Transcript of Audio File 1f86b7f8-126a-544e-9959-7892c9290fd2

Hi. So, we are going to dive into the weeds of Clay, and you're going to get your hands dirty really quickly. But before we do that, I want to try and give some perspective on Clay, um, such as the most common use cases. With the hands-on exercises, you are going to cover a lot of the hands-on the the most common use cases. Um, I also want to talk about the different phases that, um, that I view there are in Clay. And Clay's team talks about this as well. I I frame it just slightly differently. Um, and then I also want to talk about the top 20 most common enrichments. You want to know these ones exist, uh, because if you know what they are, you're you're covering about 94% of, um, the enrichments that there are by usage. So, it'll get you a long way. Um, we've had, uh, Josh and Bruno pull these from from the growth team at Clay, so, um, it's really good data. So, let's start with the most common use cases. Um, there are several, but some of the most common are outbound. This is probably the most common, um, very common with, um, agency owners. And outbound is you take a list of people or companies and enrich it and then send it to usually to Smartlead to Instantly, maybe a LinkedIn tool like HeyReach or, um, or the Growth Machine. And this is either done through creating static tables and doing one-off blasts, or doing a streaming setup where, for example, you scrape, um, job boards continually. The jobs flow into Clay and then it's automatically enriching them, finding, um, finding decision-makers and finding their email, writing a message and then sending it to another platform, uh, like the ones I mentioned, and just it's an evergreen campaign that just keeps going. Um, inbound is more common us uh with larger companies. They already have lead flow on their website and they want to respond to them quickly. Speed to lead is a real thing. Statistically, if you respond to leads within, uh, 5 minutes, you're much, much many times more likely to actually convert them. Um, so handle handling inbound well is important. It's it uses a lot of the same skills as outbound does. It's actually in my point, by my point of view, it's it's slower-hanging fruit, you because these are people who are already interested in you, and you just have to respond to them well. Um, okay. CRM enrichment and cleaning. Often people will have a CRM that's not been taking care of, or hasn't been using Clay. And when they start using Clay, first they want to say, "Let's take all my old data, my leads, my, um, contacts, my accounts, etc., and just make sure they're enriched with all this data that we want from Clay." So you might export the data, enrich it in Clay, and put it back in Clay. And then when new leads, contacts, accounts, etc., are added to the CRM, you want to continually make sure that those new contacts are, um, updated with with good data, and then there might be a a recurring refresh of data as well. So that's a a use case for more sophisticated companies. Pre-call preparation, making sure that if you're an account executive, um, or a GTM engineer, you want to have good information about the person before you're on a call with them. Um, and it takes a lot of work to do research, especially if you're on back-to-back calls. So Clay can help do pre-call preparation. Um, we have an exercise that you may get to do where you'll build build something that does pre-call preparation. As well as we might have we have exercises for several for most of these categories. So, lead scoring is another interesting use case where you have a list, whether that's a list that's continually streaming into Clay or just a static list that you upload, and you want to prioritize this list of leads and say, "I want to reach out to these ones first, these ones I'll maybe reach out to later, and these ones are junk. I don't don't want to touch them." Um, there are lots of there are several ways we can lead score. Um, yeah. So, that's that's that covers most of the common use cases. I'm sure in your roles, you'll see several more. But, um, as I said, we're going to try and cover most of these with exercises during the bootcamp. Uh, okay. So, another perspective on Clay is no matter your use case, you're you're probably going to use one of these steps. So, you start out with a list, um, then you enrich the data on your list, and then you wrangle it. Um, that's maybe not the best word, it's the word I'm using, which is you take, um, data that you've pulled in or data that was already on the list, and you manipulate it into another form that's usable. For example, in in an email. And then you send it. You might be sending this to an an external platform. So, let's talk over these stages real quick. Um, first, um, this is the most extensive discussion. Um. Okay, I'm trying to. There we go. Okay. So, ways to get lists into Clay, a very common one actually, is importing a CSV of data. Um, probably the most common way of doing this is through Apollo. Apollo is, um, people trust Apollo lists more than Clay lists right now. I would say in general, I'm sure that'll probably change, um, but Apollo has huge amount of data coverage right now, um, they have very granular ways to filter the data, and it can be pretty cheap. A lot of people work around Apollo plans using, uh, scraping Apollo, which is against their terms of service, but people do it in general, anyways. Um. So, yes, that's very common. Clay has a people search. It's free, um, because of Clay's, um, I guess I assume, uh, problems with the data provider that scrapes LinkedIn. Uh, the data completeness has been off. I know there's a lot of work I'm sure behind the scenes to to fix that. But this is another viable option. What's nice is that it's absolutely free. It's it's very convenient to use. Um, similarly, we have a company search. Um, I believe that completeness, data completeness issues aren't as as bad with the the company search, if if at all. Um, LinkedIn is not as protective of their company-level data. Um, then we have job search data. Or sorry, job data. Um. So, you can put in search for certain job titles, um, and it will return all the jobs with those job titles. There are some alternatives to this. So, right now, as I understand, Clay does not have a way where you can set up a job search, and new jobs fitting that criteria will stream in. Um, it's not possible as far as I know, and so if you want to stream jobs in, it makes sense to use something like Apify or Browse.ai, they have ways of doing this, I've done this, um, with quite sophisticated setups. Um, you can also Clay has a way to scrape Google Maps. Just one second. Clay has a way to scrape Google Maps. Um, I believe it can scrape up to 1,000 results, which honestly, I usually recommend people use Outscraper because it's not limited to 1,000 results, you can scrape thousands or tens of thousands of results. It's pay-per-result. It's relatively cheap. Um, it's they're just quite specialized in this, and so it's not just trying to work with an existing API, it's doing a bunch of different API calls, and then aggregating the data together. So, if you have a big data set, I recommend people go outside of Clay for this. Um, but I'm sure Clay is is great for 1,000 results or less. Um, Ocean.io is a way where you start with a list of companies, and then, um, you can find a list of lookalike companies. And I've heard really good things about Ocean, um, and it's really cool that Clay just added an integration so that people don't have to buy a separate Ocean subscription because Ocean is very expensive, at least word on the street is is it is. Um, Storeleads is a really good way to find, um, e-commerce stores like Shopify stores. If you want to do a campaign to e-commerce, this is a a really good provider. Also, recently added, I think. Um, PhantomBuster is different than most of these other, um, lists in that or list creators in that it will use your LinkedIn cookie or it's mostly used for LinkedIn scraping, um, as I understand. It, you take your LinkedIn cookie, it'll take your LinkedIn cookie and can scrape a post of all the people who liked or responded to the post. It can, um, if you're an admin of your LinkedIn, um, company LinkedIn, you can scrape all your followers, um, stuff like that. So, it's it's powerful for LinkedIn stuff. You do have to be careful that, um, you don't overscrape LinkedIn or LinkedIn will, um, figure out that you're scraping and shut down your your profile. Temporarily, it's, uh, usually not for more than a day or two, as far as I gather. I've I've scraped LinkedIn, and I don't think I've ever had my profile shut down, but I know people who have it. Oh, yeah, every week it gets shut down, and I'm fine. Um, job change tracking is really cool. Um, I hope we do an exercise with job change tracking job change tracking as well, because it's a new feature in Clay. But basically, you give it a list of people, and when the people change jobs, they stream into another table, and then you can automatically take action on them, whether that's sending them an email, adding them on on LinkedIn, etc. I I mean, if you want to if you want to send them an email, you send, uh, their data to Smartlead or Instantly, and if you want to send them a LinkedIn request, um, you you send them to the Growth Machine or HeyReach, etc. Um, but anyways, yeah, that's what you can do with with that data. Um, you can import data from your CRM. I know Clay has been working on their CRM integrations with Hubspot and Salesforce. Um, and at least some of these you can import a list, and as you update the list in your CRM, that data will, um, flow into Clay or maybe flow out of Clay as well. Um, so, it's a a can be a streaming setup. Apify I mentioned before, basically Apify is the Zapier of um, of streaming data. So, they have a lot of different, you can say they have a lot of different apps that are dedicated to scraping different things. So, you have several for scraping LinkedIn, for Facebook, um, Instagram, etc. You, if you can't find a way to do something in Clay, check Apify and see if they have a scraper for it, so you don't have to, um, you don't have to be, sorry, I had to zoom out. So, if, um, if you can't do something in Clay, check if Apify Store has an actor that will do that and there's a good chance that uh they will. And it integrates well with Clay, and then Webhook is if you absolutely can't do anything with any of these other these sources, there are a few there are a few more sources in Clay that are good, and so I would check, um, Clay's list building sources first there. And then, if you're like, there's no way I can I get this data into Clay from any of these other sources, Clay has webhooks, and a webhook is basically something that catches data from any source. You send it send it data programmatically and then that data will be placed into a Clay table. And so I could write code some basic code on my computer, right now, in about 2 minutes, send it to send it to Clay, a Clay table via Webhook. Um, so you can write a write any computer program or basic script and it'll send data via Webhook. We're going to do this, I know it sounds complicated, it's not as complicated as it seems, um, but, um, I want you to start knowing that that's possible. That's all. Okay. So, we have covered the the most common platforms that I want you to know about for list building. Uh, these other three stages are going to be quite short compared to what we just went over now. So, once you have that list, you can enrich the data. You can have Claygent, which is, um, very powerful. It's AI plus Google searches plus, um, a scraping, and you can basically, there's an AI that's coordinating other, um, ChatGPTs, Google searches, and, um, and a scraping, but it that this um master AI will ah will say, "Hey, I need to know this thing." Okay, ChatGPT is the right thing to do to to to to try first, and then, "Oh, that doesn't work, okay, we're going to try Googling." "Oh, we got some results from Google, now let's use ChatGPT to ah to manipulate them," etc., and it'll just keep going until either it it it costs too much, um, or um it finds what it needs, or it just concludes that it's not going to be able to find what it needs. So, Claygent is very powerful and iterative. Um, and as you'll we'll see later, this is one of the very most common enrichments that's that's used. Scraping is there are a couple of different ways to scrape, technically Claygent scrapes as well, but, um, Clay has a built-in scrape website, um, extension. Zenrows is another one that's worth checking into. It's especially cheap with um and if you have your own Zenrows API key, very very cheap. Um, and what's nice about Zenrows is for some websites, it can return data back to you after it scrapes. Not as HTML which normally, uh, that that's the native, uh, code that websites are built-in usually, uh, it can return it as JSON which is much more structured than HTML and you can pick you can pick data out of, um, out of the JSON much more easily than out of HTML. And if you don't know what I'm talking about when I say JSON, um, that's okay. Uh, you're going to learn about it. Don't be intimidated. I'm just giving a little bit of detail for those who who do know what I'm talking about. Okay. Um, Enrich, we have an email waterfall where you you say, "Hey, I have this person's company, their name, maybe, you know, some other data, and um, it'll check a bunch of email providers until it gets the data that, yeah, it'll just keep checking email providers until it's ah, here's an email that is valid. HTTP API is um, earlier, I said a webhook can catch data. HTTP API is a little different. It throws data to another another source. So if if Clay does not have a native integration with with an enrichment, Clay can still throw data to another service, any any service that has an API and if the API is working or is is is friendly, it'll throw the data back and say, "Okay, here's your result. You're able to ah, here's what you need." So HTTP API is is very powerful, um, also a very common way to use Clay, um, and it's a little bit more difficult but we're going to cover this, we'll we'll get into it as well. Um, web search, you can do automatic web searches with Clay, Google Google searches, you can, um, pull data from your CRM which is more relevant for larger companies that are organized enough to have a CRM, not all not all companies are. Um, and then we have also like things like finding people. Uh, so if you have a company name and you have some job titles you want to find, you can find the people with those job titles, for example. Um, I don't know why this one's here. Okay. So, that's Enrich in a nutshell. Then we have Wrangle. Very common, you take ChatGPT, you can here's some data, rearrange it, think about it, spit out a result. This is the most common uh enrichment in terms of uh number of times used. Another powerful one is Formulas. So, ChatGPT costs money, whereas formulas are free. Uh, you can take existing structured data and, um, and modify it. So, formulas are free, but they're not as flexible and powerful as ChatGPT, but sometimes they'll actually do um the job. In some some cases, they'll do the job better than ChatGPT can. Um, normalized company, you take a company like Acme, uh, Inc. Limited, and just shorten it to Acme, so that when you're sending, "Hey, I noticed you work at Acme Limited, Inc., whatever," that's that very clearly looks like your mail-merged data in and people say, "Ah, I don't want to read this. They're just uh spamming me." So, normalize company is um allows you to look more natural when you write emails. Same thing with uh normalizing name. Uh if if my name on LinkedIn were uh Dr. Nathan Lippy, I'm not a doctor but if I were, um if you didn't normalize my name, it would um it might say, "Hi Dr. Nathan" or "Hi Dr." And that would be immediate giveaway that ah that was, uh, not a real person did not write this. Another thing is is scoring, like lead scoring. I think I mentioned this earlier, where you have a list of data and you can say, "Given these factors, um we're going to calculate a score of how good of a lead this is or whatever." Um that's a way to wrangle data and we'll we're going to do some exercises around that. And then another thing is count. Like let's say that you pull a bunch of jobs for a company, you can count and see how many jobs were pulled. Just some basic examples. There are a lot more ways you can things you can do, especially with formula and ChatGPT, um pretty much everything you can imagine you can do with these enrichments as far as Wrangling. And finally, sending, uh so once we've done all that, wrangling, email writing, um enriching, list building, etc., we can take them and send them to Smartlead and Instantly, or Smartlead or Instantly. Um as far as I gather, most of the community leans a bit towards Smartlead as as do I. Um, it's built more for technical people. Um, in my opinion, both of these platforms have large usability issues, but they're the best that exist as far as I'm aware of, um for something called multi-domain sending, where you're in order to get around spam filters, you send not thousands of emails a day from one inbox or or from one domain, but you spin up a bunch of different, um, domains and have a few inboxes on each domain and send a few emails from each inbox every day. And, um, it looks like you have an army of people sending emails, but in reality, can just be one person or no one sending those sending those large volumes of email distributed out over several domains. So, these are the best as far as I'm aware, best tools currently for doing multi-domain sending. Um, the Growth Machine and HeyReach are good tools for sending out LinkedIn invites or messages etc at scale. They can do sequencing. Um, not just in LinkedIn, but that's what they're especially good at. They can also, at least The Growth Machine, I'm not sure about HeyReach, but The Growth Machine can also do email and Twitter and some other other platforms. Um, but the reason why I wouldn't use The Growth Machine instead of Smartlead is The Growth Machine, each one person has one inbox, so you can only send a limited volume of emails. Um, so if you're doing something low volume, like, um, I I don't know. Like, let's say you you only want to respond to the very high-value visitors that visit your website and you're screening them, and let's say you only have 10 a day. So, that's something you might want to do in The Growth Machine. Um, but something if you're sending out, want to send out 1,000 or 10,000 emails a day, um, Smartlead or Instantly is more what you'd want to use. Um, but LinkedIn outreach is extremely powerful. It's going to usually get notably higher reply rates than email. Because email, it's easier for people to send larger volumes, which means everyone's sending larger volumes, which means, um, everyone's inbox is crowded. It's easy to ignore people when you get 300 emails a day. So, uh, yeah, LinkedIn is is is much more restrictive on how many people you can message per day. So, they're they're there's less noise. And then finally, you can also send data to your CRM and the CRM itself can send forward data on to other platform platforms like Smartlead, Instantly, um, etc. So, that's what we got. We've We've covered so far the most common use cases and these four phases. Almost to the end, let's talk quickly about Clay's top about 20 enrichments, and, um, I think I said that uh Josh from the growth team pulled pulled this data. So, let's let's talk I'm not going to talk about all of them, um, I don't think, but let's quickly cover the very most common ones. Okay. So, ChatGPT is, by far, the most most commonly used one. It's a very flexible general way of Wrangling data and because ChatGPT, um, 4o-mini just came out, it's, um, about half the price of ChatGPT 3.5, but it's more powerful. And so, it's starting to make more and more sense to it's basically, it's trivially expensive. Um, so you can use it for lots of things that yes, it costs tiny fractions of pennies, but you can use it at a large scale and not rack up too much of a too much cost for yourself. Um, and it's so, ChatGPT is powerful. Claygent, also understandably, is is very flexible, um, it's a very powerful tool for scraping data when you don't know exactly what you're going after or how how you're going to get that data or it's going to be different in every case. Um, Claygent is is can iterate towards what you want. Um, then we have lookup rows in other table. So, um, if you with more complex table setups, you can store data in one table and pull it out of the of the table when you need it. Um, this is a very powerful pattern. Just to give an example, one one way that we've used this and and you'll learn how to do it is something called a query cacher. So, in order to not spend too much money on enrichments, um, let's say that you want to write a haiku about a person's city. And you have a row, uh, table of 10,000 or 100,000 people. Um, 90,000 of those cities are probably repeats. So, why are you going to spend credits on those 90,000 repeats when you can just store the results that you already got for that poem about their city in another in another table, and then you check that table first. Ah, is there a is a poem for their city in there? Yes, there is. You pull the the poem out and you don't have to use AI to generate a poem. So, um you store that that cached, um stored data in another table and then you use the lookup rows in other table, um to retrieve that data. Just a a simple, or just an example. Um, HTTP API, I mentioned before, is a way to connect Clay with any product that has an API and it's especially useful if Clay doesn't have a native enrichment column for it. Um, normalized company name, you take a company name like Acme, uh, Inc. Limited, and it reduces it to Acme so it looks more natural. Lookup company and other, let's skip that one. Um, I'm asking the team what that is and I don't, I don't know what that is. Uh, so, we'll see. Um, email waterfalls are extremely powerful for getting more coverage. Um, if you don't have, you you check Prospero, or ICPs or whatever. Um, you check one provider, if the provider doesn't return data, you check the other and the other one. Um, lookup company on LinkedIn, add lead via Smartlead, normalize URL, we'll clean, we'll clean a URL. Um, Write to table is kind of the counterpart to look up rows in other table. So when with that in that haiku uh situation, if you let's say you're going to write a haiku for Tampa, I, I check the lookup, uh, I check another table, does the haiku exist for Tampa? No, it doesn't. Okay. So, when I generate a haiku for Tampa, I have to send that data to the other table so that later it can be referenced. So this, this, um, Write to table does that. LinkedIn find person, it's pretty self-explanatory. You can take their LinkedIn URL. It'll give you data about them. Um, get domain from company name, you can give Clay, um, a company name and it'll return the domain of the company. Let's see if there's anything else I want to cover. Adding a row to Google Sheets. Uh, Clay has problems at scale. Um, and it it's hard to query large data sets, uh, pretty granularly with Clay. You can't query it with something like SQL, um, as some people call it SQL. Um, sending that your data to a Google sheet allows you to more granularly query it and not not run into problems after 100,000 rows. Um, let's stop there. So, hopefully that gives you a flavor of what are the most common enrichments used at scale. If you aren't familiar with these now, you certainly will be soon. Um, and if there's anything here that piques your interest, feel free to, of course, look into it. Every little piece of knowledge you fill and it's going to make everything else, uh, moving forward, easier. Okay. So that is really it for the perspective exercise. We covered the most common use cases, outbound, inbound, CRM, enrichment, pre-call preparation, lead scoring. We talked about list, enrich, wrangle and send. And then we talked about the most common 20 enrichments. So hopefully from this video, you've started to get a little bit more of a, uh, if you didn't have it, um, a mental framework of, uh, how to think about Clay and how Clay is used. All right. Um, we will next jump into exercises. You'll start doing real exercises, um, and looking forward to it. All right, thank you.