to find, to kind of match people up to see if there's still certain jobs. But ultimately when it comes down to it, the repetitiveness of you, like if you are sending the same message over and over, that is never going to be a huge value. You know, when you think about from the sales side and the rev-up side, all of those playbooks, which I'm sure you guys have built plenty of them where it's like add value, add value, add value, ask for something, right? You shouldn't, you know, abandon that and people think, well, it's cold email, it's a numbers game, we just got to get as many emails as possible in asking for that 15 minute meeting. And I really don't think that's the way you should be thinking about it. You know, just because you're, just because you are not getting a response on, let's say, a meta or a TikTok or something, you're not expecting everyone to give you a response. You're basing it off the end, like what's my either click or my conversion rate or my CAC? Well, why is email different? Why is email all about reply rate or something like that? You know, it is another channel that is actually arguably way cheaper to get in front of your TAM and your ICP, then it is to go pay through another platform, right? And you have a better idea, I think, of whether they're actually engaging with that based on, again, response rate or hitting the website or this or that. There's plenty of other ways to kind of measure that. So I think if it is just another channel that gets paired, you know, along with the paid side. - Okay, so your philosophy is really that a good, cold outbound starts with keeping the foundational TAM periodically up to date, constantly up to date. What were the big trends or, you know, pain points that you encountered that kind of shifted your view or have you always had that view? - You know, it's been a long road with me and cold outpound, you know, it's come a long way. AI has certainly helped you move faster. I remember, you know, a few businesses ago, we had massive teams that were doing this, right? And then when I was at some smaller companies, 'cause my background, as you mentioned earlier, as a CMO, I've been at mainly in the private equity space. So I've been CMO for two different private equity firms where I'd oversee every company that they owned and then, you know, I've deployed into them for sometimes six months, sometimes eight months, sometimes a year. And I've been obviously internal at other companies that largely private equity backed for a few years at a time, they've ranged in terms of the size, some are in healthcare, some are technology. And so how you're reaching out and how you're getting that TAM varies across all of the different, you know, if you're in healthcare, it is not so easy to get that TAM list. Because if you're marketing to other providers, that physician's assistant or front desk person is not, they may not be on LinkedIn, right? They may not even be in a database. So in those cases, you might have to go purchase the TAM from like a specialty database that they signed up for, right, to get your TAM list. And that's a little bit harder to update over time, you know, versus when you're in SaaS or you're in any kind of business where the TAM is pretty readily available through searching through firm graphic or type of graph or any kind of information that is available and, you know, zoom any of these other, you know, tools where you can just click, click, click, you know, they have this type of title, this seniority, this type of business, this revenue range. Boom, here's your TAM, that's a lot easier to keep up to date, obviously. - So when did it really become, you know, obvious to you that the problem was, you know, a lead list versus or contact list, company list, versus the messaging and the content itself? I don't think it's a versus, I think it's an end, right? You can ask that contact list out there if you're sending garbage messaging or your, your, you know, value props don't align with what the ICP is actually looking for or you just don't catch them at the right time. There's a million reasons why it won't, you know, work right away. And as we see kind of the trend of emails going up, responses will go up, right? It will happen, that is a numbers game, but if you vary that message and try to pair it as best as possible with the ICP, you're gonna be in a much better position. So I wouldn't say it's an, or it's more of an end. - No, that's a fair call out. When we're thinking about, you know, the mix of human roles and what AI agents are doing, at least, you know, from your perspective, I wanna paint a scenario for you. So let's say an SDR or a BDR working, you know, a top of funnel gets a bad record. They fix it in their head or manually in Salesforce or in their sales loft cadence, let's say, and they move on. What happens when an AI agent is doing the outreach instead? - Well, so that's actually an interesting point. A lot of people are scared when they have this master tamless. So they have 400,000 people, right? They want to reach out. and you're doing cold outbound to them. Most people don't want all of those contacts and leads called them stored in their CRM. Lots of reasons why I think there's arguments both ways. One way is if you didn't want to push them into your CRM and you want to tag them with something that allows you to do a filter at a later point to get them out, you can do that. There's been companies where we've loaded them all in. You're not necessarily paying for them because they're not in a marketing sequence. So you can have this master is massive list in there where you've got 400,000 people, but they're all tagged in a way where you know that they were brought in through an integration and the integration was from this tool and they're tagged as cold outbound. And then later every month, if you wanted to clean it out or something, you could say go through anybody that hasn't responded and this and that and boom, get them out of there, right? So you can, there are ways to kind of load them in, but a lot of people still, I understand why, I've been on both sides of this, so why you wouldn't want them in your master's CRM. We've actually built for some of our clients that we do this type of work for. We've built like a pre-CRM that we call it. So it's another tool that kind of sits before your CRM where all of your tam lives in there, someone comes to your website, we're able to use the pixels to resolve that person's identity against the database of a database where everybody has opted in. So maybe 60% of your website traffic is actually, you actually know who is on your site, right? Because obviously, when people come to your site, 99 point whatever percent are not going to fill out of four. So even though a hub spot or a sale or a part-odd or whatever it's called now, even though because they may understand who that person is because they know the IP address, they're not going to tell you who that is, right? But there are third party tools that will help resolve that identity. So if you can resolve that identity and push it back into the pre-CRM that has your whole tam, and now you know which person has visited your site, you can decide which ones you want to flow into your main CRM to allow them to now enter some kind of marketing sequence or be assigned to a seller or anything like that. So it's a way to kind of pre-score your tam, right? And that also allows you the opportunity to be updating it in one place because anybody that's in that tam list that is like, again, pre-CRM, anybody that's in there, you know, has not had enough of a score to be able to flow into your main CRM and get it to a point where you assign it to a seller. - So that's a good, I'm glad that you called out, you know, potentially not a flaw, but another layer to the CRM that needs to exist, at least in your world, because in past iterations of this podcast and past episodes, we've talked about, sometimes the CRM is not, you know, the center of everyone's world and sales and marketing anymore. Increasingly so, it's changing, it's a role is changing. Contact data accounts live in other apps sometimes. So like when you say pre-CRM, let's get a little bit in the weeds now, are we talking about like a snowflake database, like Databricks, what is pre-CRM for you? - I don't think you need to go that heavy. If you already have a data warehouse set up on a Databricks, snowflake or anything like that, you could absolutely add it as a table and they're pretty simple. You know, a lot of clients that we work with don't have that level of data architecture, there are tools that are way more AI native now, than a snowflake or a Databricks that we've been using for the data layer and the kind of data warehouse, where it is pretty incredible, where we can get it set up, you build data models, semantic vector exation on top of everything, normalization of the data, and we've been able to set up data warehouses for clients within two months, fully, like literally all of their SaaS platforms feeding into one tool that is, where they're able to talk to it, get the right answers, and build their own visualizations and dashboards without an analyst. So that, I would just say like as an aside, for anybody out there who is wondering about how to make their data AI ready, there are some incredible tools out there that are kind of beating some of the bigger companies in terms of speed and ability to get that data to be able to get married together and be able to talk to it and get the right visualizations. So if we have a tool like that in place for client, we would use that as the data warehouse for it. In other cases, like you could spin up a suitcase and have it, you know, the back end of that is Postgres and you throw all the contacts in there. We do some scoring on top, you know, using AI and have an automation that pushes it into your CRM when it's ready. Okay, so you have your pre CRM. It's obviously, you know, being updated dynamically. What's the trigger for updating the pre CRM before things are fed into CRM? I would love for you to take me through sort of like a step-by-step process. Like what's the system here? What's the trigger? What's the enrichment? Yeah. - So it varies based on your ICP. So all right, let's imagine you have a 100,000 people in your pre CRM, you've done nothing with them yet. Obviously, you're gonna need to make sure they're enriched, you grab all the information from whatever tool you're going out there, you get their phone number, their email, they're accepted, right? So now all that's in there and ideally, like I said, we already kind of narrowed down that 100,000 from the master universe to be able to say, these are the people with the right title, the right seniority, et cetera. So now they're in there. Now what you want to start doing is figuring out how you're going to do outreach to them. So one thing you can do is the cold email. We'll talk about that in a moment. But there's other ways you could think about this of like, you could take that 100,000 person list and you can load it in and use AI to actually do cookie IP and device matching. So if you wanted to say, all right, we're going to send cold email, but I also want everybody who's in this 100,000 person list to be seeing connected TV ads, right? To be seeing social ads, you can take that list and if you do the device matching, you're able to send connected TV ads to people directly and if they have the Hulu version that isn't the ad free, online media, any other kind of streaming service that as ads, you can pop your message into there and you look like a much bigger company because, oh, wow, they're doing video on TV ads. I'm watching it. Well, I'm watching this, my old madmen rerun and now there's an ad in there. Boom, now they're also getting an email and now they're seeing a social post. You'll look way bigger even though you're only marketing to that smaller segment. And so from a normal channel point of view, you could take that Tam and utilize it in lots of other places. Obviously, you could beat it to your sellers if they need a backup of leads. You just go into there, grab it, pull it out. Again, if you connect to TV, display ads, all could be done through cookie or device matching. So it matches up to your Apple TV and all those types of things. And then there's, you can get that list and load it in and match against the social platforms so that you can actually get hit the same people there. Okay, so let's say you got your Tam, you're hitting them on the other channels and now you're like, let's go do, let's hit them with the cold outbound as well through email. So you can take that list, you then go and do messaging. And what you want to do with messaging, what we like to do is when we build our master Tam, we broke it into verticals. And what we said and kind of grouped. So we looked at like, with cadre AI, we are a strategy and integration firm for AI. So imagine like companies, they're coming to us saying, we know we need to do AI. We don't really know what to do. We need help with the strategy of what we should be doing, the training on how to get our team up to speed. We need help getting our data AI ready so that the agents can actually read it properly and aren't gonna be guessing every time because it's a probabilistic model. And you don't want it to be guessing when it comes to data. They say they want help building the agents and optimizing workflows and then helping with the change management. So if we have to do all of that, you think about our Tam is massive. It's pretty much any business out there who believes that they should be implementing AI, which right now with the hype is pretty much every one of them. And then you think about like who's our ICP? Well, right now it's mainly founders, owners, presidents. Some instances you have like a chief AI officer, some instances you have the VP or COO who are sometimes is the GTM leader who's saying, "I don't want to do this," but usually at this point, what we're seeing is that this is a company-wide initiative that starts at the very top. So if we were to say, okay, we're looking for founders, CEOs, presidents of companies in this range, blah, blah, blah, we get this master list. Now we need to figure out like what's the messaging? So when we went through that Tam, we said, what are all the industries? There are 515 different industries. Like, well, we're not gonna tailor our messaging based on industry then, okay? So then we said, how can we AI help us? But we rolled up those 515 industries into like about 55 industry groups. And we said, okay, that's better. That's more palatable to do that. And then we said, let's roll that up into sectors. We both broke that down into 10 sectors. So it might be like real estate and construction. You break that down and now under there, there's like all different sectors under there. And are all different industry groups under there. And under each industry group, there's a whole bunch of industries. And then we created playbooks at the industry group level. So at the 55 of those, we created 55 playbooks that said, what's the messaging to this audience? What are the top 10 AI use cases that this audience would care about? So it's a construction audience. It's like, well, you use ProCore and you use NetSuite. And here's a way to blah, blah, blah, blah. Like, here's a way that AI can help you with this or that. And so we found all these use cases based on our experience in those industries. And we said, here's the things that we know they're better will likely resonate with them because it's things that other clients have paid us to create for them. And so we said, here are the use cases that would be interesting. And then we built in those playbooks, by the way, we have like talking points for sellers, we have, it creates blogs, it creates a ton of things. So we kind of get the messaging, the use cases, and then we put together demos. We make sure we have demos of each of those things that we're putting in the emails. In case they reach out, we want to be able to show them, we have our talking points, we have our rejection handling. So now we're ready to go. And we start saying, all right, let's take that time. Again, we said, cold app on email, it goes out through there. Let's go to create some video content that's going to go to Connected TV, let's create newsletter content that's going to go out through personal newsletters on LinkedIn, we create hyper-targeted paid ads through the IP and device matching, right? And then you want to obviously have your website lead gen magnets, your landing pages, your white papers, your webinars, your blog articles, events, we do something that's really cool. It's called the AI maturity index. So if you go to cadre.ai, there's an AI maturity index tool on there where you can go in, and it feeds through, and based on 2,000 responses of different companies, we have questions that it'll take about 10 minutes, and you and your whole organization kind of run through it, and it tells you where your company sits on a scale of 0 to 100 in terms of AI maturity, how each team is using AI, what tools they're using, how they learn better, so we can help with training. So we have lead gen tools like that that will provide a lot of value to businesses, but obviously we'll get hooked in. And so now, OK, we start sending the emails, we're doing connected TV, we're doing these things. Some of these, you can get click data, but other things you can't connect to TV, you have no idea if they're clicking or not. Outbound, cold outbound, you don't want to put a pixel in there that's going to see if they opened it, and you don't want to put a link in there. So what do you really do? You're kind of like asking them for a response. But there's a good chance they get a cold email, and even if they're interested, they're just going to go to your website. So now again, comes the, they hit the website, you do the resolve, their identity. And if there were one of the 60% of people that hit your website that have signed up for a newsletter on some other site somewhere and opted in for their data to be sold, now you're able to know who that person is. You match that back up to your pre-CRM list. And then if they score up enough, whether that is like they've clicked on, as they filled out forms, they've hit the website a certain amount of times, maybe they've hit certain pages on the website that are higher in value or intent than others. They score up to a certain level, and then AI pushes them into your CRM and assigns them to the right seller. - Something that, thank you for walking me through all that, something you hit on earlier was CTA, the call to action, you know, avoiding just the generic, like, hey, let's meet for, you know, 15 minutes. You got a quick sec, you know, this week, which is basically just throwing something into the wind, hoping they're going to respond back based on nothing relevant. So like, how do you, how do you tailor, you know, again, like talking to the audience here, like for a marketer who's looking to really create a better hook at the end of an email or a call or what have you, how do you tailor your call to action there? - With AI, it comes really down to use cases. So many people are hearing about AI these days, and they're like, yeah, I hear about agents, I hear that there's like agents that are just people are managing them, and then they're not the higher another person for it, but like they don't, it's hard to wrap your head around it, right? And when you jump in to a clot or a chat GBT or anything like that, it's kind of like, it's kind of like the early days of the internet, right? The early days of the internet, you know, 46 now, like when I was, I remember the Beboop dial up and all this stuff, right? And like, but the early days, like, it wasn't transformational in businesses, it was a personal tool, right? It was, it was, you look up lyrics to a song, you'd be chatting on instant messenger, you'd be using it for those things, and over time you're like, wow, this thing is really powerful, and now businesses are incorporating it, and it then becomes a core component of businesses. I believe AI has the same trajectory, it's just a lot faster, right? 'Cause what we're seeing with the adoption of AI, it's not the traditional bell curve, where there's a lot of people kind of like, from low adoption to high adoption, it's not like that, it's like a barbell, where there's a lot of people that aren't using it at all, and a lot of people that are using it like crazy, and not a lot of people in the middle, and the reason is because it's like, it's got insane product-led growth to it. Once you, if you start using it, you're either gonna decide this ain't for me, I'm not interested in this, I don't, I'm not interested in the hype, maybe there's, you know, environmental issue, whatever you're latching on to, if you don't wanna use it, you stay in that area, but the people that use it and go, whoa, boom, they go right through that middle section and go right up to the power users of it. They start building their own skills, and now they're just like, everything they do, they're trying to automate. And so, to us, when we think about the messaging and what's resonating for us, it's all about, how do we help a client understand how AI can help them specifically? Because if they just jump into a clutter or chat GPT, and they're like, how can you help me with my construction business, or my HVAC business, or my SaaS business, it's gonna give them this long list of things and they're gonna be like, I don't wanna read all of this, and this isn't that helpful. And, you know, the whole like a 60% rule of like, it is so easy with AI to get to 60% done. And it looks great, but to actually make it functional and work, especially not just for you, but across a business, that kind of remaining 40% is where all the effort and work comes in, and it takes way, way longer. So for us personally, it was all about figuring out how do we connect the things, the pain points that we're very aware of at each of these ICPs with the actual use cases of how AI can help. Now for other businesses, I would say there are some generic ones that could be used, like again, the ones I mentioned earlier, such as, hey, even if you're not going to an event, if there's an event in Nashville, send a blitz to everyone in Nashville. Hey, think we're gonna be at this event, we're thinking about doing a coffee chat, a Tuesday at this. There's doesn't seem like there's a lot of program around that time, and grab another couple of other people. Let me know if you wanna join, meet some other people in the industry. If you get enough bites, fly out there. - Yeah. - You know, then do it, or reach back out and say, oh, so sorry, we didn't end up going, would still love to connect. How about 15 minutes? - Yeah. - Now you've already had a conversation with them. So whether it's that, or whether it's, do they wanna contribute to a blog or a white paper to help be thought leaders? We have a podcast as well, called AI 2030. We have a lot of guests, you know, people that are influential in the space that might even be ICPs of ours on the podcast, because if they're really like, you know, true thought leaders in the space. So there's just so many ways you could think about as, you know, in's with your ICP, and I don't think it's always just about the asking for the 15 minutes right up front. - I like it, it's situational, and it allows you to say, you know, hey, are you gonna be here on Saturday? Let's grab a cup of coffee, or like, hey, I saw your AI maturity score was blank. Like, what do you think about this? Let's work with this. So, yeah, I definitely resonates with what I think is happening, and a lot of companies just, a lot of my clients, you know, struggle with just including, you know, a pretty broad call to action without incorporating some of that signal data and geodata. So let's bring things a little bit higher level now. What, where does the ownership lie? You're talking about, you know, keeping the TAM dynamically updated. The TAM is feeding the ICP. The ICP is feeding, you know, your ads, your channel mix, your emails, your cold calls, your events. Is this something that marketing owns across the organization? Is this something that shared ownership between marketing and sales? Clearly, there is a handoff there at some point, or maybe just, you know, a left brain, right brain, they're just connected all the time. - Larger companies are gonna have a RevOps team. RevOps team is usually the connective tissue, working with the automations across the board. For smaller businesses, you know, they likely, maybe they have one RevOps person, if that. You know, so I would say largely this lives with marketing. The marketing has to be, in my opinion, has to be tuned in with sales, has to be really, really close. At some of the organizations I was at, I was like, sales and marketing leader. In other cases, I was just a marketing leader. And one of us is I was just a sales leader. And it's different in the sense of the size of the business and the amount of resources you have, that's really gonna dictate the needs. You know, at cadre, we're, we're, you know, three-exing growth year over year, but we're putting all of our team members into the people that service our clients. So we have a very, very lean marketing team, right? And we have one RevOps person, and then we have our sellers. And so we've found a way to construct this, you know, using AI and allowing us to have agents that do a component of it, and then humans are doing another part of it. And we're still obviously evolving, but it is, you know, everything I described is, is where is like a North Star of where every business, if they start working towards, they will never get perfectly there. There's always gonna be new tools that come out, new ways to do things better. So it's always gonna be evolving, but I would say marketing should really own a large percentage of this, if there is not a truly well-built out RevOps team. got it. And are you aligned with having, you know, market and sales sort of, you know, connected as one GTM motion or marketing reporting into sales, you know, what are your thoughts on reporting structure organization? You know, I've typically seen it be a little bit difficult for sales leaders to oversee marketing and marketing leaders to oversee sales unless they've truly done books. Got it. And so I think it is very hard to make that call without knowing the capabilities of the people in the roles. If you have a somebody who's done pure marketing and another person who's done pure sales, I would not have them report, not have either one report to each other. They should probably roll up and at that point either roll up to like a COO directly to the CEO or if you're a large enough organization, you would have like a head of go to market or a head of growth who both would report into. Which I've seen a lot more often these days. A head of go to market, a team of go to market engineers, you know, you see a lot of FDEs reporting under them. So I think that that rings true in my experience as well. I'm thinking about, you know, something that is a large part of what I do and the clients I work for is, is, you know, finding the right KPIs to benchmark performance on what does success look like in marketing. So when you're, you know, thinking about, you know, what you used to, I guess, as a barometer of your success, like where, what moved the needle and which needles moved after you, you know, started using your pre CRM and dynamically updating your TAM as, you know, as the foundation. Like how, how did you see performance shift? This again, you know, it's so different for us versus the, the clients that we've done some of this work for. Yeah. Because everyone's going to have a different gauge of what success looks like. Sure. In some cases, there is just a math equation of we need, if we're going to spend this amount on the tokens, the humans that are reviewing the content, you know, et cetera, if we're going to spend it and the time put into it, we're going to spend this much on it. We need to get this much in sales out of it. And so there's a simple equation where you work back and you say, okay, in order to get that many sales, how many, you know, discos, do you need or how many demos have to happen? What's the, what's the conversion rate between, you know, disco to demo to proposal? Whatever the flow is for that organization, you kind of work back the math to say, okay, well, we need this amount of people feeding in from the CRM that are scored up to a certain percentage that are going to, or feeding in from the pre-CRM into the CRM that are going to be assigned to the sellers. So in some cases, it's really a, how many, you know, some people call it an MQL, a marketing qualified lead versus a sales qualified lead SQL. So you've kind of like it varies per business. And some people will, some companies will have a dedicated, you know, structure in place where we know that MQL is if they have filled out a form, you know, but what you need to think about is when you start moving to things that are less, you know, self-based input, like a form and you're looking for more signal-based things, then an MQL can't just be based on a form, it has to be based on a score. So what score is going to get them to a point where you say they are qualified enough that marketing believes that sales should reach out to them. And then the sellers are going to take those leads and do something with them. And they're going to try to get on a call. They're going to try to get a disco set up. They're going to try to move them to demo. They're going to try to move them to propose them. They're going to try to sell them. And so what you need to be measuring is are the leads that are coming in through the scored side in the pre CRM? Are they running through the same conversion ratios that you're seeing for people that maybe have a little bit more intent because they filled out a form? If the conversion rates hold true, then like you have an exact number you need to go after. If the conversion rates are lower, meaning somebody who just, you know, they saw a cold email, they went into your website, they got scored, maybe you said like that's enough. You score them up, you send them in. And all of a sudden you see the conversion rates are lower. They're not getting on the phone as much with a seller. They're not making it to demo as much from the disco call. They didn't qualify for the banter, the medic, or anything that's up front in the beginning. So now you have to decide, well, my scoring system, is that need work? Maybe just coming to our site and matching up the tab isn't enough, right? If you have a tab that is, let's say, more enterprise, there are signals that you can use AI to scrape online to say, are they posting press releases about certain things? Are they, if they're in the healthcare series, are they doing clinical trials about certain things, right? Yeah, lab scraping. Yeah. Yeah. With FireCraw and others, you can scrape the web for those types of signals. Do they have multiple locations on their website, like whatever the cues are that would allow you to say these are either a higher value prospect or have more intent, those can all be used to do the scoring criteria. And the better you get the scoring criteria, you want to get those as the best you can so that it at minimum matches the conversion rates of the form fills. If you can get to that point, then you're good. And if you can figure out how to get that scoring criteria even better to where you're getting a higher conversion rate of people, then you're at a point where you might go, you know what, I love the idea that I'm I'm feeding the best quality leads through to my sellers and their conversion converting even better than someone who just raises their hand through a form fill. At that point, you might go, I wonder what would happen if I lower the threshold a little bit, right? Because you don't want to miss out on people that that might be right at that form fill level. So kind of measuring that along the way and again, connecting all your data sources to a single data warehouse where you can have AI asking questions and doing that analysis for you is one of the biggest unlocks that we found. Yeah, that's huge. The interconnectivity making sure that the AI tools are actually connected to the fresh data is the entire thing. It's garbage and garbage out, right? Speaking of garbage, I want to hit on AI Slop and why it seems to be so prevalent in today's like I pick out five to 10 emails a week in my inbox that I just call out as Slop because there are just some things, whether it's too many M dashes or the lack of a relevant hook, where does AI genuinely help marketing content get built and how do you how do you avoid crossing the line into Slop? Okay, first off, the M dashes you're right. But also how annoying is it? Those headlines that are like five answers, nothing resolved. Like, like, there's short little quippy things where there's always a number in there, like five things, two of this, like, oh, it drives me nuts. Everyone here has reasons you should open this email for me today. Yeah, yeah, don't I tell everyone here at cadre, do not send me anything. If I think we're an AI business and if I think it's AI, take it back. Like, I'm going to be like, as soon as I start reading this and if I believe AI wrote that, I'm going to send it back. So you either need to get AI to the point that it is so good that I can't tell or you need to just put some, some, you know, elbow grease into some of these things and take the AI content and synthesize it. But to your question of how do you avoid the AI Slop? I have a pretty clear point of view on, on how I think work changes and I think it, it, it, the, I think people not accepting that or not realizing it is what is generating the Slop. So if you think about a task that took you 70 minutes before to do, right, 70 and it took you 60 minutes to do the thing and then it took you maybe five minutes to go, well, let me review it, right? And then another five minutes to be like, oh, let me send it. Okay, 70 minutes. Now if you use AI to do the work, the doing the work part take zero minutes. And what most people think is that what they can do now is say, wow, that took me zero minutes, I'm just going to take the five, the same five minutes I had before to review the work. And then I'm going to have AI go send it and it's going to take also zero minutes. But the problem is is that what you need to do is spend 25 minutes reviewing that thing, maybe 30. If AI made it for you and you saved your 60 minutes of doing it and now AI did it, spend more time on reviewing it, give AI feedback, read every line. Even if the first time you, that AI does it for you, it takes you 200 minutes to get it right. Then next time it's going to be better. And now you're reviewing it. Again, it does the work and you're reviewing it for 30 minutes, giving it feedback, getting it better. Now the next time it's only 20 minutes. And before you know it, yeah, that process is automated where it only takes you five minutes to review because you're so comfortable. You have, you have systematically over time gotten the slop out and trained it to be better. And whether you trained it using, you know, having it right markdown files or whatever the case of, you know, whether you're using a project and trying to save it in the memory, but like, you have to get to a point where that skill or whatever you're, you're using a build to how AI do the tool that agent has gotten better. It is your job to give the AI context and feedback. And if you don't do that, you are just passing along AI slot for someone else to review. And there's nothing more annoying as a person in organization to think that someone else in your team or someone even a vendor, God forbid, you know, just created something. Just took your fireflies call through and into AI and was like, Hey, here's what we should do next. Like if you, if that, that is so frustrating because you tip AI is verifying. For both, it writes long, and so you have to read all of that. And when you feel you're reading something that the person who sent you didn't even take the time to review, it is really frustrating. And so I think like there's a million ways to your question of like how can AI help marketers without Slop? Like there's a million ways AI can help marketers can help everybody in the organization. The core thing about the Slop comes down to you realizing that your job is different now. Your job is still to make, let's say your job is to make a hole in the ground. Just to do it with your hands and it took you an hour, now you have a shovel. But guess what? You got to assemble the shovel, right? It's good. The first couple times is going to take you a little bit longer to dig that hole, or maybe the same amount of time as it did before, but it gets easier, right? As long as you're willing to give AI the context and feedback, and that means spending the time on its results to actually explain to it why it's not right. We would never expect an intern to get our business to get something right the first time. And give them coaching, you explain to them the context. That's how you need to think of AI, and they will grow into an incredible employee, but they're not going to start like that. It's all in the process. It's all about setting rules, giving you the right data to work with. So Keith, thank you so much for your time today. This has been really eye-opening. If I was to ask you to kind of sum up what you've done, let's say I'm a client, and we want to come out of this with a few key takeaways to help marketers not only improve their reply rates, but just really create better engagement by keeping their tam fresh. Sum up what you do for folks at home, and whether it's one or two things, best practices, to help them out on a day-to-day basis. Number one, make sure you think about cold outbound email, not as a true pure lead generation source, but as another channel that gives you the ability to have an always-on conversation with your tam. So you should have your updated, most complete up-to-date tam list at all times. Whether you're updating it monthly, quarterly, it's up to you as to how large it is and how often you think people are moving around and people adding jobs, et cetera. But keep that tam up-to-date. Think about cold outbound as a way to have an always-on conversation, and get creative with it. Think about ways to add value where you're not asking for something, or ask for something that is not a meeting, or share something that is incredibly relevant to them. All the time spent up there, figuring out how to be as relevant as possible, will pay off at the end. Now it may not be in a reply rate, and it probably won't be in an open rate because you're not going to put that pixel on to track it, because you don't want to go to spam. But you could definitely figure it out based on whether those people are visiting your site, whether they're engaging with ads and other platforms. Use that tam to do cookie and IP and device matching to get your ads in front of just those people on other platforms. Because if you think about your tam, whether it's 5,000 people or 500,000 people, you can take a budget that you would normally spend on retargeting visitors to your website and just going through social and being like, oh, this criteria, that criteria, rather than doing it, that's start with the tam. And then say, these are the people. And even if you have to whittle that down and say, here are the high value tier ones, then fine, spend the majority of your money there, hit them on cold, hit them on paid, hit them on connected TV, just have your ad dollars working just for those people. And then do your best to resolve those identities and catch those signals so that you can feed it back and determine how to score them to feed into your sellers. Well, you heard here first, folks, add to your to-do list, have an always on conversation. I think that's the theme of the week. Big thanks to Keith Jensen for joining us, Kadra AI. I'm John Lloyd. This was Revenue Architects. We'll see you in two weeks for the next episode. See you, everybody. Thanks.