A rule of thumb about rules of thumb

In the world of software development and tech startups there are a few rules of thumb that are widely known. For the purpose of this post let’s consider just one: that startups should always host servers on a service like Heroku, AWS or Digital Ocean, as opposed to operating a non-virtual machine yourself. The logic is that by outsourcing the work of running the server, you free up time to work on your core business, and avoid tasks outside of your core competencies. It’s a good rule of thumb.

But what is a rule of thumb, exactly? According to Wikipedia, “A rule of thumb is a principle with broad application that is not intended to be strictly accurate or reliable for every situation.” (link emphasis mine.) It is not reliable for every situation, which is often overlooked. To use a rule of thumb properly one needs to understand the reasoning behind it. I’ll return to rules of thumb in a moment.

Part of being a tech worker is talking to people who don’t understand what you do, e.g. a programmer working with a non-programmer. A strange norm has emerged in many workplaces where the non-technical workers fawn over the technical workers, and act almost worshipful towards them because of their technical expertise. This is a harmful dynamic that leads to overconfidence on the part of technical workers.

When you combine this hubris with a rule of thumb like the discussed above, the consequences can be harmful to your business. Technical employees will make major decisions (“how should we host our servers? on managed VM’s!”) without properly considering the tradeoffs, because a rule of thumb exists that is widely accepted so there is no need to examine the tradeoffs. And since the worker never understood the rule in the first place they are both 1) not able to understand how to make the decision 2) confident they have made the right one.

I’ve also seen technical workers who have been “the smartest developer in the room” at their company for a long time forget how to talk to a peer who actually *does* know their stuff. When challenged on a point, they will cite their long history of working with clueless people who have make mistakes, which isn’t responsive.

If you’re about to spend a lot of money or time on something at work, don’t rely on rules of thumb. Understand the fundamentals of the decision you’re making and what resources are available to you. Understand the cognitive biases commonly suffered by software developers and how that causes companies to cargo-cult instead of being creative. Rules of thumb are popular because they’re usually right, but when one doesn’t apply to your business it is a chance for you to stand out and do better than your competitors.

Consider the example of Pinboard vs. Delicious (here comically referred to as ACME).

But where the projects differ radically is cost. ACME hosts their service on AWS, and at one point they were paying $23,000 (!!) in monthly fees. Through titanic effort, they have been able to reduce that to $9,000 a month.

I pay just over a thousand dollars a month for hosting, using my own equipment. That figure includes the amortized cost of my hardware, and sodas from the vending machine at the colo.

And if you see someone make a decision that breaks the usual rules, before you assume that they made a mistake, take a minute to consider the possibility that they made the better decision.



Latest startup launched!

My latest startup is live! Visit it here: Aardvark Server Company

It is a B2C SaaS for players of the game 7 Days To Die. It allows customers to spin up a private server and edit the server settings. Currently it works perfectly although some things I still have to do “by hand”, like shut down a game server after the free trial has expired. I will automate more functions if I get some customers, based on their feedback.

I’m 100% all-in to make this a great game hosting company. If I get some customers I’m going to “do things that don’t scale” to make sure they have a great experience.

I have had some visitors to the site from entrepreneur and hacker forum posts, and one user spun up a server on the free trial. No one has signed up for a subscription yet but since the launch I’m full of new ideas, have thrown away some bad ones, and am very focused and excited about the future of this service.

You might wonder what happened to me founding a startup every month? The answer is simply that after having some success the first month with Marginal Revolution Books, and then launching flops the next two months, I realized I needed to sit down and study design. So on the advice of some friends on twitter I worked through the book Design for Hackers, which took about a month (August). I then spent some time teaching myself how to do SEO-related research, e.g. how to answer the questions, “are people searching for this? How many? Are their needs met by what they find? Are they thrilled?” Then I spent two months (Sept-Oct) coding up the site.

Why not launch sooner and “get feedback from users?” Because this service only works in a way meaningful to customers once they can launch servers and play on them. Since this market is already liquid there’s no need to do a campaign and collect email addresses of people who are interested. I know there’s an interest. Anyone can see that by researching the market.

After the last four months, more than ever I appreciate how much harder entrepreneurship is than it seems from the outside. But now that I’ve put in the time and effort I am so excited to see what happens next!

How I got linked to by Tyler Cowen, sold 124 items in three days, and got rejected from Amazon Affiliates

tl;dr The first startup I made for the “12 Startups in 12 Months” challenge unexpectedly got a ton of traffic, it sold a bunch of stuff on Amazon, then it got rejected from the Amazon Affiliate Marketing program.

The Traffic Surge

Last Saturday (June 3rd, ’17) after spending a few nights and weekends building a side project called Marginal Revolution Books I decided I had a Minimum Viable Product on my hands and it was time to validate my idea. I posted a link to my site on the Libertarian subreddit (which got no love), and the Indie Hackers forum.

The next day I woke up to this

While I was sleeping, Tyler Cowen tweeted a link to my startup, with a ringing endorsement! It seems I had “underrated” my site. People really liked it! A few hours later (when does this guy sleep?) Tyler linked to my site from the Marginal Revolution blog itself.

This is the first time one of my projects has gotten traffic, and it was a lot of fun. I acted on the feedback I got from the site’s users and added an email newsletter, a search feature, and a “random month” button. People I had never met were tweeting a link to the site, including one in Greek!

That’s a first!

Screenshots of analytics dashboards

Here’s my google analytics

Here’s the amazon affiliates dashboard (taken the day before they shut me down… more on that later!)

The site never crashed and response time stayed reasonable. I also collected 102 email addresses from the little AJAX form I threw together pre-coffee on Sunday when I woke up and saw the traffic surge. (I think this is what a form made by Jony Ive in 10 minutes would look like, not sure.)

In the end, does it even matter?

On Thursday I received this email

Amazon rejected me! I won’t be getting paid for the traffic I sent their way.

Was my project a success?

After the site was rejected, I complained about my bad luck on the Indie Hackers Forum. The site founder, Courtland Allen himself, analyzed my situation.

[…] Furthermore, you’ve done a few really great things in a very short period of time:

  • Built something useful enough that lots of people derived some value from it, including Tyler Cowen himself.
  • Learned a lesson about having a business model that can stand on its own and can’t simply be shut off by AppAmaGooFaceSoft (as patio11 calls them).
  • Learned a lesson about entering a market where the sales you make are substantial enough to pay the bills.

It takes many people years of hard work and hundreds of thousands of dollars invested to learn those last two lessons, and you just got a crash course in a week or two! […]

Do read the whole thing. The other comments helped me a lot too! Let’s figure out whether building the site was a good use of time. (Spoiler alert: it was.)


I spent less than a month building the site on nights and weekends. In the process I

  • learned how to use the python scrapy library to spider websites. (Thanks for the pointer Cleeftone!)
  • learned how to extract product info from the Amazon Product API
  • did a bunch of tech stuff I’m experienced at, like building a rails app, hosting, design db schema
  • did some tech stuff I’m not experienced at, like using the bootstrap lib to style a site
  • did business stuff I’ve never done before, like promote a site, respond to user feedback, collect emails for a newsletter
  • built something actual people used and got value from!
  • Learned multiple hard lessons quickly
  • Can now change my twitter description from “wannabe indie hacker” to just “indie hacker”
  • Tyler liked it! yay


  • Paid $7 to Heroku. That’s almost enough money to buy a beer in SoHo!
  • I’m not getting paid for something I didn’t expect to make any money anyway
  • My career as a affiliate marketing baron is over before it started

What’s next?

I’m going to leave Marginal Revolution Books up and running. It only costs $7/month and people are still using it, why not? I’m so glad that I can give something back to a great community like MR! I will also add features to the site if you ask nicely and I already know how to make them.

As for me, in the last seven days I’ve interacted with some really cool people on twitter, including Tyler Cowen, Courtland Allen, Ryan Hoover (the Product Hunt guy) and a bunch of other cool people. Old friends are coming out of the woodwork and reaching out to have coffee, wanting to brainstorm, etc. I’m very excited about the future!

Watch this blog for an announcement of my second startup in 12 months, coming once I have an MVP built.


Product Idea Generation

I want to build a product

As an aspiring bootstrapped entrepreneur, I need to find product ideas that work for me. That means a product:

  • I can build during nights and weekends
  • that requires very little capital
  • in a market I already understand or can quickly learn about
  • where demand exists

My first undertaking as an entrepreneur has been researching product idea generation. Here is a list of methods I found, along with hyperlinks for more information.

Methods for finding product ideas

Look in the Apple App Store for an App that is widely used & sucks. Then make a version that doesn’t suck. Indie Hackers

Look in the Apple App Store for a successful App, implement non-English language version. (same link as previous method) Indie Hackers

Build Browser Extension for your own use, put in Chrome Web Store, charge $2 for pro version. Indie Hackers

Get emailed a product idea every day. Opps Daily and Nugget

“look for problems, preferably problems you have yourself.” Paul Graham

When a popular service like del.icio.us or Google Reader gets shut down, build a replacement service. Example: Pinboard capturing the del.icio.us exodus, and NewsBlur doing the same for Google Reader.

Forget about SAAS recurring revenue, and build a one-time purchase product like a WordPress plugin, Magento add-on, Shopify app, Drupal add-on, Photoshop plugin, or an ebook. Software By Rob and Hacker News

Build an aggregator/scraper of other sites. Indie Hackers

Build cross platform desktop app using Electron, the software Slack uses for their desktop app. Electron and Hacker News

“Fast Follow” an existing SAAS product with an already validated market. Hacker News

“Hipster retro businesses.” Build software for sellers of products like vinyl records, camera film, and magazines.

Next steps

My plan is to find a product idea I can build quickly, since most people don’t succeed on their first try. I will build the product and launch it. That way I can gain experience doing stuff I’ve never done before, like marketing. Hopefully the experience I gain will lead to revenue from a future product.

Do you know any ways to generate product ideas I didn’t mention? Please let me know in the comments!