YT YouTube2Text

YouTube2Text → Transcripts → How to get SO many leads you don’t know what to do with them

How to get SO many leads you don’t know what to do with them — Transcript

by Nick Saraev · 7,653 words · 1,092 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Hey, this is an exhaustive and updated
0:01list of every way that I currently know
0:03of to scrape and generate leads for
0:04services. I use these every day over at
0:06Left Click to generate hundreds of
0:08thousands of dollars across a variety of
0:09industries for clients using outbound.
0:11So, this is the sauce. There is no sauce
0:13like it. Everything that I'm showing you
0:15guys is currently working as of the time
0:17of the recording of the video.
0:18Obviously, outbound methods change, but
0:20in addition to showing you guys what the
0:21method is, how to go and do it yourself
0:23with a live demonstration, and then
0:24where you need to go in order to get
0:25these resources, I'm also going to give
0:27you guys a Google Doc that just has all
0:29of them laid out in front of you, top to
0:31bottom. So, if you guys want to, I don't
0:33know, create SOPs or give this off to
0:35your team to do some education or
0:36learning, feel free. You can find that
0:38down below in the description. Let's
0:39start with the first. So, there are two
0:41ways that currently exist in order to
0:43scrape leads. Now, when I say leads for
0:45the most part, I'm meaning email
0:47addresses, although obviously phone
0:48numbers, first names, last names, and
0:50companies are part of that as well. Let
0:52me run you through the two that people
0:53are using all day long that they don't
0:55tell you about. The first is nominative
0:58email enrichment. Basically, what
1:00happens is you get company domains and
1:02or first names or last names. Then you
1:05take the name of the person, so in my
1:06case Nick Sarraf, and you combine it
1:08with the domain of their website, so
1:10Left Click. Then you just add, remove,
1:13and multiply these in various
1:14combinations. Nick at leftclick.ai,
1:17nick.s at leftclick.ai, nick.sarraf at
1:20leftclick.ai, over and over and over
1:23again until you find one that works.
1:25When you find one that works, boom,
1:26there you go. That's your email. Why
1:28does this work? Because most of the time
1:30people follow pretty standard patterns
1:32and conventions when they name their
1:34emails. And so, you just get to take
1:35advantage of that. You know, if I'm
1:36Nick, I'm not going to name myself
1:38[email protected].
1:40Although, maybe I should start, then I'd
1:42get fewer damn cold emails. I'm going to
1:44name myself like Nick at leftclick.ai,
1:46so people know who the heck they're
1:47talking to. Okay, so this is number one,
1:49and you'll find that a lot of the
1:50methods I'm going to show you guys are
1:51going to include this. The second method
1:53is using what are called HTTP requests
1:55to then generate a bunch of data that
1:57you then scrape and parse. Now, if you
2:00guys don't know much about technology or
2:01about IT or how to do this sort of
2:03thing, that's okay. Most of the stuff is
2:04abstracted away to tools nowadays
2:06anyway, but just so you have a
2:07bird's-eye view, what you do is you
2:09request the content of a page, like with
2:11a tool, Make or n8n or Zapier or
2:14whatever, then you parse it. Parsing
2:16just means breaking it down into a
2:17language that you understand. Once
2:19you're done with that, you have this
2:20thing in a language you understand, you
2:22just look for strings or sections of
2:24characters that are in specific
2:26patterns. For instance,
2:27[email protected].
2:30Okay, so Nick at leftclick.ai.
2:34Once you have data in this format, it
2:35becomes very easy to just dump it into a
2:37list or a database or something like
2:38that, and then you're good to go. You've
2:40essentially got email addresses as well.
2:43So, they do this not only for emails,
2:44they do this for phone numbers at
2:46various formats, they do this for for
2:47domains, for all sorts of information,
2:49but we're going to be going specifically
2:51over email addresses today just cuz
2:52that's the easiest. I'll also talk a
2:53little bit about phones, too. So, the
2:55first way is to use Air Scape to then
2:57find a list of domains. You take that
3:00list of domains, feed it into an
3:01enrichment tool like An Email Finder,
3:03and then you get a bunch of emails. So,
3:04first things first, use this tool
3:06airscape.ai, which I'm not affiliated
3:08with or anything, to find companies. You
3:10then select, you scroll all the way to
3:12the top, your location. I'm going to do
3:13United States. Company size, I'm going
3:15to go at 1 to 10 and 11 to 50. And then
3:18all the way down here, keywords. I'm
3:20just going to type in PPC agency, that's
3:22way easier for me.
3:23You then click preview companies, and
3:25it'll go and it'll actually find a list
3:27of companies that fulfill uh your
3:29requirements. So, I'm looking for things
3:30that have the word PPC agency in the
3:31title. As you can see, I now have a
3:32giant list of PPC agencies all over the
3:34world I can sell to. Then you click load
3:36all companies to a table.
3:38You access the results, and you get
3:39something that looks like this. Now,
3:41most of the time lists like this aren't
3:42going to include the domains. Instead,
3:44you're going to include the company
3:45names, company LinkedIn URLs, and other
3:46fields. But what we need is we need the
3:48domains to perform that nominative email
3:50enrichment, right? So, luckily, tools
3:52like Air Scape and other ones like Clay
3:54and stuff like that as well have this
3:55built in. Just click on the company
3:56domain column, then you go down to
3:58select where you're going to be pulling
3:59the company name from in order to do the
4:01lookup search, and then I'm just going
4:03to use the same company name thing. I'm
4:05going to add the current page's rows, so
4:06I'll do this on 40, and then I'm going
4:08to run a workflow. Now, as you see, this
4:10is running top to bottom through a giant
4:12list of records to pull the domain names
4:14from these records. So, Zozimus, Voxter,
4:17Triosco, The PPC Guys, The Agency RE,
4:19and so on and so on and so forth. We
4:21just found a bunch of domains. So,
4:23that's pretty sweet, right? What we do
4:24now is we go to export, and I'm just
4:26going to download a CSV. I'm only going
4:28to do these selected 40 rows, that's
4:29really easy, and then click this button.
4:31What we have now is we have a list of
4:33companies, which is pretty solid. I'm
4:34just going to resize this so it can be
4:36uh visible, and now we have all these
4:38domains. What we do after we have all
4:40these domains is we feed this into an
4:41enrichment tool like An Email Finder, so
4:43I'm going to head over here to the top
4:45right-hand corner. I'm going to go bulk
4:46search, and then I'm going to go new
4:47bulk, and you upload, click decision
4:50maker search, select your spreadsheet,
4:52paste it in. It'll automatically map the
4:54company names and domains column, then
4:55you click process, and it'll go through
4:57and it'll actually find emails for
4:59records in this data set. So, as you can
5:01see, we've processed 30, we've got 21
5:03emails so far. Okay, and so after our 40
5:06rows were processed, we received an
5:08enrichment rate of 70%, 28 out of the 40
5:11were considered valid, and as you can
5:12see, we actually have the email
5:13addresses over here, although obviously
5:15I'm obfuscating them. Now, if we want to
5:16download them, we just click this
5:17button. This is going to charge us a
5:19certain number of credits. They usually,
5:21what I like about them is they only bill
5:22you based off of the ones that were
5:24found, like a lot of email services bill
5:26you just based off the number of
5:27attempts that they made as opposed to
5:29the number of ones that were actually
5:30found. And then once you're done, you
5:31have the name of the person, you have
5:33their role, you have their LinkedIn
5:35URLs, you even have their email,
5:36although I am obfuscating this. Same
5:38thing with valid email over here.
5:41Uh you have their type, AMF status, that
5:43just stands for An Email Finder, company
5:44name, company domain, company LinkedIn
5:47profile, location, industry, and so on
5:49and so forth. So, hopefully you guys
5:50could see, this is more than enough
5:51information to, you know, craft a cold
5:53email campaign or maybe some sort of
5:55like outbound LinkedIn campaign,
5:57whatever the heck you want. You have it
5:58all right here. The second way is using
5:59LinkedIn Sales Navigator to get a list
6:01of people. With those people, you then
6:03feed them into Vain or an alternative to
6:05turn this into a Google Sheet with
6:07leads, then you import them into an
6:09enrichment tool like Vain or An Email
6:11Finder to get the email addresses. So, I
6:13have a list here for PPC agency, okay?
6:16What I'm going to do is I'm going to
6:16look specifically for founders,
6:19co-founders,
6:21CEOs,
6:22you know, owners, all of these terms
6:25that demonstrate that they're some sort
6:26of decision maker. Then for the
6:27geography, I'm going to go look for
6:29people that are specifically in the
6:30United States. This list is still pretty
6:32big, so just for the purposes of this
6:34demo, I'm also going to filter based off
6:35of head count, and maybe I only want
6:37companies that are between 201 to 500
6:39people. I'm looking kind of enterprise
6:40level. So now, as you can see, I have a
6:42list of 62. Realistically, you should
6:44have a list of probably like 2,000 to
6:462,500 for best results. And what I can
6:48do next is I can just copy this URL. I
6:50can go into a tool like Vain.
6:53Now, with this tool, okay, I just click
6:55create order. I paste in the LinkedIn
6:58Sales Navigator search URL, I check it,
7:00and then it'll add this to a database to
7:03be queued automatically. It's found 62
7:06results. I'm going to call this example
7:08for YouTube. Now, Vain offers you the
7:11ability to enrich directly in the app,
7:13so you can get the emails of things if
7:14you click this button. Um I didn't
7:16select that feature today, so I'm just
7:17going to click create order.
7:19And now, on bottom left-hand corner,
7:21it's actually going through and it's
7:22scraping things using my LinkedIn
7:24profile. The specifics on how to set it
7:26up you can find um on Vain once you've
7:28signed up, but essentially you have to
7:30get like your LinkedIn cookie, and then
7:31you have to export the LI_AT token
7:34before connecting it that way. Once
7:36you're done, you can scrape leads
7:37really, really easily. So, this is an
7:39example list that I did for PPC agencies
7:41a while ago. Just going to click
7:42download.
7:43And you got a lot of data here. If I go
7:46um import, and let's just replace this
7:48with my new data from Vain, assuming
7:50that this is like the list that I just
7:52exported.
7:53As you guys can see here, I'll get a lot
7:55of data. I have the first name, last
7:57name, some of them even include the
7:58emails. I have phones, summaries, job
8:00titles, job descriptions, and so on and
8:02so forth. What do I do with this list?
8:03Well, it's pretty easy. I can just pump
8:05this directly into a tool like An Email
8:07Finder, like I just had before. So, go
8:09to new bulk, and then I can look again
8:11for decision makers, or more pointedly,
8:13I can look for the person name search um
8:15endpoint.
8:16The reason why this is valuable, you'll
8:18see in a second,
8:21is you allow the algorithm to use
8:23nominative enrichment based off the
8:25first name, the last name, and then
8:29the company name over here. Now, we
8:31don't have the domain name itself, what
8:32we have is the company name. So, An
8:34Email Finder, really any other service,
8:35is going to look for the company domain
8:37using some Google magic, and if it finds
8:39the company domain, then it'll be able
8:41to enrich this. So, once you're done
8:42with that, just click process, and as
8:44you can see, we're now going through
8:45that list again. We processed 60, 70,
8:4880, 90. Out of 100, we have a val-
8:50validity rate of something like 80 or
8:52so, 80%, which is very solid. I mean,
8:54how often is it that you're going to get
8:5580% of a list that you put in? This is
8:58going to fluctuate and presumably go
9:00down with time, but you get as the end
9:02result a list that looked really similar
9:04to our previous one with a big list of
9:06emails. It's just we have even more
9:07information now, since we have the first
9:09name column, last name column, and a
9:11bunch of personal stuff, too. The third
9:12method is enterprise specific. This
9:15takes advantage of a little-known
9:16arbitrage hack within An Email Finder
9:19that allows you to get significantly
9:20more emails for significantly fewer
9:22credits. The exchange rate ends up being
9:24about one credit to 40 emails, which is
9:26bonkers. Basically, in An Email Finder,
9:28you have what's called a company search,
9:30where if you put in a domain like
9:32OpenAI, Forbes, Anthropic, Microsoft,
9:34whatever, okay, you can actually get all
9:37of the emails at that company, at least
9:39the ones that are publicly available.
9:41So, let's say I open up Forbes over
9:42here. I have all these emails. It's
9:43bonkers, right? Down at the bottom in
9:45grayed-out text, you can download all of
9:47the known emails at that domain. 176,000
9:51with Forbes endings, okay? And then
9:52you'll notice that there's a
9:53differential between the number of
9:55credits you pay and then the total
9:56number of emails you download. This is
9:57about a 1 to a 40 uh ratio.
10:00Um the reason why this is so great is
10:01because you can obviously arbitrage
10:03token or a credit cost, and then for the
10:05price of one credit, you can get 40
10:07emails. This is also great if you're
10:08targeting enterprise cuz a lot of the
10:10time, like good enterprise sales
10:11involves you just targeting like one
10:13company or a small set of companies. So,
10:15you have some list in front of you, then
10:16you just dump that list into Anymail
10:18Finder, you get 10,000 emails, most of
10:20them are two or three companies, then
10:22you pitch something that's very
10:23company-specific. Usually, it only takes
10:24a few people at the company for you to
10:26be able to, you know, sell your thing to
10:28so that you then as, you know, a service
10:30seller who's looking to get some social
10:31proof can say stuff like, "Our tool is
10:33used by the team at Microsoft." or
10:34whatever. How to actually go about doing
10:36this at scale? Just head back to Air
10:38scale, make a company table looking for
10:40big businesses like this one, Worldpay,
10:42Warner Bros, Vistaprint, whatever. Find
10:44the domain name as well, and then export
10:46it, and then all you do is you just go
10:48back to Anymail Finder, and then instead
10:50of running an individual search, do a
10:51bulk search, and then it'll actually
10:53give you a comma-separated list of all
10:55of the email addresses, which is wild.
10:57Now, you're not going to have all all of
10:59the email addresses. If you run this
11:00thing in bulk, you won't have the option
11:02to arbitrage the 1 to 40, but you can
11:04usually still get, you know, five, 10,
11:06sometimes even like 15 emails from just
11:08one search per row. So, if you want like
11:09an automated way of doing it, this is
11:11definitely pretty solid. Otherwise, you
11:13know, if you're targeting, let's say, 10
11:14or 15 companies, just get your ICP, you
11:16don't need to use Air scale for that,
11:18bump them in through a company search
11:19one at a time, download all of the known
11:21emails, and then run them through some
11:22sort of validation service to trim them
11:24down to the most likely ones. If you
11:26guys want um a way then to get
11:28information on who these people are, all
11:30you have to do is just take their email
11:32address, publish them or put them into
11:34Google like this, and then a lot of the
11:35time you'll actually find the first name
11:37and the last name just directly through
11:39Google. Then you can build a really
11:40simple automated pipeline using n8n or
11:42make.com to do this for you
11:44automatically. I'll show you a little
11:45bit more about that later, but then you
11:47can extract first names, last names, and
11:49emails for a fraction of the price. The
11:50next method is using Apollo to generate
11:52a list similar to LinkedIn Sales
11:54Navigator, then exporting that list
11:56using a variety of tools like Ample
11:58Leads, Leads Rapidly, or maybe Export
12:00Apollo. This then gives you a CSV, which
12:03then you can feed into whatever tool you
12:05want. The good thing about Apollo is
12:07most of the leads already include the
12:08email address, so you usually don't need
12:09to pay for any additional enrichment.
12:11Now, this used to be the best tool, and
12:13I'd recommend it all the time within
12:14Maker School, so much so that it was
12:16actually in the curriculum of Maker
12:17School for quite a while, where I
12:19actually showed people how to, you know,
12:20get up and running with um the scraping
12:23infrastructure.
12:24Unfortunately, they recently changed
12:26things so that I got to refill my dang
12:28video now.
12:29Because Apollo was scraping uh another
12:32service, LinkedIn Sales Navigator, and
12:34then other tools, these ones here are
12:35scraping Apollo. The point I'm trying to
12:37make is this is very quickly obfuscated,
12:39unfortunately, and it just doesn't
12:40really work anymore. Still, I'm going to
12:42include it for posterity because there's
12:43some rumblings showing that maybe this
12:45still works, it's just working at a lot
12:46slower uh basis. Um so, the way that
12:49this works is you basically grab your
12:50list, okay, in Apollo, same ideas
12:53before, get a bunch of job titles,
12:55founder, partner, co-founder, owner,
12:56whatever. Get a bunch of locations, then
12:58you have a list and a URL. You copy the
13:00URL into a tool, and I'm just going to
13:02use one of these, why don't we just use
13:04Leads Rapidly for now. Once you're in
13:06the tool, in my case Leads Rapidly, you
13:08just paste in the URL.
13:10This is then going to tell you how many
13:12leads you are going to be scraping. So,
13:14I have a list of 3,040.
13:16Uh maybe I only want to scrape, I don't
13:18know, you know, 1,000 of those leads. I
13:20can then click purchase, and then it'll
13:21consume my credits right up here.
13:23Once you're done, you'll have a list of
13:24different transactions. So, one of these
13:26failed, these fail every now and then,
13:28but um your completed records, you can
13:31then right-click, go download, then
13:33you'll have a list of them just like you
13:34had before.
13:36And if you import these into some tool,
13:39again, you get a list of first names,
13:41last names, full names, email addresses,
13:44their titles, company names. I mean,
13:45literally everything that you could
13:46want. You also get phone numbers and
13:48sanitized phone numbers. We'll talk a
13:50little bit more about how to get more of
13:51these um later, but uh yeah, this is
13:53more or less like the most
13:54cost-effective way to generate leads.
13:56It's just it's very shaky. Some of these
13:58providers have taken forever, and I
14:00don't know if they're still scraping
14:01them at an automated basis, but who
14:03knows, maybe in the future these things
14:05will be back to service. Just wanted to
14:06mention them at least for posterity's
14:08sake. The next few tools take advantage
14:10of a platform called Apify, which hosts
14:12a bunch of scrapers for a variety of
14:13services. The first one is going to be
14:16the Google SERP results scraper, where
14:18we assemble a giant list of domains
14:20using Google Search, then with those
14:22domains, we feed them into Anymail
14:24Finder to get a bunch of decision-makers
14:26at those businesses. So, I'm just going
14:28to go to Google SERP real quick. There's
14:29a search results scraper right over here
14:31on Apify. I'm going to try this, they'll
14:33let you try it for free with certain
14:35certain number of tokens.
14:37All you do is you just type in whatever
14:38the term that you would search on Google
14:40to get companies like this. So, I'm just
14:41going to do Calgary dentists. I'm going
14:43to go down here to results. Why don't I
14:44just go one page and 10 results per page
14:46to keep it simple.
14:48You can select a bunch of settings
14:49surrounding, you know, how many of these
14:51leads to scrape, um how to do this,
14:53whether or not you want um paid results
14:55only or or organic results only or
14:58whatever.
14:59After you're done, you'll then get um
15:01essentially a bunch of data, which we
15:03can preview.
15:04This now includes a bunch of websites,
15:06so we have the URLs right over here.
15:09If I export this, then we scrape the
15:11organic results, click download, take
15:14those organic results, and then upload
15:16them into a Google Sheet just so you
15:18guys could see what's going on.
15:20And now you guys could see we have the
15:21title of the clinic, the URL, the
15:24description, and then we have the
15:25displayed URL. I mean, both of these are
15:27basically the same, but because we have
15:29the URL, we could take this URL into a
15:31tool like Anymail Finder. So, I just go
15:32over here, click new bulk. You can
15:34choose decision-maker or company search,
15:36depends on kind of what you're going
15:37for. I'm then going to select the
15:39domains column here,
15:41and then I'll click process. And as you
15:42guys could see, it's going and doing
15:44some scraping for me.
15:45Looks like of the five, it's found four,
15:47so an enrichment of 80%. And when all
15:49said and done, we achieved an enrichment
15:50of 50%. Now, this isn't as good as it
15:53could have been, but still pretty
15:54reasonable. On average, I get between 40
15:56to % of all of the domains that I feed
15:58in, and luckily the Google SERP scraper
16:00is really cheap. Um as you can see, it
16:02only costs something like, I don't know,
16:04$3 uh per 1,000 records. If you upgrade
16:08your plan on Apify, it's significantly
16:09less, but um yeah, you know, this
16:11economies of scale come together.
16:14And then when you import them, you get
16:15the name, the result title, aka the um
16:18title of their role in the business,
16:20their email address, then even some more
16:22information including descriptions and
16:24stuff like that. All stuff that you
16:26could use to whip up a really cool
16:27campaign. But Google SERP isn't the only
16:29thing you can do. A lot of the time
16:30we're working with local businesses, and
16:31that's where the Google Maps scraper
16:33comes in handy. So, I'm just going to
16:34click try for free here. And again, just
16:36like I did earlier, I'm going to look
16:37for, I don't know, dental clinic. Maybe
16:39this time we're going to do New York,
16:41and maybe this time I only want to
16:42extract 10 places.
16:43Then you click save and start. Now, what
16:45we're doing is we're scraping specific
16:46map listings for people in an area,
16:49which is pretty sweet. So, as you can
16:50see, I just ran this, and just like we
16:52had before, um just out of the top
16:53right-hand corner, this time I can just
16:55scrape everything.
16:57Once you're done, you can go back here
16:59to your sheet. I'm just going to import
17:02these leads
17:03just to show you guys what these look
17:05like.
17:06Now, we have 10 businesses, Smile
17:08Crafters, Floor El Vivar, W Dental,
17:10whatever. These are all dentists
17:11basically in our target markets,
17:13presumably. Once we're done, we just go
17:14back to Anymail Finder, click new bulk.
17:17Let's go decision-maker search, then
17:19feed in the exported list from Google
17:20Sheets. Now, we can see as we still have
17:22that domains column, so I'm going to
17:23upload this.
17:25And same thing as before, we're now
17:26going to find their emails. Now, that
17:28last sheet I tried running through
17:30because it only had 10 records, you
17:32know, law of large numbers came in, and
17:33I think we only scraped three or four of
17:35the leads, but just to show you guys
17:36what you can realistically accept with
17:38Google Maps, not all of the websites
17:40that you guys are going to be getting
17:41are going to be websites that like the
17:42company owns themselves. A lot of
17:44websites are going to be like Facebook
17:45pages and stuff, and obviously we can't
17:47find the decision-makers at Facebook
17:49anywhere near as easily um as you know,
17:50some small company. So, on average,
17:52you're probably going to have validity
17:53rates somewhere around like 20 to 30%, I
17:56want to say. Um I just reran this on a
17:58list of 50, and um you know, I think
18:00we're going to end up getting somewhere
18:01around like
18:02probably 15 emails or so. I think right
18:04now we have what, 14?
18:05In my case as well, I fed in the same
18:08domain a couple of times, so this is
18:09probably just not very worthwhile to do.
18:12You should probably deduplicate the
18:13list. Um we ended up with a total
18:15enrichment rate of 32%, so um nowhere
18:18near as valuable or cost-effective as
18:21the Google SERP leads. That said, this
18:23can be very valuable if you have some
18:25that's very location-specific. So,
18:26spending a little bit more money for
18:28these leads can make sense if you are
18:29targeting people in a local area for a
18:31local product or service. You'll have
18:33significantly higher rates if, you know,
18:35you can mention things like, I don't
18:37know, what's across the street or where
18:38they are in a specific geographical
18:40region. Back in the day, I used to go
18:41door-to-door, and so we would look for
18:43data sources like this in order to
18:44pre-qualify our lists before going out
18:47and then actually knocking or calling.
18:48The next method involves using the
18:50Google SERP and Maps scrapers to feed
18:52into an automated workflow. Now, I set
18:54this one up in n8n, which is a simple
18:56drag-and-drop, no-code tool, and I've
18:58already preset all of the data. I just
19:00wanted to show you guys what this looks
19:02like just for the purposes of your guys'
19:03uh benefit. Um basically, what I have is
19:06I have a workflow set up that automates
19:07the process of running an Apify Google
19:09SERP scrape, so that's what's happening
19:11right over here under the hood. I'm
19:12basically looking for a term like
19:14Calgary dentist, just using Apify, I'm
19:16just doing it through their API, so kind
19:18of behind the scenes, their back end.
19:20Then I'm extracting the data. I'm
19:21limiting the number that can go through,
19:23and then most importantly, I'm actually
19:24scraping the website. So, I'm getting
19:26the website, I'm scraping it, and then
19:28I'm feeding it into AI to get details.
19:30I'm just going to click execute workflow
19:32just to show you guys what this looks
19:33like on three records so you guys could
19:34see.
19:35But basically, [snorts] every time I
19:36call a website, I end up getting a bunch
19:39of code back, right? Now, this code here
19:41on the right-hand side actually just
19:42contains all of the details of the
19:44website for me. And a lot of the time,
19:46this will contain things like the email
19:47address, like the phone number. You just
19:49have to parse it in specific ways. And
19:51so, that's what I'm doing here with
19:52these fields. I'm turning this big block
19:54of text into a much shorter version
19:56where I can start parsing things out
19:57like phone numbers as you guys can see
19:59over here.
20:00Now, what's really cool is you can start
20:01combining this procedural mechanism of
20:04generating um you know, website scrapes
20:06and stuff
20:07with AI-based mechanisms of extracting
20:09data. So if I feed in all of this data,
20:12which is very unstructured as you see,
20:13into a model like um OpenAI's GPT-5, it
20:16can go and output something that looks
20:17like this. As you see here, we got a lot
20:19of data. We got the record ID, the page
20:21title, page meta description, industry.
20:24We get the owner name, owner title. Now
20:26um owner email in this case wasn't
20:27found, but we can find these I want to
20:29say about 10 to 20% of the time.
20:31Although keep in mind when you do, it's
20:32likely not going to be the owner email
20:34themselves. It's going to be something
20:36like, you know, info@ dentalclinic.com
20:39hypothetically. Still, a lot of the time
20:41this goes directly to the owners, then
20:42you can use information like this to do,
20:44you know, whatever the heck you want. Um
20:46my AI model extracted phone numbers,
20:49bunch of contact form notes, even some
20:51address information including um angle
20:54suggestions for pitches and so on and so
20:56forth.
20:57So pretty cool to see just how much data
20:59we can pull um in a semi-automated
21:01fashion. And after I just dumped all
21:03this stuff into a database which um I
21:04put up together a while ago. As you can
21:06see, we scraped a couple of emails and
21:08dumped them all into the Google Sheet.
21:09So we have a couple of these general
21:11ones and then uh it even talks about
21:12what those are.
21:14Is this as valuable uh scraping source
21:16as let's say AnyMail Finder's Decision
21:18Maker Finder? No, but you can go
21:20significantly higher volume. And a lot
21:22of the time the marginal cost of sending
21:23a cold email nowadays tends to zero if
21:25you guys have all the infrastructure set
21:26up. So if you're sending cold emails or
21:28DMs or whatever, you can also extract
21:29LinkedIn profiles, Instagram profiles,
21:31and stuff like that. Um also keep in
21:33mind that you can combine this with any
21:34of the other scraping mechanisms that I
21:36talked about to get not only the email,
21:38but more or less whatever other
21:39information is present on any of their
21:41web pages. The next way is purchasing a
21:43list on a data set marketplace like
21:45Bright Data. Now, if you guys are
21:47long-time watchers of my channel, you'll
21:49know that I haven't talked about this
21:50before and it's because I just wanted to
21:51purchase a couple of lists and try them
21:53out for myself before I discuss this.
21:55I want to say these work reasonably
21:56well. Leads are definitely cheaper, but
21:58a lot of the time these resources have
22:00been scraped forever ago. And because
22:01they've been scraped forever ago, like a
22:02lot of the time companies will have
22:04changed their email addresses or people
22:05just won't be using an email anymore.
22:08Anyway, [snorts] the specific um lists
22:10to buy are LinkedIn people profiles over
22:12here.
22:13Okay, Google Maps full information over
22:15here.
22:15Um you can get information from
22:18um
22:19Facebook as well and then Crunchbase,
22:21too. So if you go on LinkedIn people
22:23profiles, the way that this works is you
22:25have a basically a giant list of names,
22:27IDs, cities, and so on and so forth.
22:28They actually give you a little preview
22:29and sample data set. This is over 600
22:32million records and this is updated
22:34every quarter, so that's what they say.
22:36The actual total cost of this is $0.0025
22:39per record as you see. And then with
22:41this information, you can pump it into
22:42something like AnyMail Finder um and so
22:44on and so forth in order to like get
22:45more enriched lists. You can also use
22:47Air Skale to find specific companies
22:49instead of LinkedIn people profiles,
22:50then use that information. If you go to
22:52company information over here,
22:54um or just buy the company uh um
22:57you know, list and then filter based off
22:58the presence of websites. And then you
23:00guys can loop this into whatever other
23:01flow you want. The thing is, all of
23:03these services that sell lists will have
23:04some minimum order amount. So $250, you
23:07know, if it's like your very first time
23:08doing this sort of stuff, you can get a
23:09lot of leads for $250. You got 100,000
23:11leads for $250, but
23:13you know, do you really want to spend
23:14that much money all in one place? That's
23:16why I prefer typically like smaller
23:18scrapers. Anyway, I purchased a couple
23:20of these lists. Uh my email performance
23:21and outbound performance tends to be
23:23worse by a measure of about half. So
23:25basically, either these leads are on
23:27average half old or they've just fluffed
23:29the lists or something. Uh but they
23:31still work. I still get replies and then
23:32because of the very low record cost,
23:34obviously you can arbitrage this and do
23:35cool things, especially if you have uh
23:37tools like AnyMail Finder available to
23:39you to then get email addresses. Okay,
23:41next up you can use a tool like the
23:43Apify Instagram Scraper to scrape
23:45Instagram profiles for specific hashtags
23:48or queries. Then you can look with in
23:50those profile descriptions for uh email
23:52addresses. Enrichment rate on this is
23:54pretty low. That said, a lot of the time
23:56these include links. Because they
23:57include links, you can then use AnyMail
23:59Finder and other email enrichment tools
24:01to get emails there, too. So I'm just
24:03going to start by looking for uh posts
24:06and then I'm going to run based off of a
24:08search query over here. I'm going to
24:10look for the term automation.
24:12And then I think we're just going to
24:12search for users that have the term
24:14automation inside. What you end up with
24:16here is you end up with a giant list of
24:18all of the Instagram information
24:21And you'll see a lot of these are
24:23businesses that include things like
24:24WhatsApp numbers or email addresses just
24:27directly in the description. Um now,
24:30obviously I'm doing this in manual way,
24:31but hopefully you guys can appreciate
24:33how a moment ago I just um used an
24:35automated tool in order to go through
24:36and then filter and do a bunch of stuff.
24:38Well, guess what? You can use automated
24:39tools to go and filter and do tons of
24:41stuff for whatever your lead scraping
24:42sources. And this is what something that
24:44I would do for this as well. Uh you
24:46know, we've scraped 168 leads or so, so
24:49I'm just going to go export.
24:51Let's put this in as a CSV.
24:56Import.
24:58Drag and drop.
25:01And we have our lead list. Now this is a
25:03lot of leads. You'll see a lot of them
25:04also include things like domains just
25:06directly in their descriptions and so on
25:08and so forth. The reason why this is
25:09valuable is because now you have like a
25:11variety of different lead generation
25:13mechanisms. For instance, you could look
25:15specifically through the descriptions
25:16for email addresses and you'd probably
25:18get, you know, 5 to 10% of these.
25:20Or after you're done with that pass,
25:22once you've done your low-hanging fruit,
25:24you could then look for domains. Once
25:26you find domains, you could pass them
25:27through AnyMail Finder and then you can
25:28get even more. Now a lot of the time
25:29you're going to have first names and
25:30last names in the attributions. Um not a
25:33lot of people realize, but these
25:34attributions you could then use to find
25:36the the people themselves. Or if you
25:37have the domain name in addition to the
25:39first name and the last names, then
25:41obviously there's a lot you can do with
25:42that, too.
25:43So yeah, just one of many different ways
25:45to do this. In addition to using
25:47Instagram,
25:48you guys could do the exact same
25:49approach on TikTok. You guys can do more
25:51or less the exact same approach on X as
25:54well. Uh and so any of these social
25:56media platforms, you just take the same
25:57funnel, you scrape a bunch of profiles
26:00or posts based off of some search query,
26:02then you pump in the description, uh the
26:04profile picture, the alts uh uh codes on
26:07the images, and literally all the
26:08information into some tool, presumably
26:10like Any Data Maker probably. Or if you
26:12want to do it manually, you could do it
26:13with Google Sheets just like
26:14command-effing, but that's a lot of
26:16work. Then you have a giant list of
26:18everything that you need in order to do
26:19an enrich. You're probably going to get
26:21enrichment rates of like 20 to 40% or
26:23so, but there are multiple millions of
26:25profiles on Twitter, Instagram, TikTok,
26:27and so on and so forth in any niche, so
26:29odds are you're going to get enough for
26:30whatever your total addressable market
26:31is. The next mechanism is pretty sneaky,
26:33but basically um you guys know like
26:35LinkedIn Jobs? I don't know if you've
26:37ever gone on LinkedIn Jobs before, but
26:38if I go LinkedIn Jobs in Calgary, which
26:40is where I'm from, you'll see that
26:41there's a big list of people that are
26:43looking for hires in particular uh
26:46niches. So, you know, I'm typing in, I
26:48don't know, automation or something like
26:49that. Well, I see that there's 475
26:51people that are hiring automation
26:53people. Uh 475 companies essentially.
26:56Some are looking for Zapier, some are
26:57looking for automation testers,
26:59whatever. You know what's really cool?
27:00Turns out you can just take this list,
27:02you can go directly to a tool like
27:03LinkedIn Job Scraper. You could pump in
27:05the URL right over here
27:07and then you can actually find companies
27:09that are hiring for particular roles.
27:11Why is this so valuable and powerful?
27:13Because this is a list of people that
27:14literally have needs that they're
27:16willing to pay money for. I mean these
27:18people are saying, "Hey, I want somebody
27:20to go and, you know, do some automations
27:22for me and I'm willing to pay real money
27:24for it."
27:25This is about as close as you can get to
27:27people that have money in their hands
27:28that are willing to pay for your ability
27:29to solve their needs. Now what's really
27:31cool is do you have a company name
27:33column over here. So what you could do
27:34is you could download this list,
27:37take it back into AnyMail Finder, add a
27:39bulk decision maker search,
27:42paste in this list of people that are
27:44looking for work. Then where it says
27:46company name, just feed in the company
27:47name, okay? Once you're done with that
27:49process, and you know, when you search
27:51just based off company name, probability
27:53of you getting results will go down. As
27:55you see here though, we're feeding in,
27:57you know, 15 results. We're getting uh
27:58uh uh yeah, we're getting basically a
28:00richer rate of 100% so far. This is
28:02unlikely to persist all the way through,
28:04but it's pretty dang good already. So um
28:07you know, this is just another way that
28:08you can get a list and then instead of
28:09your list being kind of unqualified or
28:10at varying levels of qualification, you
28:12know that every single company here
28:14actually suffers from a need. Now, when
28:16you email, don't necessarily just email
28:18the CEO or whatever. If it's a big
28:19company, you're probably going to want
28:20to email a hiring manager or something
28:21like that. But yeah, as you can see, we
28:23just fed in 35 records and we have 29
28:26emails that are valid with another four
28:28that are considered risky. And I got to
28:30be honest, it's the wild, wild west out
28:31here. I send emails even if they're
28:33risky. I don't give a crap. The next
28:35mechanism is on a cold email platform
28:37called Instantly. So Instantly is a cold
28:39email platform that basically allows you
28:41to actually get leads directly within
28:43the email platform. So I ran a bunch of
28:45different campaigns the other day for
28:46one of my YouTube videos. Um in the top
28:48right-hand corner here, you can actually
28:49click on get leads, then you can upgrade
28:51to what are called supersonic leads. Now
28:53I've used this service before. This
28:55basically gives you access to a database
28:56of about 250 to 300 million leads and
28:59then you buy them for 4,500 um $97 per
29:03month for 4,500 leads.
29:05Will I recommend this for cost purposes?
29:06No. But do I recommend this for
29:08convenience purposes? Yes. Um you can
29:10absolutely just get up and running with
29:11these leads. These aren't going to be as
29:13high quality as the ones that we're um
29:14scraping with most other methods,
29:16especially the first three on our list.
29:18Um but, you know, if you just need to
29:19get a message out, the cost of the leads
29:22compared to the opportunity cost of you
29:24not sending those emails or messages
29:26today
29:27is usually very disproportionate. And
29:29it's almost always better, unless you're
29:31like really, really broke, to just start
29:33the campaign today and see what people
29:34have to say. The last mechanism is when
29:36you scrape a data source directly. Now,
29:39the benefits to scraping a data source
29:40directly are these are leads that are
29:42not easily accessible. And because
29:43there's a much higher barrier to entry
29:45when you scrape leads directly like I'm
29:47about to show you in a second and enrich
29:49them through a totally custom enrichment
29:50pipeline. The people that you're
29:52reaching out to tend to be a lot warmer.
29:53They tend to be people that haven't been
29:54burned out by tons of cold outreach
29:55before. The downside is you have to know
29:58a little bit about how to actually, you
30:00know, do the HTTP requests and the
30:01parsing and stuff like that. But just to
30:03give you guys a quick example, this is
30:04school. This is the same service that I
30:06host my community on Maker School where
30:07we guarantee your first paying customer
30:09for an AAS service in 90 days or you
30:11don't um pay. We give you all of your
30:12money back. So, the way that this works
30:14is you get skip listings where, you
30:16know, in the hobbies category number one
30:18is calligraphy school, number two is the
30:19acting lab, number three is the
30:20brotherhood of sand and so on and so
30:21forth all the way down. Well, you know,
30:23what's pretty cool? It turns out you can
30:25build a scraper to scrape this stuff. I
30:26built a scraper for as part of my most
30:28recent video. So, as you see what we're
30:30doing here is we are HTTP requesting
30:32school. And so, this is just one of the
30:34many ways that you could do something
30:36like this. But in my case, I'm literally
30:38just pinging this page first using
30:40Anedon. What I get is I get a giant list
30:42of all of the entries from one all the
30:45way up to 34. This gives me a massive
30:48list of data that I can then use to do
30:49more or less whatever the heck I want
30:51with. Now, this data is very big and
30:53it's very unseemly and so, I don't
30:55really like having the data in this way.
30:57So, what I do next is I extract it in a
30:59much easier way where I now have code
31:01that shows me what all of these elements
31:03are for. After I'm done, I've actually
31:05parsed all of this data on the left-hand
31:07side that looks really complicated into
31:08a simple structure where I have the page
31:10title, the description, I have the image
31:12link, I have the category type and name.
31:15I even have the amount of money that all
31:16these charge for their services.
31:18Hopefully, it's not a far cry to see how
31:19I could take all that information and
31:21then I could dump it into a big Google
31:22sheet that contains things like the
31:24first name and the last name of the
31:26person that runs the community, uh their
31:28Facebook pages, their Instagram pages,
31:30even things like their websites. From
31:32there, all you do is you pump them into
31:34a tool like an email finder. And in my
31:36case, when I did this, I pumped in
31:38something like 5,000 rows. I did zero
31:40work, zero post-processing or anything
31:42like that. Once you're done with that,
31:44you get a list of emails that looks
31:45something like this. And yeah, this is
31:47one of the most powerful things that you
31:49could realistically learn. It's one of
31:50the reasons why I recommend learning how
31:52to scrape if you are going to spend some
31:53time learning automation cuz you can
31:54compile quick and easy lists like this
31:56that let you send emails that achieve
31:58reply rates as high as this one is now
32:00at 11.5%.
32:02That is 11.5% of all the people that I
32:04emailed responding to me. And it looks
32:06like about a third of those responded
32:07positively wanting something to do with
32:09me, some sort of meeting or whatever.
32:11Okay, so I'm going to leave all of these
32:13methods here in this Google uh doc and
32:15then I'm going to provide that to you.
32:16All you need to do is just click below
32:18at the top of this description. Um this
32:19is free. You don't have to pay anything.
32:21You may have to enter your email
32:22address. If you do, I'm just going to
32:23send you some emails talking about how
32:24great I am. Um but yeah, that is more or
32:26less how you can get so many leads that
32:28you have no idea what the heck to do
32:29with them. Thank you very much for
32:30watching and I'm looking forward to
32:31hearing from you guys. Leave a comment
32:33down below if there are any lead
32:34scraping mechanisms that I missed out.
32:35I'll catch you all in the next video.

About this transcript

This page contains the full transcript of How to get SO many leads you don’t know what to do with them by Nick Saraev, generated from the public captions YouTube serves with the video. The transcript has 7,653 words across 1,092 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.