A Postmortem on 2021
A lot happened in 2021, and in the spirit of building in public, I thought it would be a good idea to write up a postmortem for the year. To be clear, I am using the word "postmortem" over something like "review" purposefully. This was a year of learning for me, and I've found that I learn best when being honest with myself about what worked and where I fell short.
A little accountability and humility doesn't hurt either, hence this post. It's probably going to be a long one.
I tried and failed to launch four products this year. I'll take you briefly through each one and summarize what I learned afterward.
Partake - failed project #1
The first product I tried launching this year was Partake, a marijuana ordering and delivery app. The plan was to start with the state of Michigan and branch out over time.
The big mistake here was obvious in retrospect - the very first thing I did when I came up with the idea for Partake was to start writing React Native code.
After a few weeks of development, I started to realize how big of a scope this idea was. Since raising VC money isn't something I plan on doing, I moved on to the next idea.
Ruminate - failed project #2
I've been using Basecamp professionally for about four years now, and one of the things I love most about it is how it handles asynchronous communication. So I thought - why not unbundle some version of that functionality and sell it as a service?
This time, instead of immediately starting to write code, I did a decent amount of market research. I found several competitors in the space, which was a good sign.
Before I wrote any code, I realized that pulling companies onto an entirely new platform was going to be challenging, so I pivoted from a platform to a Slack app. I pivoted a few more times to different ideas, like formal RFCs, for reasons I can't remember right now. This project became way too tangled, so I moved on.
Dub - failed project #3
The idea here was simple - host your podcast from Notion. Again, I did a fair amount of market research before starting to code anything. Podcasts are popular and growing, and other apps built on top of Notion seemed to be enjoying some success.
I built the whole thing, start to finish. I built a sync system using Lambda functions to keep Notion pages and databases up to date with Dub's database. I built a podcast RSS feed generator. I integrated with Stripe and Cloudflare.
About halfway through, I realized that I didn't know how to market Dub. I didn't know where podcasters hung out, and I didn't really know anyone in the industry even if I did. This is because I'm not a podcaster. I'll talk more broadly about this in a minute.
Storyform - failed project #4
Towards the end of developing Dub, I was granted access to OpenAI's GPT-3 API. I wanted to try and build something that would be quick to build and validate, so I landed on a tool to generate marketing copy. I wanted it to be a cheaper but less featureful version of already-existing tools in the space.
Again, I built the whole thing. This time, I started with the design in Figma before writing any code. This was the first time I'd done anything like that. It's up right now at https://storyform.ai. I'll probably move it to a different domain and use it for my portfolio or something eventually.
I ran Google Ads using keywords I got from Ahrefs. Over the course of my ad campaign (nine non-consecutive days), I got 97 signups and zero paid users. I spent $157.
Similar to Dub, I eventually realized that I had no idea how to market this product.
What went wrong?
Each of these ideas failed for the same reason: I knew very little about the industries and domains I was attempting to build products for. For example, with Dub, I'm not a podcaster and I know very little about the podcasting process.
This logically lead to a decision flow in my mind (though I wasn't aware of this at the time). It goes something like this:
- Am I knowledgeable in the domain that the businesses I'm serving are in?
2a. If yes, do I care enough about the problem to work on it for the next X years?
2b. If no, do I care enough about the problem to become knowledgeable enough to run a businesses in that domain?
If the answer to either 2a or 2b is yes, then it might be something worth pursuing. If the answer to both is no, it's probably not. All of the ideas I worked on this year were no to both 2a and 2b.
This is why I had no idea how to market the products I did finish building, and wouldn't have known how to market the products I didn't finish if I had. I'm not knowledgeable in product deliveries, retail ordering, asynchronous communications best practices, podcast management and growth, marketing, website copy... the list goes on.
That would have been alright had I cared enough about the problem to become knowledgeable about it, but I clearly didn't - that's why I didn't try to figure out how to market my finished products.
This is not a revelation - tons of people in the community have talked about this very scenario. I've heard it myself many times since joining the build in public movement back in June. I let myself believe that any gaps in knowledge or interest could be solved by software and features. There is a point at which you have to close your IDE and figure out the answer to "now what?". It's abrupt, and your knowledge of programming cannot help you much.
2022 and beyond
If you've made it this far, thanks so much for reading all of this. Going forward, I've landed on starting a product in a domain I'm very familiar with and care very much about - a zero DevOps platform built on top of cloud providers. Something a la Render or Heroku. Hook up your git repo and I'll handle builds, deploys, rollbacks, etc. I'll be building this in public, so follow along on Twitter if you're interested.
This time, instead of going directly from market research to building, I'll be doing some marketing to gauge interest in between. If there is interest, this will help me figure out messaging and branding so that I have at least some of the tools I'll need when I finally close the IDE.
Thanks for sharing your story Adam, must not have been easy going through all of that. I notice two things:
I feel like the two are closely related.
Like you, I tried paid ads for my product, and it failed miserably. But I think I know the reason now. For ads to work, I need to get everything right. From product to purchase flow to brand awareness to reassurance... So it's not well suited for an early stage product that's still incomplete.
But with organic marketing, and connecting with people on twitter and other communities, it's 10x easier. Because it makes people trust me (as opposed to an ad from a stranger) and diffuses the fear of buying for a still unknown company.
And the good news is, when you pick a target market you truly resonate with, organic marketing becomes easy. You can just be yourself, jump into conversations and bring value naturally... and it will lead to user growth.
Maybe this question can help you find your next product idea: who do you want to hang out with and help the most?
Thanks Dagobert! I agree with you about the ads - in retrospect, running ads that early on was probably never going to result in sales, though it did indicate to me that people were interested in the idea. I'm sure there are less expensive ways to figure that out, however :)