Peter chats with Frank Sondors about how his company Salesforge is making personalised email outreach faster thanks to AI.
In this interesting interview, Frank discusses the problem with traditional sales templates and cold email outreach (which is very rarely personalised), as well as how he is tackling this issue for businesses.
Frank and his two co-founders are bootstrapping. They launched their product two months ago and have so far gained 80 users and $2K MRR. He talks about how to make your email campaigns more effective and also shares his journey so far with Salesforge.
[00:00:00.090] – Peter
Frank Sondors from Salesforge.ai. Hey Frank, thanks for joining me. Can you introduce yourself? Tell us a bit about who you are and what you do?
[00:00:08.610] – Frank
Yeah, so I’ve been in B2B sales for well over a decade. Been in frontline sales doing a bunch of cold calling. I sent probably a few million cold emails in my life and then I led a sales team last year of 50 people. So a bunch of sales, development reps, account ecs, et cetera. And I realised that there’s a huge problem in the world of sales, as in folks just send a bunch of templates out there.
[00:00:33.470] – Frank
And the problem with that is that users just have generally a fatigue with looking at templates. It’s just like banner ads on the website. So when you’re scrolling through a blog, there’s a bunch of ads, you just don’t look at them. And the same with the inbox where you can just see it’s a template. And because of that, most folks do not reply and say cold email is dead unless the offer is really good. This is when you actually reply, right? However, most companies don’t necessarily have a super good offer. A lot of companies are selling like a nice to have solution and you really need to think about how you personalise every single email.
[00:01:09.020] – Frank
But the problem is when you scale, you can’t really ask your reps to personalise every email. So this is why folks do end up preparing templates. So yeah, I build a software that allows to craft unique emails at scale, actually not through humans, but through AI, essentially. We’ve been working on this product since the beginning of the year. We went live about a couple of months ago. Currently already have over 80 companies on the platform over 2K in MRR. And yeah, that’s kind of where we are.
[00:01:43.390] – Frank
And we are gradually sort of scaling in the beginning, just naturally all the founders know you need to do things that don’t scale. So for us, we spend a lot of time on, I don’t know, Reddit various sort of WhatsApp groups that I’m in. Let’s say sales communities, slack communities, where else? Like different essentially places, but ultimately doesn’t scale, including LinkedIn, actually. So founders have to really figure out how they will acquire systematically users and they have to usually choose one channel that will really work for them. So most companies or most B2B sales companies in B2C choose usually SEO versus the PPC route, let’s say, or social media.
[00:02:22.030] – Frank
However, when you do B2B, actually the channel that I would argue as well, which is the reason why I built the software in email for now, is called email, because actually sending emails doesn’t cost a lot of money. However, it is a very complicated topic because you land very frequently in spam because Google doesn’t like when you send a lot of emails. And then the problem is. Yeah, I mean it doesn’t work, let’s say, when you send a bunch of templates. So there’s many, many challenges. It’s very complex.
[00:02:47.280] – Frank
So we decided to tackle that problem and help startup founders, maybe big companies, agencies to figure out how to reach out to prospects in a way that it actually works. And arguably from a profitability standpoint, we are significantly more profitable than giving money to Google Ads or Facebook ads.
[00:03:06.970] – Peter
Hey Frank, that’s awesome. I love the way you come from a sales background. You know the problem very well. So you’ve built a tool specifically to deal with the challenges of outbound email. So you’ve launched this product, you bootstrapped the business so far, right?
[00:03:21.410] – Frank
Yeah, correct. Yeah. I guess the problem for everybody, especially right now during the downturn is none of the VCs really want to give you money pre revenue unless you’re competing against OpenAI. Just like this one company in France that raised a bunch of money. So we are like in the red ocean space, right? There’s a lot of competitors. However, all the competitors are sending templates. So we ultimately compete against templates. And so this is our differentiation in the market.
[00:03:52.090] – Frank
And naturally you see that from a perspective of that, it drives significantly high reply rates. So customers are seeing net positive impact. And by sending actually unique content in emails rather than templates, it helps you with your email deliverability because sending static content, google and Microsoft, they can see that you’re sending exactly the same content at scale. And this is where you gradually, gradually your email deliverability deteriorates.
[00:04:16.220] – Peter
Okay, so that’s cool. It sounds like you’re moving away from templates more into personalisation, more into AI, more into tactics that make cold email have an impact or improve the deliverability and the replies. Do you have any co founders or you a solo founder for Salesforge?
[00:04:33.010] – Frank
Yeah, I have two co founders. So initially the idea was mine, naturally. However, I met my co founders actually, funny enough, on Y Combinator a matchmaking platform. I thought to myself, I don’t know anybody in my network. I have a bunch of salespeople that I know or a lot of prospects who are on LinkedIn, but I barely know any engineers, engineers that could handle me, sort of. I’m quite active. But yeah, funny enough. So there’s actually two of them. They know each other from the first grade. They went to the same uni. They built a few apps but nothing really financially really took off. So I was like the missing commercial piece and they didn’t have a really good idea actually, what is it that they would love to build.
[00:05:12.880] – Frank
So they was they actually messaged me on Y Combinator saying, hey, I really like what you do, let’s have a chat about this. So we did, they’re like, oh yes, I’m ready to start coding tonight. And I said, no, wait a second, first of all we’re going to do some user interviews. So did the right way. Going to try and figure out, scope this out properly, etc.
[00:05:32.640] – Frank
So we kind of paused this whole development for a couple of months. So then I went through a pre-seed accelerator called Antler, actually flew over to Berlin for that. Spent a bunch of time with a cohort of 40, 50 people. Kind of gives you that ability to focus. It’s super important, especially when you’re a father and have two kids. So you need that focus. So you need to kind of fly somewhere else, be in the same sort of space with other sort of startup wannabes and really try and either hook up with one of them. So search for a technical co-founder in my case or just be in that zone where we’re all trying to get our kind of baby off the ground.
[00:06:13.510] – Peter
Yeah, around that energy and that focus and everyone the effect and the influence of everyone else rubs off on you too. It’s good to be in that space. Gets you productive, right? So you’ve just launched two months ago, you already have 2K MRR, so that’s not a bad start. And 80 users, are these 80 paying users or what’s your model like? Is it freemium with the kind of upgrade plans?
[00:06:39.910] – Frank
Good question. So the thing in our case is because what we do is we compute emails in real time computation and data fetching cost a lot of money. So we currently don’t do trials because of that. Because if I would do a trial, the problem is it’s going to cost us a lot, especially being bootstrapped. So unless we raise, we can’t give a trial where we can give you a good amount of consumption to kind of figure it out.
[00:07:03.940] – Frank
However, we will do trials probably by September where we will give you the ability to compute say ten emails just to give you an idea how this actually works. But really if somebody is looking to see a difference between let’s say, how they’re currently doing versus a new way of doing, say with Salesforge, they really have to scale it up. So at least sort of reach out to about 1000 leads. And that’s a lot of money for us to service trials in that way. So we’re just asking for founders to get a one month subscription as sort of a test because we don’t do annual contracts. There’s no point. We believe that the proofs in the pudding.
[00:07:42.860] – Frank
Once you see after the first campaign that your results are significantly better than anything that you’ve done, you’re going to stick around anyway because our current churn rate is zero with the product. So it gives us confidence that most folks will see value. Of course, somebody will churn in the future, but still that majority are not churning, they are seeing value, they are sticking around and we are sort of hashing out various sort of features or bugs along the way.
[00:08:07.930] – Frank
Yeah. It’s important to provide support in the early days to every single user that you on board, not just in the way of kind of setting up automated messages and no, actually proactively checking in with every single user in a channel that’s comfortable for them. So, for example, I’m checking with one user on LinkedIn with the second one on WhatsApp then Slack community that we have various different ways and I’m trying to make sure that they are super successful because I want my retention rate to be as close to 100% as possible in the first sort of three to six months.
[00:08:36.950] – Peter
Yeah. Don’t you think since you’ve only launched for two months, your churn rate might be low now because you haven’t got enough time.
[00:08:42.790] – Frank
Exactly. But you do expect some folks after the first month to churn already, so that hasn’t happened yet. So that’s a positive signal for us. Usually there’s a bit of a lag. Right. So you need to wait for another few months. Yeah.
[00:08:55.520] – Peter
And then do you think maybe once you’ve validated a little bit the stickiness of the product, maybe introducing annual plans with something like a 20% discount might be a good move for you to just increase retention or LTV?
[00:09:07.790] – Frank
Exactly. So we already provide 20% discount. Nobody goes for that because it’s still a relatively new concept, as in, hey, instead of sending templates now, I’m going to send unique emails at scale using AI. And people are still sceptical about Chat GPT because they had bad experience trying to write up stuff in Chat GPT. Whereas in this case you lose the control, as in what’s in the actual message? AI takes over. Yes, you can proofread one by one. But guess what I do? I just load 1000 leads into the software and within two minutes, with a click of a button, it’s all personalised, it all goes out.
[00:09:47.710] – Frank
And that’s efficiency that I’m able to provide to a startup found that they can’t afford employees, but still needs to scale 10x somehow in the pipeline. And I know that Google Ads and Facebook Ads are just too expensive. So this is why email is still the cheapest form of acquisition, in my opinion, not to hire folks.
[00:10:08.120] – Peter
So, Frank, when you say you’re setting up a campaign in Salesforge, what content do you need to put into your campaign to enable the AI to do the rest of it? So I assume you’ve got your list. You also need to have an offer.
[00:10:22.150] – Frank
Exactly. So there’s two things for AI to do a really good job. So the problem is that other solutions haven’t engineered really, let’s call it logic, what should be the input data. So that means the context that you provide to the eye, the constraints that you provide, the goal that you provide to the AI. So let’s say from a goal standpoint, we have engineered essentially a way for the AI to drive the maximum possible reply. So the way that practically it looks like is that it writes short and snappy emails with soft CTAs in it and it does that at scale, right? Because over 60% of emails are being read on a mobile device.
[00:11:01.800] – Frank
Now, when it comes to the input data for AI, we need two data sets. One is what is called the seller data. So meaning all information about you. So you can set up multiple products or the same products with different value props. And we ask you questions, very simple questions that you should know by heart, which is what’s the main problem that you’re tackling for your prospects? What’s the solution that you have for that particular problem?
[00:11:22.110] – Frank
What’s the cost of inaction if you don’t decide to go ahead with the product? And that’s enough information for AI to do its magic. And then you still need the buyer data, so the prospects, all the information about them. So this is where when we ask for you to upload the list onto our software via CSV, this is where we ask for the LinkedIn URL of the person. So this is where we go on LinkedIn. We fetch all the information, the headline, the about section, the current title, the job description, what’s the post maybe that they’ve written last 30 days. And then we feed those two data sets into AI to craft these unique emails.
[00:11:55.580] – Frank
So in the future we’ll be also able to do websites, Instagram profiles, Google My Business profiles, anything that’s really structured. So we call this really if you think about programmatic advertising, this is essentially programmatic outbound sequencing what we are designing. And in the future we’ll just add a couple more channels. So for example, LinkedIn for you to be able to use the same AI engine over to LinkedIn. And then we have some interesting things coming over to also have the ability to dial prospects as well, right?
[00:12:21.250] – Frank
So having a natively inbuilt dialer and because these are the three core channels of B2B acquisition.
[00:12:29.540] – Peter
Sounds fascinating. So it sounds like you’re improving deliverability and response rate, but you’re also speeding up the process, right? Because all of this AI is kind of gathering the content, writing the emails. So I imagine like one of your value props is getting doing more emails in less time, is that right?
[00:12:46.220] – Frank
Correct. Yes. So we have also built out various email deliverability features to essentially help you stay out of spam. So, just so you know, one simple example is we natively validate the email addresses before you would actually send anything to those folks. Because if your bounce rate is over 2%, Google and Microsoft tracks that. And then what they do is they use that as a variable to send you more into a spam folder. So we thought to ourselves, we have to have that as a basic requirement within the app. And then we do other things as well within the app to kind of boost the email deliverability. But on average right now we’re seeing anywhere between 70% to 90% open rate across the client base, which is pretty good.
[00:13:31.090] – Frank
Let’s say if somebody’s using something simple like HubSpot as an example, they’re usually seeing anywhere between 30% to 50% open rate, just to give you a comparison with the CRM sort of native sequencing there. And by having double the open rate ultimately will mean double the replies, double the sales output. Right. And then when you kind of fuse that with the AI capabilities by writing unique emails to everybody, that in itself doubles the reply rate in comparison to using templates.
[00:14:02.200] – Frank
So overall the sales output massively scales and you can actually add multiple inboxes into the software because Google and Microsoft, they don’t like when you say send more than 50 emails per inbox per day. So the only really solution is to have multiple inboxes. Kind of trying to figure out a way, but this is what allows you to scale as well.
[00:14:23.360] – Peter
Yeah. Awesome. Hey Frank, I’ve really enjoyed talking to you. It’s awesome hearing about Salesforge things. Cool tool there. Look forward to seeing what you do with it. How can people follow you and try out the product?
[00:14:34.050] – Frank
Sure, I mean they can just connect on LinkedIn. I’m super active on there. And then you can go on the website Salesforge.ai check out the product. If you have any questions, we have the Chat app. Just ask me questions. I’m always on Chat or any other channel, so I’m probably everywhere.
[00:14:51.360] – Peter
Awesome. Okay. Thanks, Frank. Okay, so if you’re listening, check out Salesforge.ai. Thanks for chatting!
Need a hand?
At UserActive, we help SaaS founders create meaningful products users love.
Click here to book a no-obligation call to see if we’re a good fit to help you and your team!