A journey of more than a year and two launches on PH

Disclaimer, before you get scared about the length of this post, read this: I’m gonna divide this post into sections with titles so you can easily skip the ones you do not care about :D Today we’re launching Touring on Product Hunt again after 7 months, but this post is not about that. It’s about sharing our journey that has lasted more than a year, and hopefully give you something valuable for your future creations. The idea (1+ year ago) About a year ago, maybe even more, we started falling in love with the idea of an app that you could use everywhere you go and that would behave (almost) as if you had a professional tour guide always with you. Not just any tour guide, but a tour guide that knows a lot of things about virtually any place in the world. A few years ago, this would have been EXTREMELY difficult, but we felt that the new advances in genAI would soon make it more accessible as a problem space. The prototypes (11 months ago) After a quick validation of the concept/idea in our networks (family, friends, friends of friends, colleagues, …) we decided to quickly develop a simple prototype that was doing nothing more than asking chatgpt to tell something to a tourist about a place. In some cases, this was doing extremely well, but in most cases it was nothing comparable to a professional tour guide. How did we know? well, read the next section :P P.S. the first attempt was actually even simpler: we told chatgpt something along the lines “hey, we’re at these coordinates, tell me something”. Spoiler: did not work at all 😀 I guess we were too optimistic about the power of GPT3 The initial experiments (11 months ago) We decided it was time to test our idea a little bit more, without spending too much time. We wanted to validate two things: - Are people going to like an experience similar to a tour but without an in-person tour guide? --> We hired a tour guide, and we recruited groups of people in Milan. We then had the tour guide literally sit in our office and provide narrations to the groups about their surroundings, while they were able to move freely: they LOVED it! - Can GPT output something that it’s comparable to what a tour guide would say? --> We sampled a few of the locations where the groups had been, and then used our prototype to output some narrations. We then tested the narrations of the tour guide and of the prototype against each other (on online platforms such as Prolific, and within our network) on things such as length, quality, engagement. The result: surprisingly, for many narrations, the data suggested that the prototype was comparable to the one of the guide. The first developments (10 months ago) We had enough data to at least try and develop something more concrete that could be given to people to use during their trip. Our strategy turned out to be to progress with two things in parallel: - Some way to get this idea out into the world and see how people respond to it → a launch on Product Hunt seemed like the best fit (even though none of us had ever used PH before 😛) - Some concrete prototype app that we could then give to early adopters so that we could iterate towards a more refined product → we decided to use React Native and develop a cross-platform mobile app The first launch (8 months ago) A couple of months after, we had what we needed to start showing the world our idea. We quickly developed a landing page that explained the Touring idea, and we simply asked people to leave their email if they wanted to become early adopters. We also developed a simple simulator that let users “taste” a few pre-made narrations around the globe. The results of the launch? WAAAAAY better than we ever expected. A short summary (these numbers were taken a week after the launch): - more than **3000 unique visitors** to the website (from more than 100 countries) - organically featured in **18 different newsletters** around the globe - more than **600 signups** to become early adopters - **800 upvotes** and more than **150 comments** on PH - 2nd Product of the Day! (a funny off-topic on the first launch) I remember the day before the launch I was talking to my colleague, Simone, and we told each other: “I mean, if we get more than 50 signups it’d be great!”. Deep Dive and Develop (8 months ago —> today) After the successful first launch, we knew we had to dive deeper into development and refinement. We started gathering feedback from our 1000+ early adopters and doing user interviews, which was crucial in understanding the strengths and weaknesses of our prototype. The main things we focused on in these 8 months: - Improving the quality of the stories, moving from a simple interaction with GPT to a more complex RAG system - Optimising the generation pipeline to be able to generate stories anywhere, in seconds - Improving the UX and the UI of the mobile apps - Releasing the apps on the App Store and Play Store Today, second - bitter - launch (🚀🚀🚀🚀) And that brings us to today’s launch! Today, we are officially terminating the early adopters program we have been running for the past 8 months, and we are releasing our newest shiniest version of the Touring app on App Store and Play Store! You can download it right now and use it on your next trip 😀 🙈🙈 However, for some reason unknown to us, Product Hunt has decided to not feature us, and so we are very little expectations regarding the usefulness of today’s launch. The organic visibility is extremely low, because apparently Product Hunt decides (for unknown reasons to us) to put several products in an almost hidden category. I know, I know, I know. I said this post was not about todays’ launch. But still, if you found this interesting and you want to support what we’re doing with our lives (🙏), consider checking out and supporting Touring, or even share it with your friends that you know love travelling! What’s next? (Today —> until AI takes over 🤖) While it’s incredibly nice to see people downloading and using the app you have been developing, there’s a lot of challenges on the horizon. 1. Choosing the product direction: everyone travels, but there are so many different types of travellers. We have so many ideas about features and directions for Touring, but it’s proving extremely difficult to clearly understand which one (if there’s one) is the best direction to follow. 2. Validate Product Market Fit: Everyone talks about it right, but it’s damn hard! We made our assumptions, we made our MixPanel boards with all the analytics that we believe are important, but at the end of the day, we’re still trying to figure out if Touring has signals of PMF or not. One could say that it does, seeing the numbers from above, but PMF is not as simple as seeing a couple of numbers. 3. Monetization: as ugly as it may sound, you can’t live of upvotes 😅, so this is also something we are currently exploring and experimenting with. It’s not straightforward to understand which is the best approach to monetization for an app like Touring, and there are a lot of contrasting examples out there in the market. So for now, we want with our first approach (time-limited passes) and we’ll then experiment based on users’ responses. Bonus points! In this discussion I made in November after the first launch, I also wrote down a bullet list of “things we should have done”. It might give you some good suggestions! https://www.producthunt.com/discussions/it-s-been-a-week-since-the-launch-2nd-of-the-day-impressive-results-here-some-tips I’d love to hear from you if you found something useful in this discussion, or if you have some comments/suggestions on what we have done and we are currently doing with Touring!