Update:
So I have been working hard for the past 2 weeks on my little project. There's been quite a learning curve, but I feel like I have turned a corner and my goal for the next 30 days is to generate $10k in total revenue.
Current setup:
145 warmed up email mailboxes (30x day sending capacity)
1 main "test" campaign (5k leads)
Lead scraping + enriching set up
2 cold email strategies (podcast vs direct)
Custom AI SDR reply bot (email based)
Custom AI chatbot demo (with customer's branding)
Upcoming:
150 + more mailboxes (approx 30-50 domains) to be added over the next week
Create auto-generated demos for ANY industry
Create bulk email verifier
Okay... a lot to unpack here, I will touch on all of these and get your feedback.
145 (somewhat) warmed up email mailboxes (30x day sending capacity)
These have been set up, warmed up (mostly), and are currently going through the sequence. I have it set up now to do 2 follow up emails spaced 1 day apart. So cold email Monday, follow up Tuesday and Wednesday. I did it this way to get faster feedback. Yes I realize spacing them out would be better. I wanted to see if follow ups wroked at all. They kind of do.
Roughly a 1% reply rate on all emails. A little more but let's call it 1%. That means 1% on cold, 1% on first follow up, 1% on second follow up.
This is prob not hte optimal strategy, I have a feeling that more initial emails and less follow ups ist he way to go. But I only had a small test batch of leads. Anyway will tweak this later.
This is roughly around 40 or so domains. Some domains have 3 emails, some have 5. I would say most of them are around halfway done wiht warm up, but I won't hit full sending capacity on all of them for around another 3 weeks. I only tested with around 8 mailboxes at first and let the rest warm up, but recently just started sending with the non warmed up emails but only like 1-2 a day and will scale to 30 over the course of a month.
1 main "test" campaign (5k leads)
These were the leads I scraped as I was trying to find a way to get leads for cheap. I scraped a few niches just to test and see which would be more responsive. I think I scraped dentists, business funding, solar companies, maybe 1-2 more.
Got some responses that ran the gamut from "no thanks" to "let's set up a call".
One thing I didn't realize is that with Instantly, when you set the max emails per mailbox that includes follow ups and first contact. I thought it only counted towards first contact and that follow ups didn't "count" towards the daily total, but apparently that's wrong. So with 2 follow ups, that means essentially after 2 days at full capacity, each mailbox will only be contacting 10 new leads per day. Even with 145 mailboxes that seems low at only 1450 new ppl contacted per day.
Moving forward I will run 2 strategies: general and specific. WIll talk more about this later, but basically I will have 1 front that's like an all purpose agency thing. No niche. And then I will have other specific campaigns that will be for things like insurance, home services, solar, blah blah blah.
Lead scraping + enriching set up
Leads are expensive, specifically getting them and verifying the emails. To get the leads, I use the apify scraper for Apollo to get up to 50k leads for 1.20 per 1000 leads. It's very customizable you can get whatever you want. Once I get the leads, I run them through chatgpt to filter for the fields I want. Then I run it through an email verification service since Instantly's is overpriced as all hell.
Moving forward, I want to create my own bulk email verification tool so I don't have to pay these companies exhorbitant prices for verifying emails. What a joke.
That is on the list but I think it's prob better to work on other stuff before I get that.
2 cold email strategies (podcast vs direct)
This is a trick I learned from someone else in a similar space. Basically you email your customers an invitation to be on a podcast where you interview local business owners. Many more business owners will say yes to a podcast but no to a sales call. So it's an easy way to get them on the phone.
Overall from my campaign I got I'd say maybe around 10-15 replies of people that I could have gotten on, but I didn't put enough effort into reeling them in. I had one guy book a call straight away after I set up my SDR auto reply bot, but I flaked on the call. Dumb but I was working on other stuff.
Moving forward I am going to test both. I get the appeal of the podcast idea, but I like the direct approach too. More qualified leads plus you don't spring a pitch on them later.
With the podcast, you basically are supposed to record the entire call and then feed the trnascript into a specialized prompt that will give you a custom plan of AI services that can be applied to that company. Generally it's either going to be one of a few things like database reactivation, lead gen, chatbots, etc. But you give them a nice big report with everything you're going to do, strategy, all that shit. Like "oh hey great podcast the other day, by the way I came up with a plan to fix your lead gen problem, let's set up a time to go over it."
Like I get it... but I dunno.. I don't like having to hide the fact that I don't care about their dentist office in Tennessee and I just want to sell them something.
Custom AI SDR reply bot (email based)
THIS is what has been eating most of my time for the past week. Somehow I got the idea in my head that I didn't want to reply manually to all the leads that were coming in, but that I would rather set up an AI SDR reply bot that would read the emails and reply in a way that would get them to book an appointment.
Dumbed down version:
First email (hardcoded): hey, do you want to be on a podcast?
Lead reply: maybe, send me more info
SDR: (reads the reply, sends something custom w/calendly link)
This took me WAY too long and I will tell you why right now. To start, I used N8N to make the basic reply bot. Now I do not know shit about programming really or modular workflows, but I know how to use chatgpt and youtube. So I pieced together something that works, it's actually pretty sophisticated. Here is the workflow in case you are a nerd:
I won't explain every step but basically:
Customer replies to cold email
Lots of data sent to bot (email, name, message, etc)
Bot writes a reply based on the message + data
Email is sent back to lead
Sounds like it should be way easier than it was, and I'm sure my noob status as a developer means it's not as efficient as it could be. But it works finally and does what I want.
So bottom line is now I don't have to worry about follow up, now I can just focus on cranking the cold emails out and the SDR will handle all the replies. Theoretically that's how it works, but I'm sure it will need more refining.
Custom AI chatbot demo (with customer's branding)
Ok.. I am proud of my AI SDR thing, but this is def the crown jewel so far. Now, not only will the SDR reply to the email to book an appointment, but NOW I can get it to send out a customized demo chatbot to the lead on working domain that has all the lead's branding.
For example, if I message Bob from Super Solar Company asking if he is interested in a chatbot that generates leads, and he replies with anything positive, now the bot will send over a customized AI chatbot. I will show you a screenshot.
Here is an example from a test site that I made. It needs some refinement for sure, but the core functionality is there. It says Capitalworks Funding because the test leads I was using were from business funding, but the bot/frontend was for solar. So disregard that. Oh and instead of solar bookers it's supposed to have the lead's company name at the top. So I need to change that.
The bottom line is that these get sent out with the leads name, their company name, and a working iphone display so they can see how it would actually work. I should prob add something on there about how the process works or whatever. But you get the idea. It's 98% there. All of this is hosted on a vercel server and was coded with Cursor + Claude.
I plan on refining this even more in the future, but for now it's fine.
So as you can see it's moving right along.. I need to focus more on sales and less on this backend shit. It gets addictive though. Will post an update in a week or so. LMK if you have questions.