Wengo Travel
I built Wengo Travel, a web app to check actual past weather for trip planning, not just averages.
It came from a problem I had experienced myself. Whenever I was planning a holiday, I would Google something like “weather in Barcelona in May”, and every site would just show monthly averages. But averages only tell you so much.
I wanted to see the actual daily and hourly weather from past years to get a feel for what it might actually be like when I’m there, day to day. It seemed crazy that I couldn’t find anything like this. I thought surely I could build it.
I know what you are thinking. A weather website? It sounds more like a toy project than an exciting startup.
That was part of what interested me so much.
I’m a big fan of Paul Graham’s essays and I thought this was the kind of toy-like, solve-your-own-problem idea that he says, counterintuitively, often ends up becoming a great startup idea.
“finding the problem intolerable and feeling it must be possible to solve it. Simple as it seems, that's the recipe for a lot of startup ideas.” - Ideas for Startups, Paul Graham
“the very best startup ideas tend to sound so lame, at first, that you'd reject them if you were consciously looking for startup ideas” - How to Earn a Billion Dollars, Paul Graham
“Just fix things that seem broken, regardless of whether it seems like the problem is important enough to build a company on. If you keep pursuing such threads it would be hard not to end up making something of value to a lot of people, and when you do, surprise, you've got a company. Don't be discouraged if what you produce initially is something other people dismiss as a toy. In fact, that's a good sign. That's probably why everyone else has been overlooking the idea.” - Organic Startup Ideas, Paul Graham
Now, I’m not saying that I thought this was likely to turn into an amazing startup idea. I knew the odds. I just thought it was a good plan to follow my intolerance to this personal problem and my curiosity as to why it’s not solved yet. Who knows where it could lead? At the very least, I had a hunch that people cared a lot about weather when planning travel, and that the existing tools could be much better.
I had recently moved to Sydney from Ireland and I decided to give myself some time to explore the problem as a bootstrapped solo founder, inspired by the successes of Indie Hackers such as Pieter Levels.
I wanted to approach it using the principles of The Lean Startup, which is essentially about avoiding one of the most common mistakes made by technical founders: investing time building something nobody wants.
I broke the idea into my two main assumptions.
The Value Assumption was that people wanted to see actual daily and hourly weather from previous years, rather than only monthly averages.
The Growth Assumption was that people would search for phrases such as “weather in Barcelona in May”, or ask an LLM a similar question, and discover Wengo Travel through SEO or AEO.
If those assumptions proved true, the site might eventually make money through advertising or evolve into something larger. But my initial priority was to follow my curiosity and learn as much as possible about building a product people wanted.
Even if it never became a successful business, I thought it would be a worthwhile experiment.
I built the MVP quickly and wanted to get it into the hands of users as soon as possible. I wanted to “validate” my value assumption by observing user behaviour. I ran Google search ads for a few destinations and months to get quick continuous user feedback from people actively trying to solve the problem. I also later replied to weather-destination questions posted on Reddit to try a different source of traffic.
I ran several iterations using a simple loop:
- Run ads.
- Analyse user behaviour.
- Update the product.
- Repeat.
Some of the behaviour was negative. People bounced or spent very little time on the site. Other behaviour was more encouraging, particularly as I improved the product. Some visitors spent a significant amount of time on the site, returned, searched for multiple destinations and months, browsed the daily and hourly data for different years, and answered “Definitely” to my pop up survey asking, “Hey, is this information helpful for planning your trip?”.
So had I validated my value assumption?
I realised that it was difficult to validate something like this in a clean, definitive sense. It’s not very realistic to set some thresholds for the engagement metrics, run an experiment and determine your assumption is true or false. In reality, you can keep improving the product and potentially improve the metrics. Maybe the UI was poorly designed or the product positioning wasn’t the best. But how long do you keep doing this for?
It felt possible either to give up too early or to stay improving the product forever.
I decided the more useful question was not whether I had “validated” the value proposition once and for all. It was whether the metrics were moving towards the values required for a sustainable business at a fast enough pace.
At the time, it seemed that, in order to continue, I had to start thinking of this more as a business venture and less as a product I would pursue irrespective of commercial success. I couldn’t afford to work on it full-time improving the product indefinitely without signs of commercial success.
It seemed I needed to test the growth hypothesis to determine if this was a feasible business opportunity. If people loved the product, would they be able to discover it in a sustainable way?
I did some SEO opportunity analysis. There appeared to be relatively low keyword difficulty and high cumulative search volume across searches for different destinations and months. In theory, there was a chance my pages could rank on Google and collectively get millions of monthly visitors a month. I wanted to test this.
I started optimising a few pages on my site and building a programmatic SEO pipeline. Some pages were indexed, moved up the rankings and got a few impressions. But they were not at the level of rankings or impressions needed to get clicks yet.
I quickly learned that I wasn’t going to be able to test the feasibility of ranking at the top of Google’s results for all my pages so simply. It is not really something you can test for a few pages and then scale. Ranking and impressions are dependent on your whole site, not just the individual pages. They are also dependent on factors outside of just the quality of the page content. Finally, the change in rankings and impressions, as a result of site updates, can lag by months.
Instead of running a quick experiment to indicate whether it was realistic for my pages to rank, I would need to consistently publish new pages, improve old ones, acquire backlinks and wait for those efforts to compound over time.
Testing my Growth Assumption would be a longer process than I had initially anticipated. I still wasn’t any more certain about my chances of ranking at the top of Google if I made the time investment.
I made some rough projections to try and decide whether that investment could be justified. I got monthly search volume from Google Search Console, estimated conversion rates using Ahrefs data, and estimated the RPM I might expect from display ads. But the resulting range of revenue projected was very wide and dependent on many rough estimates.
Overall, I felt that as a simple SEO content site, revenue could be anywhere from nice additional income to that of a “lifestyle” business. More interesting was the potential for it to transform into something bigger.
Genius is an example of what started as a simple SEO content site and became a YC-backed startup. The founders and investors hoped it would become an annotation and knowledge platform for the entire internet. The bet unfortunately didn’t pay off. Genius was eventually acquired for $80 million after raising roughly $76–78 million. Despite its particular big bet not paying off, it did still go from a simple SEO content site to a platform acquired for $80 million.
Despite the potential upsides, I felt there wasn’t enough evidence to be investing in Wengo Travel full time like I had been. I had committed to Wengo as a full-time startup before I truly believed it deserved that level of investment. I would not have left a job to pursue it, I had simply decided to give it some time before finding one, which amounted to much the same thing.
That created pressure for Wengo to prove itself as a business before I had finished exploring it as a product. Instead of focusing on making something a few people loved, I was constantly evaluating whether it was a worthy investment and trying to make projections I could not reliably make. Ironically, I felt treating it as a full-time venture actually slowed down progress.
This also wasn’t the only problem. Prolonged isolation was also affecting my progress. Throughout Wengo, I was working alone in a country I had only recently moved to, without the peer network of fellow builders I previously had around me. Initially, this was not a problem, I had always worked well independently and began the project at my usual pace.
What I underestimated was the cumulative effect of months of work in isolation. Even when you are working in highly autonomous environments, like studying at university, working at a startup, or working with a cofounder, you still have some shared urgency, feedback and discussions. I realised that even if I am doing the vast majority of the work alone, a little bit of engagement with peers helps a lot in terms of both speed and enjoyment.
With that said, I reached two conclusions. First, Wengo Travel should continue as a side project, not a startup idea worthy of my full-time investment. Second, I wanted to get back to the kind of environment where I do my best work.
So, that is what I am looking for now:
A team of ambitious people, working on something worth caring about, and the opportunity to build something great together.