I teach entrepreneurship at Duke and Iām publicly growing a company ā Autopest ā from $0 to $100k/year in revenue in order to help entrepreneurs better understanding the process of building startups. Learn more about my journey here.
As someone who teaches entrepreneurship, Iāve spent lots of time preaching about the challenges of being a solo founder. I bumped into one of those challenges this week. Or rather, I suppose Iāve been bumping into the challenge since the day I first conceived of Autopest, but I only just realized it was happening this week when I discovered a critical part of my product was broken, and I had no idea. š¤¦āāļø
I wonāt waste your time detailing what was broken since you surely have no reason to care. Iāll only mention that Autopestās transactional emails (the emails Autopest sends its users, like āwelcomeā emails when people register) werenāt working.
The reason I didnāt realize the transactional email system wasnāt working was because I have my email settings configured a certain way, and it never occurred to me to check Autopest on someone elseās email account. Presumably, if Iād had co-founders, they would have noticed the issue and told me. However, as a solo founder, I didnāt have anyone else who knew enough about what I was building to recognize the issue.
In other words, I was dealing with the biggest problem of being a solo founder: I couldnāt see one of my own blindspots.
By definition, of course, nobody can see their blindspots. Thatās why theyāre called āblindspots.ā We solve this by working with other people. Co-founders are the people who help us see all the things weāre too naive, selfish, ignorant, stupid, egotistical, lazy, narcissistic, disinterested, or otherwise self-absorbed to see and/or admit to ourselves.
If youāre like me and trying to build a startup by yourself, consider my stupid mistake with Autopestās transactional email system a reminder that, yes, you have blindspots, too. Donāt be like me and keep building on your own. Iām just doing this as an educational experiment. Since, presumably, youāre not, do yourself a favor and find a co-founder. Yes, it can be hard, but itās easier than destroying your company with mistakes you donāt even realize youāre making.
-Aaron
Core Metrics
WEEKLY ACQUISITION METRICS:
Site Visitors: 451 uniques (+1%)
New Free Users: 44 (-12%)
Website Conversion Rate: 9.8%
New Paid Users: 0
AGGREGATE ACQUISITION METRICS:
Total Free Users: 294
Total Paid Users: 1
Total Revenue: $15
Total Costs: $20.91
Net Revenue: -$5.91
WEEKLY USAGE METRICS:
Extension Installs: 9 (-31%)
Unique Senders: 1 (-83%)
Funnel Analysis
As I explained last week, Iām having trouble getting people who register for Autopest to actually use Autopest. To combat this critical issue, I made some significant changes to the signup process. After the changes, the process does a better job explaining what to expect when people register, and it provides more education on how to use Autopest once itās installed.
I expected these changes to impact three things:
the number of signups was going to drop because there was more friction in the signup process;
more people were going to install the Chrome extension since it was better explained; and,
app usage was going to dramatically increase.
None of that happened. Instead, signups stayed roughly the same (a slight drop, but within expected statistical variation), which makes no sense because I added multiple steps to the registration process. Thatās a cardinal sin of UI design and basically a way to guarantee your process will be less efficient. Or I thought it was. But, apparently, that wasnāt the case.
At the same time, the number of Chrome extension installs and the number of unique users sending messages dropped like a rock.
Honestly, Iām confused. Everything was basically the opposite of what was supposed to happen.
Iām trying not to get too bummed out because, to be honest, the numbers Iām working with are small enough to not be statistically significant. Maybe Iāll collect data for a few more weeks before I draw any macro conclusions about how everything we all thought we knew about user interface design was completely wrong.
In the meantime, I still need to figure out better ways of activating users because a product without engaged, paying users isnāt really a product. Itās just an inefficient way to collect peopleās email addresses.