Emir Atli on this episode
Emir Atli is co-founder and CRO at HockeyStack, a company that scaled to over 300 enterprise customers in under two years while multiplying revenue 4.5x in a single year and closing a $50 million funding round. He built his first product at 18 from his parents' living room in Turkey and has spent his career obsessing over go-to-market operations and motion design.
HockeyStack's core thesis is that you cannot layer AI onto a broken go-to-market motion, and the operating discipline behind motion architecture determines whether AI delivers ROI or creates expensive chaos. When Emir first inherited the SDR team, he discovered reps spent 80% of their time clicking between ZoomInfo, the CRM, and Outreach—doing data entry, not selling. Rather than bolting AI onto this broken system, he ripped the entire motion apart and rebuilt it from scratch with AI handling everything that wasn't a human conversation.
The reframe is fundamental: most companies approach AI experiments by delegating them entirely to leadership and ops, which guarantees failure because leaders don't know what's actually happening in the field. Instead, Emir builds experiments as 30-day MVPs with 10-person pods that include frontline reps, sales leaders, and ops—not just executives in a conference room. He measures success not as immediate quota blowout, but as hitting 70% of prior quota attainment with a deliberately broken system, which proves the hypothesis is sound and just needs iteration. Once parity is proven, the system compounds over time.
The episode also covers where AI should handle outbound versus where humans must stay (pain discovery and multi-threading never automate), why delegating AI experiments to leadership is the single biggest mistake, and how making sales engineers full-cycle solved the expansion handoff problem that kills retention expansion. The final insight: most companies don't actually know how they close deals, so Emir's advice is to study your top 20 deals, reverse-engineer the blueprint, and build enablement around that real process, not the one that lived in Salesforce five years ago.
What we cover in this episode
-
2:21
Building GTM engine for 4.5x growth How HockeyStack scaled revenue rapidly through inbound demand, then layered outbound motion on top of existing pipeline
-
7:10
SDR time wasted on admin work Emir observed 80% of SDR time spent clicking between data sources instead of selling, which triggered motion redesign
-
9:02
Building automation from Clay to custom APIs Evolution from manual workarounds to Clay-based automation to custom MongoDB API pipeline for intelligent account and title filtering
-
14:27
Cold vs. warm outbound systems Distinction between cold outbound motion (custom-built) and warm outbound via HockeyStack's prospecting agents for different ICP segments
-
19:31
The biggest AI experiment mistake Delegating AI motion design to leadership and ops instead of involving frontline reps and sales leaders who understand field reality
-
27:22
Where AI handles vs. where humans stay Automation stops at human judgment, scale impact, and talent development; pain discovery and multi-threading require human brain
-
39:24
Full-cycle SEs and fixing expansion handoff Sales engineers now own accounts end-to-end including expansion, eliminating context loss in traditional handoffs to account managers
-
46:04
Daily metrics and why forecasts lie Emir tracks daily meetings booked, pipeline movement, and fires; ignores quarterly forecasts because control only extends two weeks out
The lines worth sharing
The single biggest mistake that people do is you delegate this entirely to leadership.
Emir Atli · 9:16
80% of the time was spent between ZoomInfo, CRM and Outreach, like clicking on buttons, which again, now that I think about it, is not that surprising.
Emir Atli · 8:04
You should never experiment with your foundations unless you have a replacement for that foundation.
Emir Atli · 31:04
The line for me is: is this something that requires human judgment? Is it scale where it impacts my bottom line and or talent development? If the answer is yes to any of those questions, I'm not going to automate it.
Emir Atli · 37:44
Common questions from this episode
How do you design your first AI experiment in sales?
Build a 30-day MVP with a 10-person pod including frontline reps, a sales leader, and ops. Guarantee them 70% of prior quota attainment while they run the new motion, removing quota pressure so they actually test it instead of reverting to old playbooks.
What percentage of SDR time should be spent on selling vs. admin?
Ideally 50-75% on actual calls and conversations. If reps spend 80% clicking between data tools, your process is broken and needs to be ripped out and rebuilt with automation handling data work.
Why shouldn't you delegate AI motion design to leadership?
Leaders don't see what happens in the field daily. Without frontline rep and sales leader input, experiments will be irrelevant to actual problems. You need one sales ops leader, one sales leader, and one IC at the table designing experiments.
What should you never automate in sales?
Pain discovery, multi-threading decisions, and any CRM field requiring human judgment that impacts deal progression, compensation, or talent development. These require critical thinking and contextual understanding that erodes if fully automated.
How do you know when to stop experimenting and rebuild foundations?
Keep your top 3 revenue motions stable and never experiment on them unless you have a replacement ready. Experiment with new motions or channels. If your proven motions break, that signals a market or brand problem, not an AI problem.
Should sales engineers stay full-cycle or hand off after closing?
Make them full-cycle for mid-market and enterprise. They own expansion success after closing. This preserves champion relationships and context, avoiding the costly exercise of account managers re-earning trust with the same stakeholders.
SEO meta description
Emir Atli, CRO of HockeyStack, shares how to build AI outbound motions from first principles, delegate experiments correctly, and scale revenue 4.5x in one year.
Target keywords
Emir Atli
HockeyStack
AI outbound motion
SDR automation
sales process design
go-to-market engine
revenue operations
sales engineering full-cycle
AI experiments in sales
CRO strategy
enterprise sales motion
quota relief experiment
Read the complete conversation
Auto-generated from the episode audio and lightly cleaned. Expand to read or search inside.
Read the full transcript · 52 KB · Emir Atli
RACHAEL0:00How do you propose that people approach
RACHAEL0:03designing that first AI experiment?
EMIR0:07the single biggest mistake that people do
EMIR0:09emotional like an experiment like this is
EMIR0:11you delegate this entirely
EMIR0:14to leadership
EMIR0:15relapse.
EMIR0:16like literature in general
EMIR0:18by definition,
EMIR0:19don't know exactly what's going on in the
EMIR0:21That's just the truth.
RACHAEL0:22where do you draw the line on what you'll have AI handle versus what stays human?
EMIR0:27the line for me
EMIR0:29is this something
EMIR0:30requires
EMIR0:31human judgment?
EMIR0:32It is scale where it impacts my bottom line and or talent development.
EMIR0:36If the answer is yes to any of those questions, I'm not going to make
EMIR0:40You should never experiment with your foundations unless you have a replacement for that
EMIR0:43foundation.
RACHAEL1:06Today on the Crow Stories segment, I'm sitting down with Amir Atli, co-founder and karo at Hockey Stack. Amir started his first product at 18 from his parents living room in Turkey. Went through Y Combinator and has scaled hockey stock to over 300 enterprise customers in under two years, including a $50 million funding raise just last month. As of this recording, he runs both sales and marketing.
RACHAEL1:30Taught class on attribution at Stanford, and his company's entire thesis is that you can't layer AI onto a broken go to market motion. Today, we're talking about how Hockey Stack built their go to market engine, specifically around the outbound motion to facilitate this rapid growth. So thank you so much for joining me today, Amir.
EMIR1:48Thank you so much for having me.
RACHAEL1:50Yeah, I've been following Hockey Stack for a while now on LinkedIn, and I just love your guys's content that you put out. The videos are hilarious.
EMIR1:57Thank you.
RACHAEL1:59So, Amir, you helped multiply hockey stocks revenue by about 4.5 times in a single year and closed enterprise logos like outreach and RingCentral, companies that have massive go to market organizations of their own. Before we get into the AI portion, walk me through what the go to market engine actually looked like. That made that possible.
EMIR2:21we have an open motion. We have an SD team. We have a motion that includes like social field marketing. We are very heavy on content as like the main source of inbound pipeline together with like the inbound motion. I'm very active on LinkedIn, so I built 43,000, I believe now followers, only ten on my profile over the last 18 months.
EMIR2:49And then we have an expansion motion as well from our existing customers. So that's like the go to market motion that we run right now. The top two channels that we have in terms of sourcing that new pipeline that closed the fastest at the highest ACV are social and app owned.
RACHAEL3:06so what did the motion look like when you first came in? Was it this structured before or did you have to build this all from scratch?
EMIR3:13Yeah. So
EMIR3:14as I'm a founder, there was nothing before me. But like there were stages of or like phases of the build out of our go to market. So it first started as 100% inbound for like about a year or so from our funding to post series A, it was 100% inbound. Then a couple months after our series.
EMIR3:38We actually started to build our up on motion. So the first phase was literally my LinkedIn, the small following and then supporting that LinkedIn followers that I was like following that I was building with paid media and then referrals and word of mouth was like our growth from zero to post series and then posteriors. We saw building that function.
EMIR4:03So the first was we hired to stress and share leader. And then at the time I think if I had to do it again, I would do it the same way. Like starting with inbound. And I think because when you do that and then when you bring it on up on function, then you have a lot of inbound demand that you enable to serve, which is like the first like experiment of up on.
EMIR4:27So for example, for like a year we had a lot of traffic, a lot of followers, comments, likes, engagement. We have an interactive demo on our website. So like thousands of people checked it out, we never reached out to them because we had never seen a function. I didn't have enough time to do it on my own. So there's like a lot of that demand that option can cater to as like the first like 3 to 6 months motion.
EMIR4:53So we did that. Exactly. We did that. So the first three months was essentially like, how do we reach out to people in the shortest amount of time possible and then test our messaging because it's the first time that you actually or like the prospect actually talks to someone before the first demo. So the first three months was just like, what type of script works well?
EMIR5:14And then we found out that calls or like calling people and stuff like email or LinkedIn was the most important or like the best channel for alpine, for hockey specifically. So once we figured it out, then everything was about, how can we call the highest amount of highest number of people in the shortest amount of time possible with the best script and alchemy, hire people and train them in a very short amount of time to get them on the phones.
EMIR5:42Yeah. So then after the first three months, then we turned the inbound to option motion into more like a system that requires maybe like 20% of our time, then posts the first quarter. Then it was all about again, like, how can we call as many people as possible in the shortest amount of time possible with the script, and then we scale it from there?
EMIR6:04The first team of like three became five, eight, nine, and then some of them graduated into our A team and account management teams.
EMIR6:14And then right now, like 50 to 75% of their time is spent on calling support our sales functions. Yeah. And then after that then SDR team was on a in a good place. We of course like we continue to improve that.
EMIR6:32And we can talk about how we improve the return into a system. Then we open, we start opening up new channels and new experiments all the time, which contributed a lot to our growth. So like right afterward, right after we figured out the motion, then we opened up field marketing. Then we opened up other add channels. Then we opened up internal influencers.
EMIR6:53Yeah, just scale from there afterwards.
RACHAEL6:56So you went into building the motion knowing that you wanted to create this massive scale and likely with AI. So how did you build those foundations to make sure that AI could actually give you the ROI that you're looking for from it?
EMIR7:10Yeah. So that that was like, let me start the function. It was the first time that actually I actually personally interacted with an SD team. So we hired an SEO leader and then we hired monster first after like a couple of weeks because he was we didn't want to hire an STR, but it was a he was an incredible profile.
EMIR7:36And he actually reached out to us. Devin, who was still here as an AI in our team, was our first STR, so we had to hire him even though it was the wrong time. But he exceeded all of our expectations. So the first thing was like when we hired this team again, it was the first time that I enacted with STR team.
EMIR7:53One of the things that was really surprising to me was like 80% of the time was spent between zoom info, CRM and outreach, like clicking on buttons,
EMIR8:04which again, now that I think about it, or like if you have been interacting with STR teams, it's not that surprising. But like as a first principles thinker, it was like very difficult for me to understand why this is happening.
EMIR8:16Then I brought up this idea of like, why are we spending this time? Like, what can we do about this? And this is like end of 2024. So before all this cloud code and everything right now that we're going through.
EMIR8:27So, first, like no one believed that this can be automated. Again, this was before the zero of everything is automated end of 2024.
EMIR8:36Then I started working on myself with two other people on my team
EMIR8:41to essentially like understanding what is this process? What does this what does this process look like right now? So essentially the process is you build an account list for every year, but you see more data with themselves. Then they click on bunch of buttons to get emails and everything from zoom info to export to CRM, from CRM expand to outreach from our state right sequences, and then they send those sequences.
EMIR9:02If you have a dialer, those those contacts also go into the dialer. So when you think about it back then, this was like, I believe that this was a highly automated bull process. Like you can make this automated. So we built the version one on.
EMIR9:19Clay and cut like bunch of CSV essentially. So we picked an STR from our team as like the experiment. So I like we watched this person doing this work. We documented everything. And then I had one person who turned us into a clay table. And then from that table we experimented with it. The first couple of weeks was like very rough because there's a lot of work that an STR like an AI needs to do on the background that I didn't notice where like especially in the go to market, there's a lot of different titles.
EMIR9:57Like for example, you need to know if a senior manager of real ups, like if you need to reach out to a measure of robots and repeal probe ups, or manager of sales ups like differences between titles is very important. And that is only something that you can gather by working at a company for a long enough time, or it needs to be highly documented, which back then we didn't have it documented.
EMIR10:21So there was a like for example, another example, another great example for this is like there's a difference between review of sales or review of North America sales. So like if, if it's an Ivory P of sales, it depends on a company. But like an RV of sales might mean that they have the entire like worldwide RV sales, like or like RV of North America sales is more senior than Ivory P of West sales.
EMIR10:50So a lot of those. So the version one was like literally dozens of hundreds of prompts that like filtered those titles and made it possible for us to get to a version one. Then we bought a dialer, and then from there it was like V Corps week. We create a selection all over, like people would put in the wrong titles, wrong people, problems that they're having.
EMIR11:13And at some point we outgrew Clay. It was great for like 1 or 2 people afterwards became incredibly painful and incredibly time consuming and very expensive. So we hired a an engineer.
EMIR11:32Like working on the growth side of a side of our business to take this and then turn into more like a API led process that we stored in MongoDB ourselves. Yeah. And then from there, we essentially turn into a pipeline of huge amount of data because all of these APIs are public. So we essentially get all of our all of the data that we need from different APIs.
EMIR11:57And we have our own custom prompts and filtering that filters everything storing in MongoDB. And then we pipe that into outreach and looks like outreach we use for sales engagement and space for dialing and then everything back into our CRM.
RACHAEL12:12Okay. Were you I'm swimming. You're like segmenting these. You know, you mentioned knowing the differences between specific job titles. So you're segmenting between these specific job titles so that you're able to get the right information on them to create the right scripts and prompts for your sales team. So they don't have to do like hours and hours of research before they even pick up the phone.
EMIR12:33Yeah, exactly. And again, I was very fortunate enough to hire great people who run this idea that I had, and it was like a very simple idea. I just thought like,
EMIR12:43this shouldn't be the case in 2024, that we spent all this time and our head of sales open. Plus our engineer was working on the incredible job. Yeah, in the simplest form, it is finding like essentially you have a list that usually companies don't update at all.
EMIR13:02So we built a system that we actually update our temp every month. Because if you have a more serviceable market, then it is very obvious that new companies enter that temp or exit that time. For example, if you are targeting 100 to 3000 employee companies, there are companies who grow from 80 to 100 over the last three months.
EMIR13:23So you need to add those. Or is there a certain companies that downsized for that you need to exclude from that list? So we continue to update our temp. Then we continue to find the people that we interviewed chat to, and then we pipe all of that data into our tools for engagement.
EMIR13:39Yeah, that's like the foundation of it.
RACHAEL13:43And did you guys include intent based signals into the things that you API for
RACHAEL13:48in order to, let's say, like stack rank the accounts that you're calling on?
EMIR13:52Yeah. So this is like the full on cold up on motion. And then we run more like a warm up on motion. So the warm up on motion that we have is run on hockey sack. So we have prospecting agents that does that portion very well. So essentially like they service anything from third party signals to first party signals to people that we know and multi-threading all of those, it's more like targeted.
EMIR14:18So it lives in stack. And then we have the cold up on motion that's custom built, because that's something that I think you need to build internally.
RACHAEL14:27That's cool. So how how do you use your own product to do this? Walk me through the parts of your product that you guys are actually using to help your own about motion.
EMIR14:35So we have prospecting agents and deal agents. Prospecting agents are to prospect and deal to run deals. Our prospecting agents are essentially we have one agent per account, and the entire job is to break into the account. So all of our stars have a tap in a hockey stick or Salesforce very nicely. It shows and target the tasks that are specific to the accounts, and then they need to go into those tasks and execute or change the task and execute them.
EMIR15:06So what it looks like in practices, we have, again, because we have it in the market for some time now, we have a lot of data on usage, prior relationships, intro pads, third party data, first party data. So anything from like this person bought from us twice and another day, another ICP account and they have posted about the AI transformation to discount has been using archetype pretty extensively, but we have been talking to their seem not their sales teams, or there's an expansion opportunity there to third party signals like anything from their LinkedIn post to hiring everything in between our service to our team automatically.
RACHAEL15:49so you guys built this open motion, you know, alongside AI. It wasn't like you had an outbound motion for years and years. And then you tacked on, you know, AI tools to help it. What's your take on people who are already running companies that have, like these legacy outbound motions? And now they're trying to incorporate AI into existing outbound motions?
EMIR16:15probably get like a team. It depends on your company size or like your team size. Let's say you have 100 jobs.
EMIR16:21For this to be statistically significant, you probably want like ten people. I would get those ten people, build a version one first before getting those ten people. But the version, one of what you want to do, get those ten people, tell them they will be guaranteed to hit 100% for a quarter.
EMIR16:36The only thing that they need to do is prove or disprove that this motion works. It might take it would probably take less than three months. So if you want, you can also do one month of relief. And then I would want that team of ten people to be from diverse backgrounds. Like I wouldn't pick like all top performers, because then you don't really know if they're top performers and that's why it works.
EMIR16:56So pick like different backgrounds, different quarter profiles or attainment profiles and then put them in. A team powered their entire motion with the thing that you're building for a pond. And then in a month or so, if you move fast enough, you would see results or you don't see results. But I'm like 100% sure that if you do it right or if you don't do it right, if you move quickly or move fast enough, then you can see you will see great results.
EMIR17:24Then once those ten people, once it's proved on those ten people, then I would gradually rolled up to rolled out to everybody. I think it's impossible to like, not change anything and then just add a system like this, or AI in general to the existing system. It it also applies for anything that not just up on running deals, expansion, whatever you do, you just need to kill everything and then start from scratch.
EMIR17:48In my opinion, at this stage of the market,
EMIR17:51yeah, that's how I would do it.
RACHAEL17:53So let's back up a second. You said tell them that you like, guarantee them 100% that they'll hit 100% of quota. How do you make that guarantee when this is an experiment?
EMIR18:02I'm saying like quote like some like a quota relief.
EMIR18:05okay, they're only job is running this motion. And if they're also thinking about quota the, the hard thing like unfortunate unfortunate thing about running a sales team is a sales team is optimized to hit quota in their minds. So for example, if you say there's an experiment and you run this experiment for a month, and if you don't say like you don't need to worry about quota whenever you're not around, they will go back to the old motion book meetings.
EMIR18:32When they're done with hidden quota. Then they will they will test your pilot. So it needs to be fair for both parties where you need to go to something for them to be able to test this new thing, because it's very likely that in the first one it's not going to work.
RACHAEL18:48Okay. That makes a lot more sense. So yeah, because salespeople, you know, they got to put food on the table. So they're going to go after the thing that's going to make the money. If they're if they're judged down by coda, of course, they're going to try and go after quota. So if they're only focus needs to be running the experiment and doing as good as job as they can while following these processes, and they're going to do that.
RACHAEL19:06Okay. So so you have these ten people. You tell them, you know, you don't worry about quota like you'll be covered. Just run this process exactly how we've designed it. How do you propose that people approach designing that first AI experiment? Like what should they build when they when they're first wanting to integrate this into their motion?
EMIR19:31Like the the single biggest mistake that people do and emotional like an experiment like this is like you delegate this entirely to leadership or relapse. So both of these are highly dangerous because like literature in general by definition, don't know exactly what's going on in the field. That's just the truth. So if you say, hey, VP of sales and vPro ops, let's come up with what this motion needs to be like.
EMIR20:06And if they sit in a room for 30 minutes with no involvement from any IC, I guarantee you that the outcome or like the output of their project, is not going to be relevant to the people in the field, especially as a large company. You have different regions, different business units, different products, different line items, everything. And each of those teams have different problems.
EMIR20:25So there might be strong overlap between the problems or between the teams of problems and challenges they run into. Still, it's impossible to do it right without any ESI moment, so I would probably pick a relapse leader. Good systems, a sales leader in a pod. So, for example, let's say you pick one of the customers that we implemented something similar with prospecting agencies.
EMIR20:51This is a global company with like $600 million in revenue, with like seven different regions and lots of different business units and product line items and sales and everything. They even have different ops leaders for different teams. So we picked a region which was.
EMIR21:13A part of 15 people and a sales theater and a team lead who was working very closely with was like an icy plus, like a player coach type of station, which is very common in all sales teams. So those three people in us, we built something for them, built as in like deployed architect prospect agents. And then that pod became very successful.
EMIR21:34And then now every region, every unit is running on architect very similar to that. I would pick that team, make sure that there's a sales theater, there's an opposite or there's a player coach for knows the field and then build like a 30 day MVP plan. And then in building that, I would think about like, what are the top three challenges that we are trying to solve?
EMIR21:53And there's an emotion
EMIR21:54then run with that part of people.
RACHAEL21:57So what were some of the results that you saw when people did these experiments in terms of, you know, reply rates, pipeline generation, maybe even revenue, as much as you're able to say when they're just when they're doing the experiment portion of it.
EMIR22:13So essentially like again, it depends on the challenge that you have. But for example, for us it was let's say let's say an SD quota is ten meetings a month
EMIR22:25your team is averaging 90% attainment rolling over the last six months. 80% of their time is spent on prospecting, and 20% gets them to nine nine meetings on average a month.
EMIR22:41So then in my mind, if I can cut that time prospecting, of course we can book more than nine meetings on average, probably. But in the first 30 days, that system is going to break all the time. Not going to have the right titles, right accounts. It's going to frustrate a lot of people. So it's not going to be like, exactly.
EMIR22:59We will go from 9 to 15 in a month, but it would make like what would make me happy and satisfied in those first 30 days is can we get to like, for example, let's say 70% of our average attainment right now with within this 30 days, while not using our old systems, would give me confidence that this we are hitting 70% with a MVP version of it that has a lot of problems.
EMIR23:30Now, I believe that if I put in more effort and more obsolete resources and more like enrollment from other people, and if the system actually works, 70% is going to go to 150%. So it depends on like again, the quota team, like you cannot expect 30% quota to go to 100% in the first month with a broken system, and it's going to be broken in the first 30 days.
EMIR23:50There's no way around that. But I would like average at like 70% of prior quota attainment average rolling 6:06 months. And if we're story days and if we can hit that or come close to that, that means that the hypothesis is correct. We just need to make it better. And when you hit like parity. So your system is at parity in terms of quota attainment, then it means that it's just going to get better over time from there.
RACHAEL24:24Right. And even if you're hitting parity you're probably using less resources I'm guessing. So less time maybe needing less capacity. Would you say that?
EMIR24:32Yeah.
EMIR24:32like my thought process. There is I don't believe in do more with less. Like I would prefer to do more with more all the time. So for example, if ten of my reps are hidden quota with less time than I can add 20, 25, 30 more and then do even more than that.
RACHAEL24:48I like doing more with more. Yeah,
RACHAEL24:51yeah, everyone's trying to do more with less these days, but it's not always possible. I mean, with AI, I think it is possible to a degree, but if you're able to do more with less, you just add more and then you can do exceptionally more.
RACHAEL25:03So when let's say in the case that you're running this experiment and, you know, quota attainment isn't on, on par or higher than what you would expect it to be, and something is breaking down the experiment. Something's not working with the way that you've implemented AI in your systems. What is the first place you look for what's broken and how do you diagnose this and renew the experiment with fixes?
EMIR25:31So there might be a couple of reasons why it didn't work. One, and again, if it's purely for our pond list might be wrong, context might be wrong. There might be like an ICP mismatch. Essentially there might be a messaging problem. It might be a personal problem like those that pod that you selected are actually lower performing low performers.
EMIR25:52So no matter what you do, maybe they're not going to get to that level. So I would like to think about what's the mismatch between the problems that we identified and what this problem, this there's a solution is not fixing in the original set of problems that we had. And then from there I would diagnose that problem that way.
EMIR26:11So it's essentially like the good thing about option is it's very systematic. So and its essence like if you have the right people and if you have the right message and if you have talented individuals, you will get to an outcome. It's not rocket science. It is difficult, but it's not like like, for example, running an enterprise deal might have 1000 different variables while you lose a deal versus not booking ten meetings a month is a much simpler problem I'm making.
EMIR26:45I mean, some people might get upset about this, but it's essentially like, this is what I believe in. So essentially I will look at those components. Do we have the right people the way of the do I have the right context and right a console? We have the right message, and are we able to supplement enough of those to people on our team?
EMIR27:00And do we have the right team? So if all of these are right, then you should get the results. If one of these are not run right, then you're not going to get to a parody or to the level that you want. Yeah. That's what essentially comes up. Like for example, from our experience, the first 30 to 60 days, it didn't work very well because again, like when you delegate this work to humans, you don't really get to see what's happening behind the scenes in their brains of like, looking into an account and selecting people is actually a very difficult problem.
EMIR27:32And that's like a human judgment problem. And you need to essentially teach the machine how to think like your best steps to choose those people. And it makes a lot of mistakes. So it's more like I approach this more like a broader scale and exclusion filters where versus including titles like when you're building a list, we start a more like broader and exclusions and then got like constant feedback from my team to execute over and over again multiple times a day to get to the truth.
RACHAEL28:04Quick pause. Everything we talk about on this show. Diagnosing go to market ops. Prioritizing projects for revenue. Impact processes. Metrics. Insights. Building a predictable go to market engine. We've built frameworks for all of it. They're free and undated on our website. Union square consulting frameworks. The link will also be in the show notes, so make sure you check that out.
RACHAEL28:27All right. Back to the episode.
RACHAEL28:29At what point in the breakdown of of an experiment for just for example, do you say we need to go back to the drawing board and actually look at the foundations here and, you know, are we even targeting the right ICP for this segment? Are we targeting it with the right messaging? Do we have to go back to how we're divvying up territories or how we're ranking our accounts?
RACHAEL28:54When do you know that it's like a fundamental process problem and not just a problem with the the way that you've trained the AI?
EMIR29:04it is like overall when I run teams or like run revenue or sales teams in general is I like experiment like I like experiments. I think like I like to run dozens of experiments at all times. But it's very important to not change a couple core pieces of your foundation before. Like for example, let's say it depends on your revenue goals, but let's say you're targeting $10 million per quarter and you have six different revenue motions, and you have like attainment or like progress, progress to go for each motion over the last six months.
EMIR29:48And then those are the, let's say, 4 or 5 motions that you have. And there's going to be difference in how successful there. And if you need five of those at averaging like $2 million per quarter, and some of them will be three, four, five and some of them will be like 500 K to a million, just like how mature they are in terms of revenue motions.
EMIR30:08I would never change the top three without adding another three motions that would replace them, and I will never experiment with those top three motions. The reason I'm saying that is because like the on, in order to understand if the foundation is broken or like AI is broken or whatever is comparing those top three motions against an AI version of those.
EMIR30:33And that is what would tell you if something is broken in AI or something else, versus the mistake that I see with a lot of revenue functions is they just like experiment with every single motion at all times, and you never understand if it's like something that you're doing wrong or the market is wrong, or there's like sentiment around your brand that's wrong or buyers don't want to buy anymore.
EMIR30:55Now you don't get to know it because you're experimenting with your foundations.
EMIR30:59You should never experiment with your foundations unless you have a replacement for that foundation.
EMIR31:04So that's how I approach it. Again, like for example, if you take it, take that example of 30 days, instead of experimented with new titles for albums, I would experiment with what are our top three, top three account filters and top three titles and two scripts that we know it will work every single time,
EMIR31:22if you don't know those, then I wouldn't bother with AI because you have a bigger problem.
EMIR31:27Yeah. So yeah, that's how I would approach it.
RACHAEL31:30And where do you draw the line on what you'll have AI handle versus what stays human? And has that line moved since you first drew it.
EMIR31:39Yeah, one of the I would say the single biggest problem in sales right now in building teams is as you do this work more and more, you lose your talent pool from the star team to a team. What I mean by that is like one thing that no sales data that I talk to is thinking about is like we have if you're running like detergent, I said to a motion, and if you're a good sales theater, you probably think about talent to open internally from start to team or start convention team if they want that.
EMIR32:18And emotional connection motion are highly more like dependent on human judgment and decision making, critical thinking and creativity. Like you have to be creative to close a 300 K deal. I'm not saying that it will close a 300 K deal in their first year as an AI, but they need to get to that level eventually. And if they're if they're there for one off year at your company, and if every single thing that they're doing is automated, then they become an AI, then they work is not 100% automated, and it's never going to be automated.
EMIR32:51So they have this like muscle that they lose. So one of the things that I changed my mind on, or like being more intentional about, is
EMIR32:58I'm drawing the line on
EMIR33:01automation, and I'm like being very intentional with my teams on like, these are the things that we need to use our brains and not automate, because otherwise I will have a team of robots, but I don't want that.
EMIR33:13So for example, for this team, we spend a. So for example, we went from like 100% just call calling all the time and repeating the same script over and over again, objection handling to reducing that time and then spending 2,530% of our time every day on like, how can we break into our top 50 accounts ourselves without calling?
EMIR33:35Like for example, we have stars. Like for example, if you have an CI Max who essentially printed, we have a concept called blueprint. In our product, we essentially create the blueprint of your sales process. He printed custom blueprints for his target accounts and thought about like what would like, let's say, rippling. What would rippling sales process look like?
EMIR33:56He printed those out at our office, wrote custom notes, shifted to their offices, and then like, record videos and things like that. This is a simple example of like using our brain. And on the side, for example, multi-threading is not automated. It's not going to be automated. Certain CRM fields will never be automated, like for example, pain.
EMIR34:21P will never be automated. So like I that is something that I've been very intentional about over the last couple of months because like, again, I, I'm running a not just a sales team, I'm also running a talent pipeline. And at that talent pipeline needs to use our brains.
RACHAEL34:43And why specifically pain? I mean, I think I can guess the answer, but I want to hear it in your own words.
EMIR34:49Yeah, because the first time that we essentially like, thought about, okay, we have a lot of, I don't know, a lot of.
EMIR34:56Like metric fields and we should automated then we automated all metric fields and all qualification. Then one of the important things about that qualification, that's not just like pipeline review, is asking who is the economic buyer and what's the pain is it's impacting escalation rates, skill conversion rates, SDR compensation. Because like we have a bar of like SQL criteria that's impacting compensation.
EMIR35:24So if you have enough of the metric fields or qualification, whatever criteria you have, then they get paid on opportunities and we get to we got to a point where like we have a lot of opportunities and you can like if you give any transcript to AI and if you ask them to describe pain, if you have like a better than average AI, you will get some sort of pain.
EMIR35:49Like any buyer, any persona and B2B tech has a pain. Does that pain mean that they will buy? Usually not. Does that pain really urgent? Usually not. Is that pain something that we can solve? Probably. Maybe. I don't know like someone needs to use their pain to understand is this pain. That's something that we can solve. Are they willing to pay for that pain?
EMIR36:12Is that pain something that they want to tolerate for longer? If so, why? We need to be able to understand those. And it got to a point where, like every deal, every call has pain. And then you read it, it's really very well because it's AI. And rather than like making that process perfect, like for example, if we prompt it enough and if you build a good enough system to say, okay, if you're not to solve that, don't fill out this field, it can get better.
EMIR36:38But also it's like like when you create a business case, everything prior to the business case is automated. AI needs to go in, intentionally read pain metrics and everything that is automated, and no one does that. So we get to business case and I'm asking Dave, what's the pain? What's the metrics I focus on? We don't know. It's automated.
EMIR37:03Yeah. That's one of the like stakeholder like multithreading. Same exact thing. Most multithreading just sucks in enterprise sales. It's just like, hey CMO, I am talking to your team fii no response needed. I know I'm not going to respond to an email myself. If someone is trying to me, why would I do that? And it's like, that's something that we need to use our brains because we need to get to power at like 5000 employee company.
EMIR37:31I'm not going to give AI to that. So the line for me is, is this something that requires human judgment? It is scale where it impacts my bottom line and or talent development.
EMIR37:44If the answer is yes to any of those questions, I'm not going to make it.
RACHAEL37:48Yeah, that makes so much sense. Like, I use AI pretty much day to day and everything that I do. And I also try to be very careful with the type of tasks that I give to AI, because you don't want some of these skills to erode. Absolutely. And pain is so, so important to understand. Like you absolutely need to be able to use your own brain to hear what somebody saying to you and pick apart the nuances and read between the lines of what they're saying.
RACHAEL38:14Also to understand maybe what's underneath and not what they're just saying at face value, even. And that's something that maybe AI could probably do. But so important to know on a human level, because you can't just like maybe one day this will happen, but hopefully not anytime soon. But you can't just sit there and have like an AI, you know, feeding you everything to say on a sales call while you're talking to another human being.
RACHAEL38:42At that point, it's just you're just you are the robot. So you need to be able to have these human connections and conversations and be able to think with your brain.
EMIR38:52Yeah, exactly.
RACHAEL38:53Yeah. So you also expanded the AI role so that they own accounts end to end. So not just when the sale closes but afterwards as well. So they stay at the camp forever. That's a structural decision that I think most crows would debate. And if you might actually make, what problem were you trying to solve there? And why did you believe the traditional hand off model wasn't the way to go?
EMIR39:21Especially in our like strategic strategic accounts.
EMIR39:24Does a recent change that we made. So like we have two years minimum of accounts like hundreds of accounts that we didn't run this motion on. So essentially my hypothesis is it's a little bit early to see results because it's been like a couple of months now. But essentially like when you have a very strategic account or like anything that's on an SMB account, SMB accounts, if you're selling semi doesn't matter.
EMIR39:48Assembly expansion should ideally be product lead, not like an elite like mid-market enterprise. And any all of those accounts, anything from 3 to 12 month sales cycles. The AI has developed a champion or a set of champions, and no matter what we do, the handoff document is not going to be perfect and the account manager is not going to be able to get that document and then turn it into like an expansion business or an expansion mindset.
EMIR40:21And the problems that we ran into were like account managers or customer success managers are comped on retention and expansion. So on the retention side, you can also make the metric like, I don't know, net revenue retention, for example, that would be tied to expansion to a certain degree. But it's very difficult to be able to balance the retention motion and expansion motion with the one function or one team.
EMIR40:53Let's say a lot of customer success managers, because again, like people can take shortcuts. So for example, if you say Net drive a new tension, one account can just renew at 15% because they love the product.
EMIR41:05That would mean that you don't have to sell in other company 15% more SKUs or seats or products, whatever, just because that company loves the account and the product and they just accept the renew at 15%, that's one second is again, that context.
EMIR41:20So we have again, hundreds of accounts where we built the strongest champions that the AI has never talked to again because they're just focus on new business. So then like it essentially means that we start a net new relationship building exercise with an account manager who is different than an AI with the same set of champions, and the champions need to trust as account manager again and learn about them, and they learn about this person or people.
EMIR41:49And we just go through this entire exercise again when we already did it once. Together with this, I also made sales engineers full cycle, to which I believe very. I'm passionate about this. So before we had sales engineers doing scoping and then once it's closed or like scoping and qualification and selling like technical selling essentially, then we handed off to the implementation managers and implementation managers exactly the same thing.
EMIR42:19There's a lot of context and trust that's built with the sales engineer, who was perfectly capable of implementing the account as well as we passed off the implementation managers. So right now the the motion that we are running is sales engineers are full cycle with this help of implementation measures. So they're like owning their success. And turned. And implementation managers are helping the sales engineer implement it with their context and knowledge.
EMIR42:44And everything about the account is very similar. It is our A's job is done when we successfully expand their count. When we expand the account, we can expand them multiple times. Their job ends. If unfortunately it turns, which is can be a combination of different reasons. But the A is essentially the CEO of the account until they can't turn.
EMIR43:12We have a sales engineer on the technical side. We have an account manager who is keeping the relationship and like doing adoption and like team usage and seats, is the owner of the dollar value of the account in terms of new skills and adoption, expansion. All of those account managers more like
EMIR43:30I need to get adoption from this account and to sell more seats and more like adoption focus and gross revenue retention, essentially, logos and then sales engineer has resources like we have similar to forward deployed engineers, we have engagement managers, and we have implementation managers, and sales engineers work with those people to be able to deploy and get to success.
RACHAEL43:53actually like that model because it takes the pressure off of, you know, a customer success manager to, you know, be a salesperson also and try to expand and renew these accounts and learn about product white space, stakeholder white space and everything when they have so many other things on their plate. As a CSM, which is a team that's historically under-resourced as it is, it's really cool.
RACHAEL44:22So if a zero is listening to this and they're they're thinking about doing something similar or making similar moves, what's one thing that you'd say that they need to get right before they implement this on their own, with their own team?
EMIR44:36Yeah, I would start from like not like change everything all at once. Again, keep the foundations right. I think I would pick like a certain segment for us. It was strategic and it came at the same time as like a new product launch, which is a great opportunity to do new things. So we launch our new product, new product module, which is revenue agents, which includes prospect that deal agents.
EMIR44:59As we start running POCs and pilots and closing customers, it became very clear that the old model is not going to work. And then we changed it for the new motion. And now we are step by step. We are bringing that set account model to like enterprise and mid-market at different levels and different strategies. So I would either like if you're a new product, that would be the perfect time.
EMIR45:22If you're not launching a new product or a new motion, I would pick I would start with like Shadow Enterprise and then go from there and SMB, you don't need it. So it's usually a permit market and enterprise and that.
RACHAEL45:34don't have a ton of time left just coming up on the hour here. So I just wanted to get a couple more questions out. So the first one is result metrics. So you talked before about being really focused on result metrics. What are the specific metrics that you hold your revenue team to and which ones would you say are, you know, metrics that a lot of crows usually put on a pedestal or think very highly of, but you don't really look at that much?
EMIR46:04We have TVs all around our office. We work in person, so all TVs have a custom dashboard that I built with cloud code. So every single day it's right next to my desk. So whenever something happens, it's like notifying everybody. So I look at meetings booked daily against daily goal meetings booked weekly against weekly goal S2 pipeline. So stage two plus pipeline is qualified.
EMIR46:33Pipeline daily, weekly, monthly change. I look at fires. Fires mean for me. Fires mean it counts with more than a certain ACV that has not moved to a different stage. And like more than seven business days or for ten business days depends on segment. So those are like daily things, daily metrics that I look at. I don't I'm not going to say like I will think about what's the metric that everyone looks at that I'll look at.
EMIR47:01But I think one of the things that everyone looks at, I don't look at that often is quarterly forecasts, because I think you can only control, in my opinion, the next two weeks of your business and you can maximum you can control this month. If I'm starting a net new quarter, in my opinion, there is no real reason or like there's no reality.
EMIR47:29And like what's going to close this quarter three months from now. So usually I ignore the quarterly forecast. I'm just focused on the next two weeks. What can we move right now? And then of course we work on like longer cycle deals and everything, and we try to close them on the closed dates that we put in as goals.
EMIR47:52But like the quarterly forecast metric, is like a leading indicator of success for a business. And my opinion is misleading because forecasts in general mostly guesswork. So yeah, I don't look at quarterly forecasts that often. I look at the next two weeks maximum this month. And what are the leading indicators of success for our quality forecasts?
RACHAEL48:18Okay. And just the last question here. If a crow is listening to this today and they want to, you know, increase revenue that they're seeing from their outbound motion, increase pipeline generation, increase predictability in their forecasting, what's the first thing that you tell them to to focus on?
EMIR48:38Yeah, most companies were mostly arrows. Don't know how they close deals. So this is something that we hear all the time. Like you talk to us or like you talk to us. And they're like paying consulting company millions of dollars to create their sales process. I just like this reality that's happening in the market right now. I think that is the right instinct.
EMIR48:59But it's the wrong move because these companies take three quarters, usually like nine months. And when they build you a sales process that you're running already, it's nine months late. So the market changed. Your people like how you sell changed in the last nine months. So that going to be accurate. What I have done for myself, which I would encourage all sales to do with their team, is you take the top 20, let's say 20 deals at the highest AC resource time, like shortest time to sale.
EMIR49:24And then you create a blueprint of those deals. So you study them for like hours and hours, however long it's going to take. I did this myself. It takes a long time. But you create essentially like how do we close deals and to maximum efficiency and what do we do? And when you do that, when you watch those calls, when you listen to recording, when you see how they did the multi setting, they come up with like an actual real process that doesn't live in your Salesforce.
EMIR49:48Between those five stages that you decided five years ago, you will see all the details in between. And then I would encourage everyone to turn it into an enablement and like run it every single day. That's how we build our product as well, what I call the blueprint.
RACHAEL50:36Awesome. Amazing answer. Thank you so much. And where can people find you? If they'd like to follow what you're doing with Hockey Stack or check out Hockey Stack in general.
RACHAEL50:47All right. And we'll have a link to LinkedIn in the show notes as well. Thank you so much, Amir.
EDDIE50:55Thanks for listening to the episode. If this resonated. Please give us a five star rating and a follow. It helps us reach more people, and you get our latest and greatest content without having to search for it. And if you're looking for hands on help and go to market strategy and or rev ops, please reach out to us.
EDDIE51:09We help our clients with everything from annual planning to improving processes and go to market, implementing systems to support those processes and go to market AI. We're always happy to offer a free consultation to help you identify the best opportunities to improve your go to market engine, with or without our help. You can find us at Union Square consulting.com and the info will be in our show notes.
