SDRs on Autopilot: How to run programmatic outbound campaigns (for next to nothing) - Harrison Rose

Those of you that I've not met, I'm Harrison. I previously co founded a company called Paddle and I'm now working on a new venture called Goodfit. My first company, Paddle, I started 1112 years ago. Hundreds of staff, 300 million in VC, raised a unicorn. Our most recent fundraise. I ran all of Go to Market revenue and strategy there for a decade or so. I've been working closely with notion since 2017 when they led paddle series B fundraise and I'm now working as a Notion expert, sharing learnings with folks just like you guys. And then more recently I co founded a new business called Good Fit. Born out of some of the unique ways we were going to market at Padawan and using data. But you'll hear more on Good Fit in just a moment. In terms of agenda for today, there's a huge amount of talk and hype around the future of Outbound, the impact of AI on sales and our outreach, and it feels like there's some fundamental shifts in the way that we're going to market and the different roles humans versus AI are going to take moving forward. I really got the chance to experience this firsthand at Good Fit in the last kind of month or two where we began Outbound campaigns for the first time, and I really got the chance to compare the experience, the setup, the infrastructure as well as results to when I ran and scaled this at PADDLE a decade or so ago. I guess. Spoiler alert for right to the end of the webinar. With no humans, no human touch points, and effectively zero cost, we're driving some really strong results on the Outbound side in an incredibly scalable way. And the disclaimer is it doesn't feel like we're doing anything particularly fancy or groundbreaking. There is no generative AI, if that's what you're looking for. Maybe I'm not the cutting edge webinar you should be attending. There's some GPTs and LLM in there, but just to make some tasks a little bit easier. But I think that the good news about this being simple is that I do think it should be something you're able to replicate yourselves in your own businesses. It's not rocket science. There'll be a clear set of steps that you guys can kind of follow and to get very specific on the agenda. I'll set the scene for kind of the state of outbound today versus 2017 when I was running this at paddle. I'll share then some context on Good Fit and the constraints we were working within when trying to set this up over the past couple of months. I want to then walk you through step by step how we and you guys can set up this kind of zero human touch outreach yourselves and then I'll conclude with the results that we saw with once we've talked you through the setup and then hopefully at the end we should have some time for some questions and I'll give you an idea as how I'm expecting to iterate and move forward with this over time. I think it's going to be tight to get this done in an hour. I talk unhelpfully fast which will help us get it done in an hour. Maybe not for you guys following. I'd suggest and encourage you guys to put some stuff in the chat if you have any questions and I'll try and feed them into the content or I'll have a look through them at the end should we have some time and we'll go from there. So I will get started if that works for folks. So let's take a step back in time to begin with, right? The year is 2017. Everything is good. Harry Styles released his first solo album. I didn't know what a Black Swan event was. I had a lot less gray hair and I'm running padawan and things are going pretty well. Right. But nobody knew who Paddle were. Nobody knew what we did or how it was different to competition. That was kind of my world. We made the brave decision to go out and find some customers ourselves because they certainly weren't looking for us and we were looking for those we felt would be a match for our product offering. We then went and contacted them and sold to them via cold outbound email. And this is in 2017. We did it all manually. I was running this out of Close IO, our CRM at the time and this was before kind of the rise of salesloft and outreaching and owning the category. And honestly, like things are going pretty well. Paddle was named the fastest growing software company in the UK. We were named the sixth fastest growing company overall in the Deloitte Fast 50. Something that happened twice in a row was quite painful because you get a bunch of perks if you get in the top five. That kind of hurt. Still not over it. Behind this kind of 3,000% growth that Deloitte calculated for us, we were running outbound email and I can't emphasize enough how this was our exclusive channel for years and we got really, really good at this stuff. Like we would indoctrinate team sequences, how to handle responses, how to get folks on the phone. And there are a bunch of like key learnings and principles we kind of built out when doing this. You know, a while back now they were things like the first goal of Outbound is to get a reply and only then book a meeting. Use closed CTAs in your outreach because they're really easy to reply to, as easy to reply to as yes or no, because that maximizes your likelihood of response. We develop these things called like show CTAs, which really accelerate your route to a call. Now a lot of those learnings and all those principles I think are still applicable today. And we've covered with Notion previously in a webinar called like Hyper Relevant Email. I think that content is still super, super applicable. And if you're interested in this topic, I'd encourage you to go and have a look at that. Maybe Claire can dig it out in the Notion kind of library and archives and stick a link in the chat. But that's kind of what we learned doing that a little while back. If we Fast forward to 2024 and the state of Outbound. I wanted to layer over some like hardcore screamo rock music on this slide to really amp up the terror, but thought that might not be my most professional move. But in 2024 everything is bad. Covid has happened. The good times have ground to a halt real fast. Budgets have been cut, Growth at all costs is over, valuations are worsening. In short, people are really struggling at the moment and most people are certainly struggling with demand. I'm sure this is a scenario that a bunch of folks are familiar with. And during this time, two perspectives I think have come to the fore. Perspective one is Outbound is dead. Like it's kind of hilarious how many LinkedIn articles we've all read telling you outbound is dead in the past year or so. Like my TLDR viewpoint on that is Outbound is hard, really hard. You can no longer make up for being bad at it by just sending a lot of volume. There are more competitors and bad people sending a lot of email than ever before. But if you do it well, Outbound is very much alive. The other perspective that's kind of comes in the fore is AI will change Outbound forever. And there's been some fear mongering around this as well. I think now on this, I think AI is inevitably going to change the way in which we go to market. It is interesting what different folks classify as AI versus not. I think a more accurate statement though is how we go to market is evolving. It has always evolved. And if you don't evolve yourself, you as an operator or your business, you're going to get left behind. Right? And that will always be true. Try not to get hooked in by like the pithy statements. Try and be aware of what's happening and take advantage of of the new tech and the opportunities that arise. But such was the extremity of these perspectives. Like I think I got stunned to a bit of stasis for a while, but I'm hoping today's content will give you some practical steps for you to be able to roll out some some new stuff that that's working really well. Just before we get there, why isn't outbound working for folks? Is it that instead or AI has ruined it for everyone? Again, I think the reasons I actually covered in a fair amount of detail in the previous webinar we ran Steven's helpfully helpfully linked I think the TLDR is like there's a ton of tech businesses out there today and the barrier to creating them is lower than ever. A whole category of tools has been created to make sending things like email and a lot of it really, really easy. And as such we're receiving a lot of email and that extends to other channels and touch points too. There's some stats for this on screen. Like this was a year and a half, maybe two years ago, we were receiving 121 per day. I think that's only increasing and I think AI is only making this worse. Right? One of the things that people have tried to do is to stand out is to deliver really highly custom, highly personalized email. We talked about hyper relevant advice. This is hyper personal in the previous webinar. But the introduction of AI has only increased our capacity to create personalized email and amp up the number of emails being sent and received. And again, that's true for other channels too. I think there's tools out there for AI driven ads, automated LinkedIn touchpoints, some of which we'll take advantage of in a second. But I'm not sure I'm at the perspective where AI has us all doomed. And again, having built zero human successful programmatic outreach over the last couple of months, I'm actually just going to give you a quick summary on what I think it is and isn't going to do and then we'll get into how you set this up yourselves for your businesses. Ellen's asked around outbound email versus LinkedIn messages and CTAs. I'm going to tell you exactly the different channels and touch points that I use in just a second, how we set them up and the results that we saw. So just bear with me and we'll get into that level of content in just a second. And thanks for the question. Keep coming. I'll either answer them as we go or get through a bunch at the end. And in terms of what I think AI is and isn't going to do, I can't tell the future. Right? I don't know what's going to be automated versus not in the future. But having tried and succeeded quite frankly, to set up this kind of zero human touch outbound in the last month, I'll give you my perspective on what I think you're going to be able to use AI for today, but you probably can't yet. But this will continue to evolve certainly, right? The thing that I'd certainly clarify, AI isn't going to work out a repeatable, predictable, successful way for you to run outbound. One of the questions that was floated during an SLT meeting I had at Good Fit was we wanted to start experimenting with outbound. It was something we'd not done before, but we didn't have any SDRs or BDRs. So someone smart in the team was like, oh, why don't we just go hire an agency or an ADR tool to go and aidr as they're called, Siri tool to go and test Outbound for us and see whether it works. That isn't going to work for anyone, I don't think. Like AI isn't going to work out your product market fit for you. You need to understand the audience of customers that you're serving who have a common pain. You need to have built a solution for them. And you're going to need to work out how to communicate the benefits of your solutions to those people. Now, in answering those questions, AI is going to help, right? It will help you draw insight from customer conversations, give structure to unstructured data. But you still need to go and have those conversations. AI is going to help you along the way. At this point I'm still not sure it's better than a founder's intuition, Having spoken to 25 companies or whatever, what it is going to do though it is going to help scale something that's already working. The ability to automate well defined tasks, decisions and activities that humans are able to do is certainly possible. And it's that fact which allows us to run much more of our top of funnel activities at higher volumes with lower costs and Less human involvement than ever before. So let me explain to you actually what we've done over the past couple of months so you can replicate it yourselves. So I'm going to be using Good Fit as a case study here. This is my latest business for some context. We provide data to sales and marketing teams. We're in a highly competitive space. There's new entrants emerging all the time, hundreds per month. There's some big incumbents like ZoomInfo. And honestly, we're in an industry where people have been burned like hell before buying data, tend to be happy with the data that they've purchased. And everybody likes a challenge, I guess. But despite those challenges, we're growing really fast. We've acquired some great logos. We've not spent any money on sales and marketing aside from attending some events. And the sales team is literally me and my VP of revenue. That was super small in terms of headcount. Historically, growth has been driven by inbound. It was referrals of existing customers, VP sales that have bought us moving from company to company, as well as VC intros maybe. Thanks notion and this got us to like well past a million dollars of pounds. So in error. And then I've been ready in the past few months to put the burners on and accelerate things. We had a couple of false starts, mainly due to capacity, but had some real breakthroughs on the outbound stuff recently. Again, to be really honest with you, I was quite scared to get started again and I tried to unpick why. I think some of it was ptsd. Like I personally sent thousands and thousands of emails manually when I was working this out for Paddle and getting back into that that mode was probably quite difficult. And there was also some imposter syndrome like I advise people on go to market, including outbound specifically almost every day. And you kind of need to prove that you still got it in this latest world that we're living in and working in, I guess so reputation was on the line a little bit. But we did what all good commercial people do, showed up and led into it and we got there in the end. Before I talk about what we did, I want to talk to you about our limitations. The limitations that I had doing this this time around with Good Fit were very different to 2017 at paddle. But I think those limitations really accelerated the learnings. Generally speaking. Like, I think having less resources as an operator makes you a better one. Right. If you're presented with paths A, B and C and you only have resources for one of those decisions, you can't hedge your bets between A and B and hope that you made the right decision. You need to pick one of them and. Or you need to work out smarter ways to do things all the time because you just don't have money or time. Right. I think having less resources makes us better. Our limitations in the case of wanting to run outbound though were we had zero people in sales. Like no SDRs, no BDRs, no A's, me and my VP of Revenue. And we had zero budget. Right. Literally no money spent and pretty extreme constraints. But helpful because I think they're similar ones to what a bunch of folks are facing right now. Smaller teams despite that, higher targets and less budget than ever before. Right. This is. This was a challenge I was up for though. So what did we do and what was the biggest anti climax ever? We did everything that also made sense eight to ten years ago. Right. Like the core concepts for Outbound remain the same. You need to map your market of companies and contacts you want to sell to. Within that market. You need to build an audience of companies and contacts you want to reach. You then need to configure your sequences, campaigns and touch points to the question that's in the chat right now. Then we put the tooling to work. This was a change. Rather than RSDRs, we put the tooling to work to get folks enrolled and get these sequences in front of folks and we iterated and scaled from there. Now it's like everything has changed the world that we're in, but absolutely nothing has changed over the last eight to ten years. Honestly, I think the things that would have made you successful at this ten years ago applied well, are probably the same today. But let me get into the weeds of each one of these five steps so you know what we did and you can replicate it yourselves. Starting with market mapping. So the first step in this entire process is mapping the market. Excuse me, of companies that you want to sell to. And I talk about mapping the market of serviceable companies you want to sell to. Again, we've done a webinar previously on how to map the market and why you should focus on your serviceable kind of addressable market first and foremost in the context of outbound. One of your constraints to outbound and revenue overall is the size of your market. It's why I encourage people to get a handle on how many accounts are actually out there. They can service all the time and as early as possible tool. Once you've done that, you can do things like apply some funnel assumptions to the size of your market to start getting some expected outcomes or even begin building the bones of your initial outbound experiments. Right. Like if I know there's 5,000 companies in my market, I think I can get a response rate of 5%. It's like 250 responses. If 50% of those responses become meetings, it's 125 meetings. But I also know I can't contact all 5,000 in one quarter. You can see how this back and napkin maths starts helping you just set some really early expectations and and the bare bones of your initial planning. But more importantly, you actually need to map the market to build the list of accounts you're going to sell to in the first place. Right. Having a list of highly qualified accounts is critical because your outreach, no matter how good your messaging is, is only as good as the people you actually send it to. I think this is something that people forget. You want that list as qualified as possible equally for the purposes of zero touch outbound. And this is a change. You need to map the market of companies you want to sell to build that list of companies you want to run your outbound at. But you also need to map the contacts in advance and en masse that you want to reach within those businesses. And you can't have a rep browsing sales nav and working out who within an organization they want to sell to and manually pulling their phone number or email address because we don't want any humans involved. Right. I've got more on how to do that in just a second. But that is a slight change for folks. Tldr right at the start of this process, we need to map the market of accounts that we want to sell to and the contacts within them. A Good fit. We used Good Fit for this. Right? I'm super familiar with it. There are other tools available and many of them of which I'm sure you're familiar with, Zoom info, whatever it may be. Let's get practical on market mapping though, right? So again, in front, in front of me there's a screenshot of Good Fit. This will be whatever tool it is that you're using, slight cheat code that we could use our own product. So I wanted to map my market of serviceable companies for Good Fit. The people who could use my product today, the people I'd be happy to sell to if they came in boundaries if you're not Good Fit. But like an expense management tool. An example I'll use later, maybe your map to market is every company in the regions that I serve with over 500 employees and three people in finance. If you're an employee of record, like deal maybe it's I want a list of every company hiring for remote staff or outside of office locations. This is going to look very different for every company and you need to work out what's true for you guys. But you want to get as tight here as you possibly can. A good thing. We've got a deep understanding of what's a serviceable customer for us. I need them to have an established sales team. So I've got sales count above five here. In my filters. I don't want people selling to enterprise only, I don't want people selling to local markets only and I don't want people selling to niche industries. These are things very unique to Good Fit. Right. Rather than establishing a set of filters to filter those people out, this is what most people would do. I basically trained a model to disqualify companies that those things were truthful in terms of how that worked. I got a list of 100 companies that were a good fit for good fit, 100 companies that weren't a good fit for Good Fit scraped the homepages of all of those companies and then trained a text based ML model to basically identify companies that describe themselves and look like the companies that I want and disqualify the companies that describe themselves more like the companies that I don't want. And this works so like 90% accuracy. I can throw any domain in it. It looks at how that company subscribed itself on its homepage, it looks at whether it looks like the ones that I want and if it is awesome, they're a qualified account for us. Most people would just use filters here. Don't let that confuse you. I just want to explain to you what you're seeing on the screen because in my filters here I've got is fit probability greater than 0.5 or 50%. This is basically include anyone who passed the Good Fit model and you will almost definitely just use filters map your market. However makes sense for you. If this is something you're struggling with, I can help you or good it can help you or whatever. It will look different for each company. Now most importantly, this delivered me 3639 companies that I could go and work. This is the size of my serviceable, kind of addressable market. I can flex it up and down based on some of those filters. If I reduce the sales count to three, it increases a little bit. Right? And having mapped your market and worked out how many accounts there are, you need to ask yourself a set of questions, right? The first one is how often per year do I want to ping someone or recycle my accounts? Conservatively, I said twice a year. I'm happy to contact one of these companies and contact them every six months. This meant that I can reach out to 7,278 companies per year. I've just multiplied my market by two. Now, if I assume I disqualify 10% of companies. Again, market mapping is never perfect. The companies that are in your list, you're not going to want to work every single one of them. My disqualification rate is 10%. My model that I just described works at like 90% accuracy. So 9 out of 10 times when it says someone is a good fit for good fit, it's right. But it doesn't mean I disqualified 10%. So again, I reduced the size of the market by 10%. It ended up at 6,550 accounts. I can work per year around 550 accounts per month. Assuming that I'm recycling my market twice per year. Hopefully that makes sense. Shout with me in the chat if you're not following. Now, having done that, given some assumptions again around funnel performance, I can start to predict how many replies and meetings I'm going to get off the back of this. Right? Things to note are with zero touch outbound, the number of accounts you can physically contact per day or per quarter is no longer the limitation. When I was running the sales team at Puddle, it was like, okay, if I hire a BDR, they can work 300 accounts per quarter, for example, in 2017, that was my constraint. In 2024, doing this without BDRs, your constraint is just largely the size of your market because we can automate getting in front of them. Right? And again, you can expand the size of your addressable market through product development and. Or you can just improve funnel performance. Right? That's the two ways to get more and more out of this approach. Next, I'm going to talk about audiences. So on to step two. Right, You've mapped your market of serviceable accounts. Now, within that market of serviceable accounts, we need to determine who we're going to outreach to. There's lots of different audiences or campaigns that you might might want to run. An audience should match a campaign. Many folks start with like competitive displacement campaigns. That's where I started way back. Way back. When you're using competitor X, here's where they're weak, here's what we do better. Some people run lookalike campaigns. We have a customer X in your industry. Here's what challenges they had and how we helped some people get as far as value based campaigns, right? Targeting companies who look like they have a pain that you solve for with messaging that suits to give you an example of a really effective value based campaign for those of you that don't know. Paddled as payments for software companies basically. And one of our strongest, most highly converting, highly resonating value propositions was helping people sell their software product globally. So the audience I built was get me a list of every software company with global traffic but only supported a single currency and a single payment method. Right. That was my audience to talk about internationalization and I could send them a message saying, hey, you've got a global audience in your site. Can I show you how we could optimize your checkout experience for that global audience? Right, That's a value based audience and campaign. Hopefully that makes sense. A good fit. We started with lookalikes and I'll explain why when it comes to the learnings in just a second. But I'm going to give you some examples of how to build an audience to begin with or how this looked like for us. So a good fit. We started with lookalikes again for some unique reasons. I'll cover in just a second. We have since started running some other experiments too. One of the lookalike campaigns we worked on was in cybersecurity. We had a number of companies working with us in cyber, so we wanted to find more companies in cyber that we could go and sell to. So within my addressable market, I need to build my list or my audience for the campaign. Again, I did this in good fit. You could do it in your CRM if the data lives there or in whatever tool that you're using. Building this audience for me looked like this. So I've still got my filters applied to ensure that they're a serviceable company. So fit probability again is still above 50% and sales count is still greater than or equal to 5. I then used G2 subcategories which I had enriched against every single account in my serviceable market to list all of the cyber security related subcategories. Right. I actually accelerated this a little bit with GBTS and LLMs. You could manually just click all the subcategories. To be fair, you only need to do it once. There's not that many of them. It's not rocket science if you look at some of those subcategories. It's things like Vulnerability scanner software, penetration testing software, a bunch of stuff in digital security or cybersecurity. Super simple, right? It's a lookalike campaign. This left me with 114 companies, right? There's 114 companies matching that set of filters that I just talked about that I want to be included in my lookalike cyber security campaign. Super simple. I'll give you an example of a different and simpler audience in just a second. But first, just some tips on audience building. Right? Yeah. Build the audiences in whatever tool makes sense for you. It's wherever the data lives. It could be in zoom info, it could be in your CRM. You want to be able to tag the companies that are in your audience with the audience name or the campaign name that you want to run them at or have them in like a specific list. Again, different tools on different CRMs have different nomenclature around this stuff. It does tend to be quite simple to do that though. But we basically want to be able to tag the people you want to run a particular campaign or in a particular audience on the contact side that I referenced earlier. So a good fit. We map the market of every serviceable account you can sell to as well as the contacts that you care about in those companies. Other tools don't work like that. They normally charge you per contact most of the time. So you can go into the tool filter for the accounts that you want to be included in a particular campaign or to be within an audience that you want to run a campaign at. And then you're normally charged per contact that you then want to enrich or gather against each of those companies. And you're normally charged on a per contact basis. That's not always the case. Rather than enriching your entire addressable market with every contact, that's going to be cost prohibitive for you. I'd only bulk enrich contacts against the companies that you want included in a, in a campaign. It's just going to keep your costs lower. Most of these tools have a way to bulk enrich a set of accounts with contacts. Not all of them are great. Some of them are good, some of them are bad. And again, this is something you can work out and I can help you with with just a little bit short for time. But I'd encourage you to only enrich the accounts that you're working with contacts just to keep your costs down. And I said this wasn't going to cost anything. I kind of assumed you already had a data provider. Maybe I'm exercising a little bit of Artistic license there a good thing we use Good Fit. So it did cost us nothing. I'll do more on cost later as well for when we do the ROI analysis. Hopefully that's making sense. To give you another example of like a value based campaign, I've been chatting to a bunch of folks in expense management recently. A few of them are buying Good Fit right now to run zero human touch outbound themselves and with them. I propose this as an example of like a value based audience. I assumed a big value proposition for expense management tools like Spendesk or Pleo is that you don't need to scale your accounts payable and accounts receivable team linearly with the number of employees in your company. If you're using their expense management tool, they make all the processes around receipts and invoices much better. Right. So the audience that I wanted to build, if that were true, were the audience would be get me every company with an existing accounts receivable or accounts payable team that's currently hiring for folks in accounts receivable and accounts payable. Why is that a good audience? The fact that they have an accounts receivable function means that they're qualified, they have a need for accounts receivable and thus accounts receivable tooling to enable those folks. But because they're hiring in that function and it's growing, there's an ongoing problem or an ongoing need there that they've put budget behind solving. Right? Really simple stuff. This is the type of company I could probably run a really, really simple campaign at. Again, just. Just like with mapping your market, what audiences you build are going to look very unique to you based on your value propositions or what type of campaigns that you want to run. This is just some examples of ones that I've seen built out and key learning. So around audience building, I think this is actually the part that you need to get right the most. We've already had a question around touch points, channels, messaging, CTAs. I think the thing that's even more important than all of that is your audiences. So let me just talk through some key considerations here. Number one, just try out a few different types of campaigns and different audiences, right? Try lookalikes, try value based. Try displacement of competitors panel. Value based campaigns and audiences worked really well for us. Why did I choose lookalikes that good fit? Honestly, I think the reasons were quite unique to us. People have been outbounded like data, which is what we do a hell of a lot. Most people are selling the same bad stuff Good fit is immediately and viscerally different to status quo based on the data that we supply. And if I know what industry you're operating in, I know that I can give you three example data points which are super useful to you that you've never had access to before, just on site, like it's just easy. And that means I can just stick them in every touch point I'm going to put in front of you and it's going to convert really well. Again, so very specific to good fit. What is absolutely critical to all of you is the tightness of your audience. Right. If there's one takeaway you have here on how to run outbound, well, it's your audience that you're running the campaign out or running the emails or whatever touch points you're using are. Your audience should be so tight that the message writes itself right. In our case, in goodness case, you're an X industry, so I can offer you Y data points that you don't have. Does that sound good? In the paddle example, it was you have international traffic. You aren't localized for it. Can I show you a localized checkout? In the expense management example I gave you, it was, you have a growing accounts receivable team. Here's how we help that team be more efficient or stop that team growing linearly. You should know what you want to say to that potential customer just because they're in the audience that they're in. We want our audience to super tight and the messaging will follow. It's super easy. The other benefit of this when doing the programmatic outbound, is that with tight audiences, you're going to have less dynamic content in the, in the sequence. It makes things less risky. If I know what I want to say to someone just because they're in a particular audience, I can have less dynamic and changing variables kind of in the sequence, as I said before, and I'll talk about that more in a second. People in the past, I think got away with loosely defined and bad audiences and then just threw massive volume the problem. And that's no longer going to work. So where are we at? Right. We've mapped our market, in my case, about 3,600 companies. I can ping twice per year. I started to build some audiences within that list of 3,600 that I wanted to ping. I went for lookalikes. I did like people in the ATF space, cybersecurity, space dev tool space, sales tax. And because they're in those and operating in those sectors, I know what message I Want to send them now. I need to go and write and configure those sequences and those messages now. I love and hate talking about messaging and sequences mainly because I think this is part of the outbound process that's been massively over indexed on over the past five years. There is so much talk about sequences, messaging touch points, channels, and nowhere near enough time is spent ensuring that the messages or your value propositions that you're talking about are being delivered to the right customer first and foremost. It's a huge bugbear of mine. I talk about it all the time. Sorry for those who you have heard me say that a million times. But anyway, the starting point for the messages again, I had some real epiphanies around this stuff and thought about it very differently and Michael, I've just got your question. I'll talk about that when it comes to the results properly. Thanks for asking. Keep them coming. I started to think about sequences really differently actually this time around it was very different. So in approaching the need to run outbound with no budget, no SDRs, no VC backing. Yeah, I had to have to do it in a different way. Like typically when we're trying to build our sequence or write our message, we're trying to write the very best sequence or campaign, whatever you want to call it that we can. And I realized really quickly that I don't have that luxury. Right. I don't have an SDR team to phone anyone. I know having been in a bunch of round tables that notions posted with sales leaders that phones are working quite well for people. I don't have anyone to call there to do the phone and I'd have loved to add to my customized loom or videos on the data that we could supply someone. As a part of my outreach, I didn't have the manpower to record the videos. Right. I recognized the cheapest and most scalable way I could run outbound was exclusively via email and LinkedIn. Touch points. These are channels that I can utilize without humans needing to drive the activities. There's tooling that can automate touch points across those platforms for me. They're cheap and I can scale them without humans and largely because there's tooling out there to automate messaging on those channels. Like my constraint becomes. Yeah, like the size of my market as opposed to the tools themselves. There are some constraints to those tools which I'll talk about in a second, as well as risks. But just bear with me, I'll get to those as I get a little bit nearer to the results. I did start to kind of pull on this rudder scalability and cost a little bit further though, right? Historically, BDRs and BDR managers, when they've been writing these sequences, have targeted an OPS created, right? Or sometimes meetings booked. Sometimes they're the same thing people. And as a result, I think it's in their interest to have the highest conversion they possibly can from contacted to OP created. That's literally what they're targeted on. They don't tend to think at all, or at least as much on the cost of any activity that they're doing. They just want as many people as possible within their campaign to convert into a meeting book or an OP they are thinking about or realizing that every dollar spent creating a loom is a dollar spent or time spent that could be spent elsewhere, maybe enrolling more folks or with another touch point of a different type. Every phone call, right, takes time. This costs money that has opportunity cost, right? And again, I tried to really dumb down the point I was trying to make to the extremes, right? Like as a CEO, I wouldn't want 10% more opportunities or meetings booked from from a particular audience as a result of said phone calls or loom videos for 50% higher cost or 50% lower throughput as an example, right? And I'd rather remove those costs and just reach out to more folks. And again, I've really dumbed that down. But I hope that's making sense to people because it was a real change in thinking for me. Like, rather than the starting point being like, let me go and write the best sequence I can, it was let me write the cheapest, most scalable one that I can. And this would be my suggestion actually, for anyone designing outbound campaigns moving forward, the very first thing that I would do is start with the cheapest, most scalable touch points that you can ask. This was linked in an email. Maybe it's something different for you. And in just utilizing those touch points, those cheap scalable ones, see what your baseline performance is, right? This should be the very bottom level of performance that you should expect because they're the most basic kind of touch points you can deliver. From there, you can add in, in a test, some more expensive touch points, having established that baseline. And when I say expensive, I mean in the sense of time and or actual cost. And then you can see the lift the phone calls, for example, give you and accurately determine whether that is a good use of time. What's the impact of adding in a phone call to the sequence? How many less people can I enrol or not? And what is the uplift in meeting book trait or whatever and you can make your assessment from there. LinkedIn does have CapSmarter. I'm going to get to the constraints in just a second. Let me talk to you about the sequences and how we wrote them. Then I'm going to talk tooling. So yeah, moving out of the theory to the practical of the sequences that we wrote about. So I wrote a sequence which was made up of six touch points over 11 days. I think to ANSW one of the early, early questions and honestly I just wrote something that felt pretty familiar to me. Again, there's some content around how I think about this stuff that's been posted earlier. I'm happy to run anyone through the end to end sequence or just share it. It's six touches across 11 days, four emails, two LinkedIn messages, nothing groundbreaking at all. All plain text, nothing in there. Now, because my audience was so tight, the message was easy for me, right, you're an X industry, you probably need Y data points. Can I enrich an example account with it? It was the same message in every one of the lookalike campaigns that I sent with slightly different data points depending on what industry you were in. Example on the left here in HR tech we have a customer in Pistonio who are in the HR tech space that I reference. And because you're in HR tech, you're probably interested in data points like the size of recruitment and HR teams, the volume of hiring someone's doing their ats, even keywords and employee sentiment on Glass Door, right? That's true for everyone in the ATS space. I can send that exact same message to everyone. My example on the right was in the expense management space that we talked about before. We've got customers like Moss and Clara and expense management. People like them are normally interested in size of finance and IT teams, open jobs in finance ERPs as well as proxies for spend like ads and tech being adopted. Right now there's no dynamic content. This is just the first touch, right outreach. But there's absolutely no dynamic content in that at all other than the company name and contact name, which is quite difficult to get right. And importantly, because the audience was so tight, I can deliver that same message to everyone in the audience. There's no dynamic snippets or dynamic examples in there. Based on some logic, I could try and compute like which vertical someone's in and then pre fill it with some sample data points. If I was trying to get super technical on this stuff, honestly it could just break or it's likely to break. The more I get into that stuff, the more likely things can go wrong. But with a tight audience, if you can just keep it this simple, you're much more likely to see some success and make some less mistakes along the way. I'll get more into the results and how I operationalize this later. I talk a lot about sequences in the webinar that's been referenced before. Next, I want to get onto tooling and constraints. Some of the questions that we've asked. So I've mapped my market. I've built some audiences. I've written some sequences of campaigns using LinkedIn and email touch points only because they're cheap and scalable. Next, I want to put the tooling to work. Right? This is the difference. I previously put my reps to work, now I want to put the tooling to work. How do I do that? To start with, every single account that you're going to work should be in your CRM if they're not in there already, right? This is just golden rule of all sales outreach or sales and you want them tagged with the campaign that you want to run them in or make sure they're in a clearly defined list with the campaign that you want to run them at. A good thing we use HubSpot. All 3600 accounts that are in my addressable market live in HubSpot. And then I use HubSpot's list builder to build a list of folks that I wanted in a campaign. In the screenshot that you're seeing here, I'm filtering for anyone with ICP DevTools. This was my tag for anyone in the Dev Tools lookalike campaign. I had a specific sequence written for the VP Sales and BDR Persona. So I'm filtering for that too here at contact level. I probably also add another filter in here saying that I don't want them contacted in the last 180 days or so to ensure that I'm not pinging them too many times. Right? Should be. Should be pretty simple. The thing that is interesting here that you want to take away is that you want to build your audience an account level. You need a certain number of accounts in there for it to make sense. I think I had a question earlier on how many accounts should be in an audience. Actually, some of the tests that we ran were really, really, really narrow. Like down to like 30, 40 companies. And honestly the answer to that question is based on two inputs I think. One, what is the level of effort you're putting into actually setting up a sequence, right? Like, the higher the effort, the bigger the audience needs to be for it to be worthwhile. And then two, what is your funnel performance? Right. Like, if you can apply your funnel performance to the number of companies that are in a campaign, you can work out, is the output that I'm going to get running a campaign audience size X based on my funnel performance worthwhile? It's all an ROI on time, quite frankly, and you should get to grips with that over time. Importantly, when some of these sequences work, you almost want them to become evergreen. Like, if my lookalike sequences work, I can continually enroll new companies that emerge that fall into those categories or. And when I recycle them, and again, maybe we can, we can dig into that a little bit later. Thing to take away, though, is that you want to build your audience at an account level. Make sure there's enough accounts in there that make sense for you, and you want to tag those accounts with as the audience or with the campaign name that you're. You're running them in. But then you want to create your list, usually at a contact level. And that's because the tools that do the enrollment normally work at a contact level, not an account level. So the list you want to get the list of contacts that you want to enroll at, the companies you've tagged with your campaign, that makes sense. You want to export this list of contacts to wherever you're running your sequences or your campaigns from. So we actually started in HubSpot. We used HubSpot to automate all of our emails for us, and we handled the LinkedIn touchpoints manually to begin with. As of the last few weeks, we moved over to Lemlist because it can also fully automate our LinkedIn touchpoints, both the connections and the messages. I always knew that we could do that. I just didn't want to invest in working out how a new tool worked before I saw some results. So right now we export our list of contacts at the companies that are tagged from HubSpot and import them into LendLists. Let me just check this question in the chat before I keep moving, Mike. Again, it's all to do with the sequence that you're writing, right? Like, if your audience is super tight, you should be able to deliver them, hopefully, a message which resonates and maybe we can, we can talk more about that at the end. There's no golden bullet as to how to do that. But ultimately, if there's an audience that you want to sell to, that's super tight. You should be able to deliver them a valuable message based on them just being who they are. That tends to be how I think about things, Lemon. Keep talking about tooling, though, in the meantime. So I'm not going to demo you Lemlist just for the sake of time. There's other tools out there as well. We set up our sequences in their product. You can see one mapped out partially on the screen. These things get quite big and complex, but ultimately you're using lemlist to enroll accounts and contacts from your list of folks that you want to work with from your CRRM or from CSV. However you choose, choose to get them in there, right? The concept is simple. You build a list of. You build an audience of companies that you want to run your campaign at. You export the contacts that work at that company that you want enrolled in the campaign, and you upload them to the tool that you want to use to automate your LinkedIn and email touch points. We chose Lemlist. There are a bunch of others out there. There's even tools that specialize, right? There's some that specialize in the LinkedIn touch points, like Phantom Buster. There's some that specialize in email, there's some that do both. I pick Lemnist because it did both and it just made my life a little bit simpler. And we've had some questions on constraints, and I also want to talk about risks, right? As much as we want infinitely scalable outbound, we're no humans. There are some constraints, right? And also you're doing all of this stuff at your own risk and you want to scale it based on your risk Appetite. On the LinkedIn side, a single LinkedIn account or a person on LinkedIn can connect with 100 net new contacts per week. You can actually, if that's a 200. If you also connect with people on the mobile app, they're separate accounts. We're using lemnis to automate LinkedIn connections and even automate our messages. But based on the volumes that we're sending and volume of people that we're connecting with, we think that's super low risk, right? Some people may choose to manually send out their LinkedIn touch points to de risk this. Some people are fine with automation. The biggest risk, as far as I know, on the LinkedIn side, is that you get shadow banned on your posts if they're important to you. Your impressions will start to drop if you start doing this stuff at crazy volume. For us, this was a risk we were willing to take. So we went automated because we're sending really low volume, so don't think it matters. We're not sending anything from my LinkedIn profile. We did during some of the tests, but not anymore. And mine's the one with the reach. Also, if you're really worried about that, you can buy LinkedIn profiles that work for this stuff that look very good if you, if you're going really crazy. Email is more scalable than your LinkedIn touchpoints. But again, when sending significant amounts of email per month, you should definitely seek advice on this stuff. I'm feeling good, a good fit. I'm going to talk about volumes in a second because we're sending under a thousand emails per month, basically. I'm not an expert here on some of these constraints and some of the risks. I'm learning a lot each day. Some advice that I give to all of you though is that you tend to want to run your outbound cold email from a separate warmed up domain than your primary workspace. So for us, our OODFIT IO kind of email and Gmail workspace, you can do that yourself or you can like, you can buy workspaces, you can buy domains, things like Lemnis help you warm up those domains. They're very good at delivering email. Just do your research. Again, based on the volumes that we're talking about, I don't think we're ever going to get banned on Gmail from sending emails. Like we're not sending tens of thousands of emails per month. But again, we just felt that Investment Giver is going to take us a couple of hours. Made sense. Just risk versus reward. And there are some really good companies out there that can help you with this. So I'd recommend a company called Allegro. They're really strong at deliverability and knowing about this stuff as a bit of a disclaimer. They're a good fit customer and I'm an investor. But I chatted to another great company about this just last week. They're running outbound at crazy scale for really big brands like Reddit and Localize have set up the infrastructure to send a million emails per day. They're called Afternately to know really, really I'm learning a lot. This does, this stuff does come with. If you do it, don't go out and send hundreds of thousand emails in two days. Let me keep it going though because we can talk now. Rich, Beside Lennox, there's quite a few. There's like Phantom Buster and Expandy are two others. Expandi, for example, are one of the examples that allow you to connect with people, both mobile and web. Cool. So we're basically there, right? We're nearly there. And it's a little overwhelming, right? No generative AI once, right. I've written a sequence for each one of those audiences. I actually went down to a Persona level. I wrote a sequence for each Persona within the audience and then I started to put my tool into work. Right. I'm using Lemlist, not Styles and send out my LinkedIn messages. My voice is getting a little bit, a little bit slow. So I'll just slow down for a second and hope you can catch up. I can tell you, Phoebe. So Allegro are really good at deliverability. The company that I spoke to about this stuff was called Atom 8 Leads. They'd built the infrastructure to send out a million emails per day. They're a very enterprise solution. They work with like Reddit and energy companies on this stuff. But they taught me a lot about the risk around sending a lot of high volume email. There's so much content on this stuff out there online. So just read into this stuff. Don't go out and send a hundred thousand emails like tomorrow because you'll get in trouble. So just think about your risk tolerance and volumes. Volumes is the main thing. We're basically there though, right? Like I'm at my market, I built my audiences, I've written some sequences and now I want to put my tooling to work. And it's the tooling that I'm putting to work to connect with people on LinkedIn, send my LinkedIn messages and send my emails, not my reps. Next, you want to measure how you're getting on an iterate, right? So I can start to tell you about some results now and give you an idea of the volumes also that we we're talking about. So in month one, and this is my first month, we enrolled a cohort of about 250 companies and 480 contacts. As I said, low risk in terms of volumes. This is, this wasn't huge. They were all enrolled from Sophia, a colleague of mine called Waheed, my VP of Revenue and my own LinkedIns. Due to my ego, I actually thought my LinkedIn connection rate was going to be higher than this. It wasn't. The results are pretty consistent. The three of us shout out to Sopia, my founders associate. More broadly, by the way, she saved me a lot of personal effort on the heavy lifting on some of this stuff, including the analysis. A real superstar TLDR. I saw a 6% actionable response rate on the on the outbound that we did this was responses that we could do something with. And honestly I was pretty happy with these results. When I was building this for the first time and only using those cheap scalable touchpoints in LinkedIn and email, I assumed and I guessed I'd only get to like a 3% response rate. I wanted to get to 5% in my end state and I assume that I need to do some kind of minor human bits or intervention in order to get there. We exceeded that on a very first attempt and it could actually just go up a little bit further. There's a few lagging replies going on there for context. Back in 2017, with completely manual outreach again templated through sequences, we were seeing highs of like 20 to 30% response rates. Those days I think are gone. But it was also a much more expensive and less scalable process powered by humans. I expected a 40% meeting book trait. I felt that was pretty feasible based on my experience. So 40% of the people that replied I expected to be able to book a meeting with. I was pretty much bang on with my expectations there. We're just under 40% right now, but with some meetings being scheduled that could lift to about 57%. They're currently being scheduled right now, so that's pretty good. I was less sure on my meeting to operate. Right. I'm actually though outperforming our inbound funnel there. That performance of like 40 to 50. So from call to op informs at 40 to 50% inbound. I'm actually 80% on my outbound and this kind of makes sense, right? Like I've got a reference customer. My data or product fit for the customers that I'm outreaching to is very strong because I've hand picked them. Like if you're an employer of record or in the HR tech space, I know I've got killer data for you that you don't have. It's the benefits of being really specific on who you're targeting. So that that is pretty phenomenal that 80%. Now this is our first go at this, right? We've since continued since this first cohort that I've analyzed here and I'm seeing consistent performance like this has continued whilst experimenting with some new different campaign types. Again, my enrollment numbers are pretty small. Initially it was just a test, but the funnel performance is really, really good. Getting 5 ops from 250 companies and 125k pounds worth of pipeline based on good fits. ACB would be different for you is pretty good. Moving into July, I want to enroll 500 companies up, up from the 250 that we enrolled in the initial test. And that's actually my monthly limit to assuming that I don't want to touch companies. Well, I can recycle them every six months based on my current market size. It's also well under the scary thresholds for volume. If you're sending out less than a thousand emails per month, I think you're pretty much under those scary thresholds. But again, seek advice on this stuff if you're getting close to that. And I do think I can do better than the 6% given my learnings from this first test. Quite frankly, we're kind of into the optimization phase either way, headline numbers are 250k pipeline created per month. If I'm enrolling 500 companies instead of 250 can feed good fit for a really long time. Especially with our win rates which are super, super high in terms of cost. Right. Like call it about 30 grand. Good fits average contract value. There are some higher, some lower. Is about 25 grand and I'm paying lemlist 5 grand per year. So for a 30k investment I'm getting what, what like 3 million in pipeline per year? Like pretty good ROI. I'd say less than a cost of a single SDR. We're super excited to keep scaling this internally and I really wanted to share it with this group because I just feel like people can replicate this stuff. I think the real proof is in making sure that you're able to map your market effectively, then build some really good audiences within that market. Like what are the campaigns that we want to run that we think will resonate based on what we know about these companies and then use the tooling to automate those top of funnel costly activities that humans were doing in the past. I see there's loads of questions appearing in the chat. I'll come back to all of those because I think I'll have a bit of time at the end. I've got one last slide. Right. So next I want to scale and monitor my performance. As I said, I'm continuing to run lookalikes to see if that performance is maintained and I think it is. I'm testing some other types of campaign, like competitive displacement. Yeah. And optimizing from there. Some stuff that I want to experiment with, which I haven't mentioned is just auto enrollment. Like this again removes a lot of the heavy lifting for this stuff. I basically want to auto enroll a company in a sequence when it's been tagged and not been contacted. So for example, I want to auto tag any company that emerges that's in dev tools if I'm not and automatically enroll them in a campaign if I've not spoken to them in six months. This will catch the net new companies and the new contacts in those companies and those ready for a cycle, et cetera. Just removes even more human involvement from the campaign coordination. I know trigger based campaigns also work. I'd like to experiment with those. Like when someone starts hiring for SDRs, a good fit. This is a great sign for us to contact them and I'd want to ping them around feeding that SDR team. These are some other things I'll experiment with. Ultimately we'll iterate on what's working while trolling other campaigns and at some point we may even look to broaden the size of our market to again just get more and more out of this particular process. That's working, working pretty well.