The Exchanges

Q Nice. I love that. Let's talk about the first acquisition that you went through on Acquire. I'm sure as each acquisition went, you learned something. The process was easier the second time and the third time. That's usually how even just building startups go. Um, what was the first, uh, startup that got acquired?

A Yeah. So it's, it's quite a random idea. I mean, I actually developed it during my paternity leave. We were talking about our kids. At the beginning of this, uh, my, my little girl was born about a year ago. That was in summer, just as the pandemic was starting to fade away and travel restrictions were opening up. So everyone in the UK at least was trying to get their passports renewed because it all elapsed over the kind of two years in COVID because no one needed them. So there was this crazy rush for passports and no one could get appointments. And I realized that from done some basic web scraping stuff before that I could ping these appointment sites with a government passport service and find out When appointments are available, where they are and these kinds of things. So I quickly spun up a kind of quite simple consumer kind of SAS product. They would allow people basically to get alerts for when appointments came out for passport renewals. And this was in super high demand and it just took off. And at its peak, we were doing about a thousand plus dollars a day, just piling money into Google ads as well. Cause it's just a brilliant strategy, at least for that business hasn't always worked out by the ones, but It just really took off, but obviously it was quite seasonal. We managed to kind of capitalize on that and package it up in a nice enough way that it could be rolle…

AI assessment note: “simple consumer kind of SAS product. They would allow people basically to get alerts”

Q That's an efficient business. Congrats. Also a great use case too, just in general, helping people go see their families again, you know. That's awesome. Okay, so let's move on to the, what was the second startup that got acquired?

A Just after that first one, I got introduced to my current co-founder. A guy called Mike, and we started to bat a lot of ideas around. Initially we were there to meet up and talk about freelancing, but we never got onto speaking about that. We just got tons of ideas in our head and we're both very similar people. So we're just like, you know, screw freelancing. Let's just go and build out our own ideas and just run with a couple of things. So we started playing in the AI space when the GPT three AI APIs were coming out from open AI and it was all becoming a lot more accessible. And the first use case that we came to maybe questionable ethics, but it was a tool to help students write their university applications in the UK, which is called a UCAS statement. Um, and the kind of generic generative model was fairly average. And when we started to train the model on a bunch of examples, we realized that we could create this really powerful, quite unique tool to create these pretty bespoke and high quality, um, University applications. We actually wrestled with open AI a few times because we were like right in the gray area of like what's acceptable for doing with generative AI, at least when they, when they were coming out with it. Um, and we suddenly paused and thought, okay, maybe the bigger opportunity here is actually with the fine tuning because fine tuning an AI model is quite …

AI assessment note: “a tool to help students write their university applications in the UK”

Q street. I'm going to be potentially, you know, selling my business to you and you say X, Y, Z. I'd like to just verify that. Um, so just a tip there. Okay. So you, you found, you found the buyers and now you move into, uh, due diligence. How did you go through, um, Due diligence. How did you think about it? How did it work out with the buyers?

A Yeah, so I'm fortunate in that I've come from a background where I've raised money before, spoken to investors, had to put together data rooms for, for previous businesses, you know, data room for listeners who don't know, you know, essentially a collection of all your important documents as a business from your financials to pitch text to sales pipelines, whatever is relevant to you, but basically, you know, a bit of a dump from all your internal systems and knowledge repos. So I had some experience putting these things together and knowing what makes a good data room. Obviously these weren't as sophisticated as ones who were trying to raise, you know, much larger amounts of money. Um, but the core components of answering all the questions about how you, uh, how you're acquiring your customers, what are some of those economics look like in terms of the costs? If you're running Google ads, maybe some exports of some of the campaigns you've done there on Facebook, for instance, that kind of stuff. Obviously your financials, whether you're tracking that yourself or in my business's case, always Stripe. We always just had the latest Stripe extract. It was a little bit cleaned up and organized with some charts as well so buyers could just jump in and see all of our data and even play around with themselves. And we, we had to do a sharing link basically instead of giving them Stripe…

AI assessment note: “So I had some experience putting these things together and knowing what makes a good data room.”

Q looking for. Product market fit. You found it. Uh, you successfully went through due diligence three times. That's, you know, quite the feat. Um, what are, what are like your top three tips to people, you know, looking to sell, um, a startup? If they had nothing, they knew nothing about acquire, like what are maybe the three things you would say that you have to get these things right?

A Yeah. I think the first two points may contradict a little bit, but the first one is linked to my previous point, which is around timing is you need to have some kind of numerical objective traction that a buyer with no knowledge of your business or space or belief around the hype. Or vision that you have, or all the potential that you may see in your business. If you strip all that away, you need to be able to see some, some hard cold facts about the amount of money you've made, the growth of those over a period of, I would say at least kind of three to six months, but even, you know, in general, I think with your platform, you're kind of looking for, for the last 12 months is like a really important metric for businesses. So I think waiting to get to that point, um, is crucial and, and, you know, you can, If you list too early, I guess you can pull it down and you can list again. Um, but I think you'll benefit from kind of strengthening the business and thinking about, okay, what am I going to get challenged on here? Like I'm spending way too much on paid ads and the conversion's terrible. Like that's going to be a difficult sell when, when it comes to speaking with a buyer, if you can focus on improving a couple of those key areas and kind of beginning to think about how this thing is going to be packaged up and presented and answering those questions in advance. Um, then I …

AI assessment note: “you need to have some kind of numerical objective traction”

Q head in the right, right direction can, can help a lot. So I, I agree with you. Um, I, I guess my last, and thanks for all those tips. Those are really, really good tips that I, I agree with entirely. Um, so I, I gotta ask now you're, you're working on, um, a new startup. Um, tell me about it. What is it? What does it do? Um, yeah.

A So on the back of the fine tuning business, which was the last one that we sold, we started to learn about embeddings. Which is a kind of AI model or a technology that's provided by OpenAI and a bunch of the other big providers, and this came out just in the middle of when we were building this other business, and basically what it allows you to do is AI-powered semantic search, which is basically finding pieces of information that are similar to a question that you have or another piece of information, and where it's really useful, I'll kind of explain in a second, but this new technology came out as we were building our other business, and we just literally Took a step back and we're like, this is game changing in terms of what, what kind of capabilities you can allot for businesses and for customers with this. So that's why we actually sold the business because we just saw a bigger opportunity that was completely, I mean, related, I guess, in the AI space, but fairly different. Um, and that's what drove us to sell it, not, not really anything else. It was just like, okay, we want to pivot to this bigger opportunity. It's great that there are platforms like acquire.com where you can kind of put a pin in the business you've built. Sell it, get some seed cash, and bam, you know, you move on to the next one. So that was pretty cool. This new one now is called My Ask AI. We're ru…

AI assessment note: “This new one now is called My Ask AI. We're running it since about February”

Q position these with buyers to attract the most interest or maybe like some initial, and we'll go down like the timeline, um, from interest and due diligence to transfer. What are some of the things you think you did really right when you were speaking with buyers or positioning it when buyers reached out after signing NDA? Um, how did that initial phase go and how did you handle it?

A I think the first thing was that I definitely learned through to the three stages was the timing in terms of creating the listing. And by that, I mean, you know, how much traction did we have when we put the listing together? The, the UCAS, the university statement writer one was actually one that sold last, even though that was listed on the platform for maybe two or so months. And the main reason for that is when you compare it to the other ones, The traction was, it was a good proof of concept from a technology standpoint. Um, and we only really launched it and ran with it for a couple of weeks. The revenue was basically nothing, a couple of hundred dollars. So we felt like, hey, this is a good enough proof point for people that want to kind of take this tech and then use it in other applications. But looking back and comparing it to the other two sales, it's like, that doesn't really fill a buy with confidence that you've only run this thing for a couple of weeks and generated a few hundred dollars. Like that doesn't really prove That much. The buyer that ended up did buying it was really interested in the technology path. And so I think we got a little bit lucky with that side of things and they weren't really buying it for the kind of revenue generating like normal traction, uh, metrics that you would expect. So that's one thing, timing, you know, don't, don't put it on t…

AI assessment note: “timing, you know, don't, don't put it on there too early, get it”