We have decided to close down Soulie, the app where you can create your own discovery algorithm to find content online that is actually good for your brain. We truly believed in our big purpose of stopping the big social platforms from feeding their algorithms with our data just to keep us scrolling for longer to show us more ads.
It didn’t go as planned. However, we learned a lot on the way, earned the love of users worldwide and succeeded in understanding when it’s time to stop. It took us 2 years, €700k investors’ money, countless working hours of six talented people, mostly highs and sometimes lows. We hoped that we would figure it out sooner than later but this time, we didn’t. Here’s our story — we loved living it, and we hope you’ll love reading it.
To set the scene — when we started in early 2022, the world looked a bit different. Everybody’s attention was on crypto, Elon Musk had not yet bought Twitter, and people were leaving their houses again after the long battle with the COVID pandemic. Even though one pandemic was over, we knew another one was creeping — the pandemic of the mindless scroll.
The validation
When we first came to our grand vision — a world where everyone owns their personal algorithms and has control over what they’re shown on the internet — we didn’t rush into building the product. We first wanted to understand if people were even aware that they had a mindless scrolling problem, and if giving users personal algorithms to curate their own feed instead could be the answer. In September 2022 we spent €800 on Instagram ads that led people to a simple landing page introducing our idea where they could sign-up to a waitlist.
The results we got were stunning. In mere 3 weeks we got to 2000 waitlist signups from 25+ different countries. Every third visitor to our landing page joined the waitlist, and a further third of them also joined our Slack community where they would share how they hated that they were stuck with scrolling and didn’t get anything meaningful done. The majority of them were Gen Z from Europe.
The users who joined loved the concept of having a personal algorithm and even confirmed that this was exactly what was missing in their online lives. Encouraged by the visible results from real end users, we decided to go ahead building the app.
The app
Having a convincing story, we raised €500k from angel investors and got a grant of €112k from Enterprise Estonia. We started off by building an MVP where users could choose the topics they’d like to see content about, set up filters for their algorithm that they could later change any time, and get fresh content every day that we scraped for them from 400 sources we had picked out ourselves. The sources were free-to-read and text-based.
In November 2022 while our first prototype was in the hands of our test users, a drama unfolded. Elon Musk walked into Twitter headquarters with a sink in his arms and started firing people left and right. He had just bought the blue bird that he later turned into X. Everybody was furious, many ex-Twitter employees promised revenge and started to build their own versions of Twitter. We were approached by several investors who smelled the opportunity in the social media sphere but as we had just closed our round, we didn’t need more cash.
We too felt that the momentum was there. However, it had become harder to shine as new, experimental social platforms emerged like mushrooms after rain. And when Artifact, founded by the creators of Instagram, was launched in January 2023, we started to feel that the competition was getting serious.
But as we were 100% committed, it didn’t hold us back. It made us think even harder about how to create value to our users. Building the core feature of our product — a recommendation system that would use the users’ personal algorithm settings to recommend content — was more complicated than we first thought. Our initial plan was to launch in early 2023 but we ended up launching our first Beta in June 2023 to our waitlist of 8000 people.
Rapid experimentation
Hundreds of people from our waitlist started to download the app. The first part of the user journey still generated impressive results. Our Instagram ads were converting like they had been for over half a year, bringing us a steady influx of new users at 0.42€ per user. We were also impressed by the activation rate — 85% of people who signed up to the app went through the activation flow where they could create their own algorithm, and read an article that their personal algorithm brought them. But the harsh truth was that only 5% of them retained. We thought it was a cold start problem and that we just needed to keep on experimenting with different features, eventually leading us to the missing piece of how to keep people using Soulie.
We still had some cash in our bank account, although the end of the runway was not far. With the low retention rate we knew we wouldn’t attract VC money and we had to push really hard to get to better results. Besides, the ecosystem had gone through another massive change. After the launch of ChatGPT, everyone was now into AI and pitching a new social media app was a turnoff for investors. That got us thinking about bootstrapping.
The decision
By January 2024, we had improved Soulie’s user retention by 3x and decided to go all in. Enough of beta testing, it was time to make Soulie a subscription app and see if our target audience was willing to pay. We had just received an extra €100k from our investors and had submitted a research grant application for €900k.
We made a deal within our team that it was time to do everything in our power to get a powerful signal from the market. Based on the results we would get by the beginning of April 2024, we’d be making a decision if we should continue building Soulie or not. We defined three possible scenarios:
- If we get a convincing number of users to pay the subscription fee and a positive decision regarding the €900k grant, we would definitely continue;
- If we didn’t get either, we would close down Soulie;
- If one of the conditions was met and the other wasn’t, we would use common sense to decide if it was reasonable to continue.
Our team put their everything into the next two months to launch the full-fledged Soulie app that supports subscriptions. We launched it in the beginning of March 2024 and within the first weeks we got over 700 new sign-ups, out of whom 90% started a 7-day free trial and 43 started paying, taking our MRR to €200. And then we got the negative decision on our grant application.
Even though some users started to pay, the signal was weak. The retention was still poor, and there was a notable churn of paid users. We reacted quickly, making changes to our paywall and released a few new features, but we ourselves were not convinced anymore. With the negative decision on our grant application and weak signal from the market, the decision had been made.
The learnings
- We were in beta for too long and didn’t launch before we absolutely had to. That took valuable time away from testing Soulie in real life.
- We fell into the feature creep trap. Our initial idea was to develop a recommendation system that would be able to serve content to users that they preferred from all sorts of sources (news, social media, blogs, RSS feeds, podcasts etc). However, a lot of the sources we considered had closed access to their API by third parties and the algorithm itself required a certain amount of training by the users (reading, reacting, giving feedback) to deliver precise recommendations. Most of the time, the first feed users were served on Soulie was not convincing enough, and it’s all about the first experience that makes people come back. Therefore, we started to take feedback from users on what would be other features (gamification, sharing with friends, commenting etc) we could build, and that took our attention away from the core — building a solid rock recommendation system that was based on the users personal algorithm that they could change any time. We don’t know for sure if it would have changed the journey for us if we had stuck purely on the original idea but looking at hindsight, it seems that the more features we added, the more features users started to ask for.
- Building a consumer app and getting it off the ground is really hard. Especially when the app’s essence is about making people change their scrolling habits on social media. You need to have a huge marketing budget or a really addictive product. We had neither.
- Investors go where the buzz is. When we started, we were pretty much on our own and crypto was the main talk of the town. When we had just raised money and Elon Musk bought Twitter a lot of investors suddenly got interested in the social media space. A year later it was all about AI and our industry was no longer exciting.
- Timing is everything. It doesn’t matter if you have a top-notch team; if there is not enough business opportunity in the market, there are also low chances to build momentum for your startup. That was also confirmed by the fact that in January 2024 Artifact, our closest competitor, came out with the news that they were closing down due to lack of market opportunity, and hence low chances of becoming the next decacorn. They were acquired by Yahoo a few months later.
All in all, working with our amazing team on a very challenging problem was pure joy, and we learned a lot in those few years.
And from the very first second of building Soulie, we had the support of the community — thousands of people from all over the world who believed in our mission and wanted to see us succeed joined us on Discord and participated in building Soulie. We are so proud that so many community members have been with us ever since we tentatively set up our first Slack community in September 2022, wondering if anyone would join. Thank you — we sincerely could not have done it without you.
Lastly, our investors. Your continued trust in our team has meant the world to us.
What’s next?
We’re keeping Soulie on the App Store and Google Play until 1st of July. It’s free to use until we shut it down, so do go explore!
And if anyone in the media or recsys industry is interested in learning more about our product and technology itself, feel free to reach out to me personally at triin@soulie.io, or the team at hello@soulie.io. We are more than happy to contribute if Soulie could help others on the same mission.
Triin Kask
CEO & Co-Founder of Soulie
on behalf of team Soulie
