Funding Innovation with Netcapital
Tips for entrepreneurs and founders for raising capital. Whether you are interested in raising funds through equity crowdfunding, angel investors, or VC's, we'll share our best practices for list building and customizing your pitch for each audience.
Funding Innovation with Netcapital
How I Raised Millions of Dollars Through Email Marketing - With Archil Cheishvili
Archil Cheishvili has used some audacious strategies to raise over $6 million dollars for the company he founded, GenesisAI.
His story is one of relentless experimentation. After trying out unconventional platforms such as Reddit and Quora and even bus shelter advertising, his fearless approach led him to email marketing, where he built his own email server.
Learn his secrets for building an email system that offers greater control and efficiencies. But you don’t have to build your own email server to learn his tricks. This episode is packed with actionable advice, including best practices for subject lines, A/B testing, and strategies for crafting an optimal email sequence.
You don’t want to miss this!
If you want to raise capital for your business visit us at https://netcapital.com/
Hello everyone and welcome to Funding Innovation with NetCapital. I'm Kathy Krasler, the Director of Marketing here and today, my very special guest is Archul, the founder of Genesis AI, who raised over 4.6 million on NetCapital, and much more than that in total, and part of the way he did that was through cold email outreach, and we're going to talk about exactly how he did that. So, archul, it's so nice to have you here and welcome to the podcast.
Speaker 2:Thank you, kati, I'm very happy to be here.
Speaker 1:So my first question is first tell us how to correctly pronounce your name and then tell us about Genesis AI and what your company does.
Speaker 2:Yes, so yeah, archil. Here at Genesis AI, we have been building an AI API marketplace, basically, websites that crowdfund different AI APIs and packages them in a way that they are accessible and usable by, primarily, software engineers, and last 12 months, we have been very proud of one of the AI tools we developed, which is the email marketing space, so I'm pretty sure we're going to talk a lot more about that.
Speaker 1:Yes, and so can you walk us through your company's fundraising journey, from traditional rounds to platforms like WeFunder and NetCapital, and how much did you raise in total?
Speaker 2:Yes, so in total, we raised roughly six million. We started raising using traditional means. Then we did our first crowdfunding round on WeFunder. We raised around $630,000 on WeFunder. Then we did our next round on NetCapital and I have had very good experience with all of that. We raised in our first NetCapital round net capital round, I think just 1.3 million and then 2.6 million. So, yeah, after that we did uh, reggae plus, in which we raised around 200 000, and now we are doing, uh, our another raise on this capital, where we are around 800,000.
Speaker 2:In terms of how we raised, we have experimented with lots of different approaches here. We have been mostly reliant on Facebook and Google Ads and the customer base of crowdfunding platforms, both on NetCapital and WeFunder, and our own user base as well On Facebook and Google Ads. Initially it was really good in terms of ROI. It was very high around 3, 4x, high around 3, 4x and after that starting from around January 2022, roi dropped. This was time when iOS 14 privacy related to data tracking changes got deployed, and then the time when we started to experiment with other kinds of marketing channels.
Speaker 1:Yes, and now I want to tell our listeners when I first met you, you were a client of ours and we were starting to help you with Facebook ads and I had been helping a lot of other clients with Facebook ads and doing really well, but I actually ended up learning a lot from you. I had never met anyone like you. You were spending at least 50, 50, 50 to 100 times more than our other clients. You were fearless. I mean, talk about go big or go home the way you think. You're a grand thinker, so I just wanted to get that. I mean I learned from you. That was amazing, but so okay. So uh, ios made changes. Facebook return on ad spend started to really decline. So where did you go from there?
Speaker 2:Yes, first of all, thank you for compliments, kati, and second, yes, so where we went from there when Facebook and ROI started to drop, of course, we started first to optimize, to figure out if there was any way we could have gotten back up to around 3-4x ROI. I remember I ran many experiments with creatives, with audience targeting, with platforms on which ads would be displayed, all kinds of things. Second thing we did was okay, this is still not working as well as we want, so let's experiment with other digital advertising platforms like Reddit or Quora or Snapchat, and we ran some experiments there on Twitter Also. Like no crazy results there. So I started to think about okay, how about SEO? How about email marketing? How about, for example, running, but also try to go to conferences and find ways to network with people there? There are all kinds, both sort of virtual and physical right. So what really and I can talk a lot more about each of this, but what really worked well for us was an email marketing- actually Okay, but wait a minute Before we go there.
Speaker 1:You had told me previously that you experimented with what we call out-of-home advertising, like at bus shelters. I think that's so amazing. Like just fearless, you just tried absolutely everything. And one of the things I love about you is that you're like the apex of marketing expertise and like computer science, but we'll get into that later. But tell me how that out-of-home went. How did you track it? I mean, how was that experience for you? How did you even place it? Did you have to hire an agency? How did you do it?
Speaker 2:Yes, so a bus shelter. And what we did was I found a marketing agency in Miami. They were offering different places where you could have put advertisements, basically like very small billboard advertising. I think we paid very little, probably like eight hundred dollars in total. We just wanted to run an experiment. Our ad showed Genesiai. It showed the URL and QR code so that people could. The UI showed the URL and QR code so that people could scan the QR code and they can go directly to our NetCapital offering page, and we were also tracking how many people were scanning that and how many people were clicking that. Based on our analysis, in the end ROI was not good enough to continue scaling that part out, but it's usually just a good experiment to run, you know.
Speaker 1:Yeah, yes, that's what I love about you You're fearless and willing to try absolutely everything. So then, what you ended up doing was you experimented with a lot of different email providers like MailChimp, et cetera. Experimented with a lot of different email providers like MailChimp, etc. Can you talk about how many you tried and what were the limitations that you ran up against, and how did you get around them?
Speaker 2:Yes, for sure. So email marketing-wise, we experimented with pretty much all kinds of email platforms, ranging from large-scale senders to smaller-scale ones. So to start with large-scale ones like MailChimp, MailGun, SendGrid and those guys. Those are great if you are emailing to your own subscribers. If you are emailing to people who have not heard about you before, those are not ideal for that. They don't like to see people who are, you know who are not already in your own contact database.
Speaker 2:So they do have restrictions there and you figure out and learn about those restrictions when you go through this process of sending emails, about those restrictions when you go through this process of sending emails and then one day they tell you that you actually can't really send cold emails. So we learned this hard way that you cannot really do this on this platform At first and second. There are smaller scale ones like G Suite, for example. You can have multiple Gmail accounts that you can use to send small scale emails maybe 300 emails a day per G Suite account. Google is pretty lenient with allowing people to email to people who are not in their contact database already. But problem here is the scale. It's going to be very hard to be sending tens of thousands of emails. For that you will need like 100 domains or so, right, or more probably. Sometimes some people have 500 domains Very hard to maintain. So the ideal solution was I can talk about what solution we came up with, but Ideal solution was.
Speaker 1:I can talk about what solution we came up with, but it was very challenging for us to do successful cold email marketing using existing solutions. So just to clarify for our listeners, can you talk?
Speaker 2:about the difference between a cold and warm email outreach and how each one of those impacts response rate. For sure, for sure. That's a great question, kari. So cold emails are emails that are being sent to people who have not interacted with you previously. This means that they are not, for example, your existing website users or they are not people who signed up to receive information from you. Warm email recipients are people who have had some sort of interaction with you through different channels either a website, or maybe you met on a conference or many other ways and those people are at some degree, expecting to receive an email from you, and those people already know who you are.
Speaker 2:In terms of open rates, it's a lot more difficult to get good open rates on cold emails versus on warm emails, because it has to do with engagement and complaint rate that you are going to get for different kinds of emails. For engagement, if somebody knows you, they are more likely to respond to your email, ask questions or say something positive. If somebody doesn't know you, it's less likely they're going to respond, more likely they're going to subscribe, or more likely that if they're gonna respond, it might be something negative in nature, and that's engagement results in lower deliverability and lower inbox placement and for uh. When it comes to complaint rate, complaint rate here is defined as a percentage of email recipients that mark your email as spam, and warm emails have substantially less complaint rates versus cold emails and this substantially affects domain and IP reputation. That is needed to get good open rates.
Speaker 1:Nice, Thank you for explaining that and so, but we skipped over. We haven't gotten to yet my favorite part of this whole story. So you found that the email service providers just weren't working for us.
Speaker 2:So my objective was first to reduce the cost, of how much cost we were spending on Mailchimp and similar solutions. And second objective was to make sure that we would be able to send large scale cold email campaigns. And I was like, ok, maybe, how about if we build our own custom mail server? So what is custom mail server?
Speaker 1:Go ahead. What is a custom email server? Go ahead.
Speaker 2:Yes, so custom email server is basically you're going to find a way to rent a virtual computer in the cloud and you're going to set up an email system on that and in the cloud and you're going to set up an email system on that and in the end, you are going to act like a mailchimp. So you have basically your own small version of mailchimp. So I remember we did this um before. It took probably like around a week to do this manually. Yes, computer engineering involved and we started sending out emails and it really worked. We got results. We got great open rate, great clicks, great conversion rates from email recipient clicking to an investment and we're like, okay, maybe we should. We can take a next step here. Yeah.
Speaker 1:Okay. So first I'm guessing that the cost was much lower the cost per acquisition of an investor. But before we get there, how are you getting these email addresses?
Speaker 2:yes, so for to get email addresses. There has been, uh, so few things. First, emails that you already have, right, definitely we use it. Make sure sure that you're gonna send multiple emails to each people across different links. Second way is to experiment with different lead generation strategies. For example, on your website you can place a exit pop-up that's going to pop up whenever a person is about to exit your website, right? Or you can have some sort of give us your email if you want to see our company deck. So that's the second channel generating leads this way.
Speaker 2:Third channel is generating leads through advertising. Lead generation of Facebook and Google Ads is very cheap. You can have cost per lead less than one dollars. Fourth way is to find a way to scrape different websites all right, to collect this email address data. Fifth way is to actually predict people's email addresses based on first name, last name, name and domain name combinations. We made an incredible number of predictions on this and we were looking for. So I can talk a lot in details, but it was really all about looking for most common full name combinations, then looking for most common domain endings and then doing some sort of mix and merge to guess a person's email and then going through the email verification system to figure out whether the email was valid or not.
Speaker 1:And use it, yes, to improve it even more and you could do that, you could experiment in that way, because you were using your own custom server, right, so nobody was uh throwing you off exactly, exactly, yes.
Speaker 2:So when you own a custom mail server, you are in control, so platforms like Mailchimp cannot suspend your account anymore, because this is your own mailing servers that you are renting, so ambitious.
Speaker 1:OK, ok. So let's talk then about how that cost worked out for you, because I'm guessing it no longer costs you four or five dollars per lead, or you know whatever it was on Facebook yes, exactly yes.
Speaker 2:The call on Facebook, you know like, sometimes cost per click was was so high and cost per lead that, uh, definitely we, we got cost per lead we got through email marketing was substantially less cost wise. On MailChimp, for example, if you want to have 100,000 contacts in your database, I think that costs around $1,200-$1,500. If you want to be able to, if you want to have 100,000 contacts in our own version of custom mail servers, our own version of custom mail servers, our own cost was like probably like 40-50 bucks, so like 95% less as that cost on MailChimp substantially less. And in terms of some of the product requirements, kari, it was doing very well. You know we were able to choose scale, a simple UI, and also we were able to a simple UI and also we were able to, you know like, do all these A-B tests and iterations that we were doing on MailChimp.
Speaker 1:My gosh, that's so amazing. Okay, so let's get into sort of best practices. You've done a lot of experimenting, so can you share any tips on crafting like the best email subject line, email content that kind of maximizes open rates and engagement?
Speaker 2:Yes, for sure. That's a great question. So we did probably around 50 different A-B tests. Initially I was thinking to do specific A-B tests, for example okay, is this a subject line? For example, which emoji to use in a subject line, right?
Speaker 2:But then, when I started to think about, I was like this is probably not going to be helpful because a best emoji you can use today will most likely be different best emoji you can use six months from now. Because, uh, you know, whenever people find that, for example, emoji x is working really well, they are going to overuse it. And because they are going to overuse it on the recipient side, people will be like this is weird. I have seen an email like this. So people will engage less. And then Google's and Yahoo's machine learning algorithms, which track and optimize for engagement, they will penalize emails. That's going to include those emojis. So it's very much iterative process.
Speaker 2:So I was like I just need to learn conceptually what kind of emails with what kind of subject lines and email body gets better delivery. So it was more like, for example, our longer subject lines better than shorter subject lines? Not specific subject line is better or not than the another. Only high is longer better than the shorter one, or is more casual tone better than more formal tone? So we ran all these A-B tests around. You know, like how concise you should be, professional versus casual using where should you place a call-to-action pattern? Should it be in the beginning? Should it be in the end? How many call-to-action patterns should you place? Should you be focused on, for example, like a mission of your company vision attraction team, or about what right Should you like how much? Oh, we read so many A-B tests and this helped us to improve our conversion rates on every single part of the funnel.
Speaker 1:Yeah, and then did you find that you there's a certain number of emails, like if you sent out a sequence. Did you find you should be sending three or four emails instead of just one?
Speaker 2:Yeah, that's a great question. So we did experiment on that too and we found that four emails in total is an ideal number that balances. First, not annoying users right. If you're going to send them 10 emails, some people will get very angry. That's very important for us to make sure that users will not be annoyed. Second part is some people can see your email one and then they forget about it, even though they are interested. So following up, reminding them, usually is good not only for you but also for a prospective customer or user. And third thing is there is a time at which you know sending more emails doesn't give you anything Like. We experimented with sending fifths and sixths email, and those ones had almost no conversion. 50% of conversions happened when you sent out email one, and the rest of the conversions happened when you were sending mostly like email two, three or four. After that, almost no conversions were happening.
Speaker 1:Interesting and okay, so one final question here. What are some of the key lessons you've learned about managing bounce rates and avoid being flagged spam in large scale email campaigns?
Speaker 2:When it comes to bounce rates on a high level, you want to make sure that you are going to have bounce rate less than 5%. Anything about that is a signal for Google and Yahoo that you are collecting these email addresses in a way that most likely these emails will be called recipients. Most likely these emails will be cold recipients. Because here's the thing if people were giving you their emails, your bounce rate might actually be zero. But if you are, let's say, for example, I don't know like scraping these emails from somewhere and some of those emails are outdated, then Google starts to figure out that, because of bad bounce rates, these are cold recipients. And if Google thinks that you are sending out cold emails, they're going to automatically give you a less engagement rate, basically less deliverability, right? So one way to do this is to definitely use email verification tools. Make sure that bounce rates basically how many emails are not valid that rate is less than 5%. When it comes to complaint rates we had on a high level, your goal should be to have complaint rates less than 0.3%. How can you choose this? How can you achieve this? You can achieve this by on a high level is really going to go down to either optimizing your cold email marketing emails so much that complaint rate you get on cold emails are less than 0.3 percent.
Speaker 2:Our second way is to split emails you're going to send between warm and cold.
Speaker 2:So let me talk a little bit about the first one.
Speaker 2:Usually, when we test it, if you're going to send an unoptimized cold email, you're going to get around 0.5, 0.6% complaint rate. So this is a higher than 0.3% recommended complaint rate threshold, right? So what is important is to do lots of AAP tests to make sure that you can bring this down closer to 0.4%. It's too hard to bring it down closer to 0.3%. So how you can go down once you optimize your cold email, how can you go down from 0.4 to under 0.3? To do that, it's really I would highly recommend it to split number of total emails you send. For example, 100 emails will be sent to cold recipients and let's say 0.4% of those people will mark your email as a spam. And then another 100 emails will be sent to your own subscribers and in that case, let's say, zero people will mark your email as a spam. So in total, your combined complaint rate, average complaint rate went from 0.4% when you were just sending cold email to 0.2%, and so that way you can keep it under the 0.3% threshold.
Speaker 1:You really figured out how to work the system. I love that and, by the way, everybody he has created a product out of this. If you'd like to use this, you can get in touch with him, and so Archul. If people do want to get in touch with you or learn more about Genesis AI, how can they do that?
Speaker 2:Yes, so we do. Thank you, carissa, for giving us this opportunity so you can go to GenesisAIio. We have two products you can sign up there. One is for Marketplace, if you are interested in using some of our AI tools, and another is this email marketing tool. At some point, this email marketing tool will become part of the marketplace, but right now we are experimenting on this as a different sign up registration process tool and I highly encourage.
Speaker 2:if anyone wants to do like large-scale cold email marketing, I personally believe that it is literally the best tool out there for that this has been so informative and, honestly, this is my very last question.
Speaker 1:How do you pronounce your last name?
Speaker 2:oh yes, it's a chase really.
Speaker 1:Chase really Archil, chase, really Everybody. I feel like we just had a master class in email marketing. Thank you so, so much. It's been amazing, oh, thank you, God.
Speaker 2:I really enjoyed this.