Enrichment Module teaser with Yash Tekriwal (Clay)

Let's see, here we go. This is a, this is the first of our module previews for the upcoming GTM Engineer School. And the school itself is going to kick off with a module around enrichment and all the other wonderful things that millions of people are doing right now with the new tools out there, new capabilities. And Yash is going to be some spearheading that as he, as he did with the last cohort and everybody really enjoyed it. So I'm going to do a couple of minutes here just telling you what GTM Engineer School is and then going to kick it off to Yash who's going to lay down some knowledge and then we're going to do some Q and A from attendees. Okay. So yeah, we do have a bunch of other webinars coming up. These are little teasers and there's a few things that you're going to learn each one. So even if you're not able to make go to Market Engineer School, no worries, attend these webinars and soak up some knowledge and then yeah, if you do want to attend the school, the registration ends September 22nd and the school starts the 29th. There are five core modules in the school. Here you can see from research, enrichment and scoring to messaging, workflow, automation, content, LLM optimization in the age of AI. And so we've got a bunch of great instructors and we've got a bunch of wonderful TAs that are gonna be helping in the breakouts. The weekly schedule looks something like this. On Wednesday there's a two hour lecture that kicks off each module and that is instructor led. It is from 8 to 10am Pacific Time. And if you can't make it, we will be recording that, sending that out to our students. Then the rest of the week there's going to be ASYNC projects for you to build and live debugging sessions with our TAs. The TA debugging sessions are going to be an hour long. They're going to be spread throughout Thursday, Friday, Monday, Tuesday. You can use the weekend to catch up and go deeper on some of the things that you've been learning. And we're going to tag the, the sessions is either sort of beginner track or intermediate track. So you can kind of pick which one is most appropriate for you, where you're feeling on that, on that module. And let's see a couple of things that are unique here. Think about go to Market Engineer School. It's cross tool expertise, kind of a, you know, a taster of the various things that you're going to be encountering. Even if you don't use these tools specifically, you'll get a sense of of how things are laid out and, and the general patterns that are going to be useful for you moving forward. It's all hands on learning. It's a little bit of theory, but a lot of just doing. If some of the best experts in the world in here training, you get free credits and pre built workflows and curated guides and then there's a slack community of all the students where people are sharing concrete use cases. We've had a lot of graduates come out of the first cohort saying a lot of nice things and doing incredible things at work and getting new jobs and all this kind of thing. We got a whole bunch of TAs either they're experts because they've been there from the beginning. Like Yash here is what they call them, the OG GTM engineer. And then we've got a lot of other TAs that are running clagencies and outband agencies and other doing other interesting things. And yeah, you get a lot for your money, we think in this course. If you're interested and you have some questions, let me know. I don't want to take up any more time talking about this. I want to kick it right off to Yash who, who I'm going to hand the screen over to you, Yash. And so I'm going to stop sharing and let's see, you want to take over there? Perfect.

That working? Yeah, yeah.

Can see your whole screen. You can go into slideshow mode if you want, but yeah, we can see your screen.

Amazing. One sec, let me fix one thing. My Internet is about to crash on. Do do do do. I might break out for a second. All right, everyone, see me. Sweet. So in the spirit of not touching on a lot of theory and getting straight into just learning what cool things exist in the world, I don't want to go through a lot of slides today, but I do have a quick question which feel free to go ahead and drop it in the zoom chat and I'm just going to scan it. How many of you feel like you have a say? Yes. If you feel like you have a decent idea of what clay does and how it's supposed to work at a theoretical level and then say no. If you've like heard of clay but don't really know anything about clay. So many yeses. That's amazing. The marketing team is doing great work and that means that we don't have to go through a bunch of theory stuff right I've spent a lot of time doing these webinars and there's always the like people have heard of clay but they don't really know what clay does. And so I spend some time trying to explain it at a high level because I do think the high level understandings are really critical to being able to work in the tool effectively. I do see a couple of middle slash, maybe no responses. And I also see actually that's what we're going to get at what Jeffrey just said in the last message. Right. Has heard but finds clay not very easy to use to start. And I will just go ahead and affirm that that is true. And then the one thing I'll say before we jump into it is I think clay. And actually I would probably go so far as to posit many of the other more flexible GTM tools you guys are going to learn in different webinars and in the program are different from standard SaaS in the sense that it's not 100% purpose driven or built. So for example, if you've used Superhuman for email, Superhuman is designed specifically for you to get through your email faster. And that means that they've built an entire onboarding flow and UI and all of the controls to very quickly help you understand how to get through your email faster. Because it's laser focused on one mute case. A lot of GTM tooling for decades has been similar. You go to Outreach IO to do your Outbound sequencing or Salesloft, you go to HubSpot or Salesforce to do your CRM. And even within HubSpot and Salesforce, as many of you likely know, you can go really, really deep down the rabbit hole, but at least it's zoned in on the CRM use case. You go to ZoomInfo or to Apollo for lead list building. And so I think because software has been purpose built for different use cases, it's been at least a little bit easier to stomach or digest what you're doing going into it. Tools like clay, and when I say like clay, I also mean N8, air ops, cargo, etc. Are far more flexible in what you can do. So you can come into clay today and you can do data enrichment, you can do outbound, you can do inbound, you can do all sorts of things. And that part of that is what makes it so intimidating to get started. And also we're not doing outbound or inbound. And just like a purpose built, like end user does the work type of way, we're building systems to do all of this work for us. And so it requires this additional layer of computational thinking and understanding how the language of programs and computers works, even if you're not actually writing code yourself. So that's my quick teaser of I actually highly recommend anyone who's interested in becoming a GTM engineer at some point, take, if you haven't already taken like a CS101 course on code Academy, and just understand what if statements are loops, variables, et cetera. Because that foundational understanding makes learning a lot of these tools far, far easier. And then if there's other questions that you guys have, I'm going to try and go through this quickly and then we'll leave time at the end for Q and A. So without further ado, for all of you who have some level of familiarity with clay, this shouldn't look too, too radically different. What we're going to go through today is not everything that you can possibly do in clay. I would need several hours for that and still probably not do a good job of it. What we're just going to cover is the basics of what I think 99% people get value out of when they're starting in clay, which is enrichment, custom data and clean data. So I'm going to start with a new workbook in this table. And I think just for everyone's knowledge and understanding, there are tons of sources you can use in clay. I actually think one of the least appreciated parts about the product often is just how many sources there are to use. Here, where you start your build from has a huge impact on the rest of your workflow because it's limiting or expanding the total number of data sources. You could use 80 to 90% of people just use find people and find companies, and you can get away with that still being incredibly valuable. That's essentially all of like Apollo or ZoomInfo or a similar prospecting product built into one feature in clay. So you can do that super, super effectively. But then also I think if you want to expand your GTM engineering capabilities, again, this extends outside of clay. I would also encourage you to think about where can you get data that other people don't easily have access to? How can you get data that people don't easily have access to? And what data is really well suited for your ICP and needs? So a couple of examples here that I won't go through live, but just to think about it, how do you think about turning static lists into dynamic signals? So now instead of you pulling a list of people and then trying to find which ones change jobs in the last one to three months. Can you set up an always on table that looks at people within a specific icp? Maybe you're hiring for engineers or you want to reach out to engineers who move from an existing customer to a new customer. Every time someone changes that then you want to get that data in a table and then you either message them with a follow up about congratulating them or about probing for a conversation. You could do all sorts of signals across websites. New hires, promotions, news and fundraising possibilities are truly limitless. The other probably most popular one, because I know that this menu can be overwhelming to people when they first scroll down it's is small businesses and businesses that are not on all of the B2B SaaS platforms. So if you are going after companies that don't have a huge LinkedIn presence honestly tends to be the corollary. For example lawyers, accountants, dentists, warehouse workers, manufacturers, all of these companies are much better found through alternative data sources. Think Google Maps was the initial way that a lot of people started building these data sets. We recently launched openmart. People have also actually been using reviews of companies to go ahead and prospect for new customers. And so there are tons and tons of ways to actually pull data at first into your clay table. And I highly encourage you all to explore these if you're going to start dipping your toes into the clay waters. For now, just to get through the bulk of what we want to in this module, let's say we go through a table quickly and do have our initial list. So I'm just going to put in software development companies. I'm going to make them slightly larger companies. For those of you that have familiarity with prospecting tools, this UI should not be altogether too crazy or too different. You can do a ton of other things in here and I can filter them. Let's do this last one to United States. And then if I wanted to, just as a note, I could exclude companies. If I have a DNS list, I could download it from my CRM into play automatically and then use that table here. I can also go ahead and look for lookalike companies in this modal. I can also start to look for products and services that companies might offer. These are all newer features that we've been adding that are AI, I would say. And then someone's getting some messages. Nice. I will limit this just for speed's sake to 100. Even though I have a thousand potential results I could get in here. I'm just trying to show you guys how this stuff works. So I'm going to import this to a new table in the workbook that I'm in right now. It's going to take a second to do that and then you'll notice this Actions Modal automatically pops up. This is also something else that I think people get lost in in clay is a lot of people feel the pull to just keep adding columns, which is one way you can do stuff here. But almost always add a column, add enrichment, or use the Actions modal, same sort of thing. Now, once you have your initial data in clay, there's a lot of creativity that we are not going deep on in terms of how to source data in clay. But I showed you guys the menu of options. The next biggest, most important part, probably the thing that people still spend the most time doing in clay, is getting accurate filled data for all the companies or all the people that you might have in a table. So as an example of this, I'm just going to run one enrichment on this company data set. Let's say I have a lot of large companies and they all might be in software and that can be interesting. But actually I really care about companies that have a high concentration of engineers. And so that should be almost all of these because they're in software. But if I really want to check that, I can look up, I think the action is employee count by criteria and then I can select this enrichment. Notice that it already tells me it's a default waterfall. Most things in Clay tend to be waterfalls because we are aggregating 150 plus different data providers to give you the most accurate data that we can. And then I can even see what the waterfall providers are here. This icon stands for like our native CPGA data set. This icon, I believe is People Data Labs. So now I can configure what domain, keywords, etc. I need in order to filter by employee count. I am going to do the keywords of engineer here or let's just do software engineer because I want to look at the headcount of people who are only software engineers at this company. I can also keep this down to just people in the US Since I filtered that to companies here. And then we actually have a role filter for People Data Labs as well. So let's put that in engineering. And then if I really wanted to get more complex here, I can come over to full configuration. Note that this exists for all waterfalls, but it's helpful more as a debugging and troubleshooting step than it is a running step. And then the Last thing that I do here that is very, very important, I always save and run on 10 rows whenever I am building a new column in a clay table because a lot of people don't want to waste credits by accident. I make sure that I'm getting the types of results I want on a sample set of 10 before I go ahead and roll it out to the rest of the table. And it looks like we're getting pretty good results here. People are. I'm getting like, yes, this makes sense, right? These companies should have a high concentration of engineers. And then if I furthermore wanted to. Right, let's go ahead and run this across the rest of the table just to see where it works. You still see this waterfall semi in action here. We're actually finding good hits on all of the first data providers, so we're not needing to run the second. But if we didn't get it, we would then run the second and then run the third. And I'll show you a more powerful waterfall in just a second. What's now cool is that I'm able to start getting more and more data about these companies so I can be even more targeted in the way that I might want to go after them. And let's just arbitrarily say, for example, I only want to actually talk to people at companies that have more than 400 engineers. I could also do that coming in and adding a formula. This is using AI, so I could just say return. Let's use this for flow. Return. True if employee count is greater than 500 and false if it is less than 500. I'm just going to turn this into a variable, generate the formula. And so what's cool about this, just to walk you guys through, what's happening here is. And false if employee count is less than 500. Seems like the formula is a little bit off, so I want to Change that, but AI will go ahead and create some. It's technically JavaScript, it's technically Play script. Like you can't access full, full JavaScript in here, but it will go ahead and return that number to us. And I'm getting a lot of falses here, but let's see what's happening. Oh, it's because it's taking employee count, not the final one at the end. This is actually another good clay debugging thing. So I was just doing employee count from these total employee like companies overall criteria. What I actually want is the output column of my waterfall and so that employee count would have a range. Right. That's why I was returning this issue for me and this is a good remark on in general for all workflow automation tools, hygiene, what you name your columns, what you name your variables, how you organize your workflows very drastically improves readability and debugging possibilities after the fact. That looks a lot more like what I was trying to get. So I'm going to go ahead and say output is correct, save formula. Now, only some of these companies will pop up and some won't. And then I'm going to go ahead and add a filter here to only find companies where this box is checked. Now I've taken my list from 100 down to 23, which you can see in the top right here. And just to show you guys how this connects to people data as well, I'm going to go ahead and come to this actions panel, hit sources, find people at these companies. And just to keep things moving quickly, I'm going to do this fast. I'm going to put in engineering just as a job function. I would probably get more discreet with all the different variations of job titles that I actually want to go after if I were doing this more fully. But I'm going to limit this to 100 as well. And then I am going to go ahead and click continue, import this to a new table, give it a second to load, and then once it does that, I will get my list of people in this new table and again have the same problem that I had with the company data, which is that this is a great starting point, but if I really want to action this data, I'm going to need a couple of core things and then a couple of additional nice to haves. I need at the very least these emails and or mobile phone numbers, depending on the mode that I'm going to contact them with. So let's take email because I think it's still the most popular and most used by far. I'm going to take work email and I'm going to note that I could come in here and configure even further. I could add more email providers to this table if I had other data points if I wanted to increase accuracy. I almost always tell people to just max out your email waterfall because we refund you if we don't find emails. And so this one I know is going to work. I'm just going to save and run 100 rows just to show you how it works across the whole table. And the other cool thing about the email waterfall in particular that you can see right here, we have validation built in so even if a data provider returns you what it says is a valid email, we are going to double check to make sure that that inbox exists before actually returning you email addresses that might hurt your domain deliverability scores. So this is actually a beautiful example of the power and magic of waterfalls. We're finding a bunch of emails from the first provider and we're also missing a bunch of emails from the first provider. So where that first provider doesn't work, row two here as an example, we refund the credit and then we go ahead and then run the second provider. Once that second provider works and validates the email, we don't run the rest of the data providers. We save you on the cost because you've gotten the data that you need. And then we automatically combine all the emails we find into this last column here and you can start to just visually get a sense for how three to four data providers tends to increase your coverage by orders of magnitude. Like, I would probably have less than half the results here if I was just using one of these. And then I have definitely double from the first two. And I'm getting sort of this like diminishing marginal returns, but still full cohesive. I'm at least getting results back from the rest of these providers. I've often found if I had gone in and added a couple of other email validation providers and finders in here, I would get close to, if not at 100%. And so we're getting still pretty close to 100% from this list, which is impressive. And so that's a quick debriefer on waterfalls, which are the core, core enrichment step in play. Part of why I wanted to move so fast through that is to get to the last thing I'm going to show you guys from an enrichment standpoint today, which is the thing that everyone cares about, which is AI, right? Everyone wants to use AI because AI is the future and it's working really well. My one caveat that I will tell you guys is AI is really good at certain things and it's also not so great at other things. Case in point, don't use AI to try and get emails unless you happen to know for a fact that the email is easily discoverable, it's findable on a website, and you can give it a reasonably structured path to get there. Otherwise, what is going to happen is AI will hallucinate email structures for you, you might not remember to validate them, and then you will send a bunch of emails to a list of emails that don't exist and your domain will go down the pipe and you will have to work very hard to restruggle or buy new sending domains which will make your life easier, but still a mess in the short term. So.