TinyPilot: Month 14

Returning focus to the bottom line.

Highlights 🔗︎

• A redesign of TinyPilot’s website seems to have increased sales.
• TinyPilot now has a European distributor.
• After three years, I’ve earned back my investment in Zestful (and I might sell it).
• I’m still ruthlessly delegating every task I can.

Goal Grades 🔗︎

At the start of each month, I declare what I’d like to accomplish. Here’s how I did against those goals:

Help TinyPilot’s EU distributor achieve his first sale 🔗︎

• Result: The distributor made his first sale on September 6th
• Grade: B

I was hoping we’d earn the first sale within August, which unfortunately didn’t happen. Still, we got a sale within 10 days of launching the EU site.

Finalize the design of the Voyager 2 🔗︎

• Result: The design is taking longer than expected, so it will take another month
• Grade: C

I thought the chip shortage was the last big hurdle to overcome, but the electrical engineers are running into more complexity than they expected in the design. We’ve now broken the project down into individual steps with time estimates for each so that we have a shared view of the overall schedule and how delays affect our timeline.

Publish a sample chapter of Refactoring English 🔗︎

• Result: Almost finished a sample chapter, but I’m still editing
• Grade: C

The thing about publishing a book about effective writing is that people expect you to be a good writer. That puts some pressure on me. I’m about 70% ready to publish a sample chapter to publish, but I still need to edit it more.

TinyPilot stats 🔗︎

Metric July 2021 August 2021 Change

Unique Visitors 5,234 4,194 -1,040 (-20%)

Total Pageviews 9,730 8,864 -866 (-9%)

Sales Revenue $23,954.64 $30,191.04 +$6,236.40 (+26%)

Enterprise Subscriptions $48.00 $48.00 $0 (0%)

Total Revenue $24,002.64 $30,239.04 +$6,236.40 (+26%)

Profit $-9,713.34 $-10,140.95 +$1,664.13

Ouch! That’s three consecutive months of ~$8-10k losses. Worryingly, I increased revenue by $6k from last month but only reduced my net losses by ~$2k.

Part of the problem is the chip shortage. 2 GB Raspberry Pi boards are sold out everywhere, so I had to upgrade to the 4 GB model at almost double the cost, even though TinyPilot gains no benefit from the additional RAM. In addition, I have to buy the boards on a fixed schedule, irrespective of how many TinyPilots I sell, so I’m accumulating more than I want.

Outside of raw materials, my largest cost is software development. I paid freelance developers $13k in July and $11k in August. I have a great development team, and they’re worth what they charge, but the last few months of losses are making me realize I need to choose development tasks more intelligently.

For the next few months, my focus on TinyPilot’s software development will be on things that directly affect TinyPilot’s bottom line:

• Offering a SaaS complement to TinyPilot
• Improving the website so that a higher proportion of customers purchase the product
• Reaching out to larger customers to find out what features would allow them to deploy TinyPilot at scale

TinyPilot website improvements 🔗︎

When I first launched TinyPilot’s website, I meant for the design to be a placeholder until I had more time to invest in polishing it. I made small improvements over time, but the overall design has barely changed.

One of the hurdles was finding a skilled developer to work on the website. After going through six hires who didn’t work out, I found a developer in July with design ideas and worked with him throughout August on implementing them. I think the site looks more professional as a result:

In August, I worked with one of TinyPilot’s developers to redesign the TinyPilot homepage.
Original design (left), revised design (right)

One promising indicator is my revenue per unique visitor. The site tends to earn $4-5 per unique visitor, but in August, the number spiked to $7.20, a 60% increase over the previous month.

It’s certainly possible that the increase was a fluke, as I didn’t do a rigorous A/B test to prove that the new site made a difference. I’ll feel more confident if the trend continues through September.

I’ll go through the changes piece by piece.

First, we adjusted the navbar so that it now lines up with the content on the page:

The feature brags are now more compact. You can’t see it in the screenshot, but we removed the animations, which were a bit distracting.

We changed the reviews from a stack of boxes to a carousel that rotates every few seconds with a new review:

And lastly, we redesigned the footer. It was a jumbled mess before, but now it looks deliberate and organized:

This is just a first pass. I’m going to hire a professional designer to iterate on this. If you have recommendations, let me know.

Adding a European distributor 🔗︎

One of my biggest projects for the last few months has been arranging a distribution partnership with a German company, punkt.de. They launched their website at the end of August, and they’re now accepting orders.

TinyPilot Voyager now also ships from Europe.

Sales have been slower than I expected. Historically, I’ve sold 10-15 Voyagers per month to European customers, so I thought it would be easy for punkt.de to exceed that, as shipment from within the EU means faster delivery and more transparent pricing.

I announced the European sales site to TinyPilot’s 400+ mailing list subscribers, but none of them purchased. Finally, this week, one customer purchased a Voyager, but we’re not sure why the volume is so low.

Still, I’m hopeful that this partnership will be a big step forward for TinyPilot. I’m glad to have another entrepreneur with a vested interest in the company’s success and who can promote TinyPilot in additional markets.

Carving out more time for myself 🔗︎

For the past few months, I’ve been struggling to find time to grow TinyPilot. With so many moving pieces, I sometimes get stuck spending all my time filling the gaps between my teammates.

Now, when I have available time, I invest in sales, but I also reserve time for delegating or automating tasks that others could do in my place.

Automated license checks 🔗︎

One of the silliest things I was still doing was verifying user licenses.

If a user wants to reinstall TinyPilot from scratch, they need a disk image. I don’t offer the image publicly because otherwise, users would have no incentive to purchase it. Under the old system, users emailed me their order number, I looked it up to verify they were a real customer, and then I sent them a download link.

It’s a classic “do things that don’t scale” task. I never prioritized it because it only happened 5-10 times per month. Still, it was a needless disruption, and I disliked making customers wait around for access to software they already purchased.

In August, one of TinyPilot’s developers added a serverless function that checks purchase records from our Shopify store, and we implemented a simple web UI for it. This means that customers can download the latest disk image instantly without me being in the critical path.

TinyPilot customers used to email me when they needed a replacement disk image, but now the process is automatic.

QA testing 🔗︎

A month or two after I began selling TinyPilot, I published a bad software update that broke keyboard input under certain conditions. Ever since then, I used a checklist of features that I test manually before each release.

My checklist started out as a handful of items I could check in five minutes. As we’ve added more features, the checklist has grown longer to the point where testing a release takes me almost half a day.

I’d love to automate QA testing, but some of the workflows are unfriendly to automation. So, I designed a system and documented steps for my local staff to run manual QA tests on spare hardware while recording their screen.

In addition to eliminating a time-consuming task from my workload, delegating QA gives us an opportunity to fix bugs earlier. I noticed a pattern where I’d discover bugs during final QA testing that were not significant enough to delay the release. Now that the QA process isn’t blocked on me, we can run QA cycles mid-release and fix those minor issues.

Can I afford to keep writing? 🔗︎

With the added time I gained from delegation, I’m doing more important but non-urgent things.

• I’m spending more time reaching out to customers proactively rather than talking only to customers with support issues.
• I’m exercising more regularly instead of cutting it out of my day because I feel too busy.
• I’m back to writing for an hour each morning.

I enjoy the habit of daily writing, but looking at my TinyPilot sales for the past two months, I don’t think I can afford it. There are ways that I can potentially increase TinyPilot’s sales that I can’t pursue due to lack of time. So, I’m planning to publish a sample chapter of my book and then scale back on writing until TinyPilot is consistently profitable.

Legacy projects 🔗︎

Here are some brief updates on projects that I still maintain but are not the primary focus of my development:

Is It Keto 🔗︎

Metric July 2021 August 2021 Change

Unique Visitors 39,568 30,439 -9,129 (-23%)

Total Pageviews 96,494 72,340 -24,154 (-25%)

Domain Rating (Ahrefs) 13.0 12.0 -1.0 (-8%)

AdSense Revenue $438.07 $358.43 -$79.64 (-18%)

Amazon Affiliate Revenue $59.65 $43.73 -$15.92 (-27%)

Total Revenue $497.72 $402.16 -$95.56 (-19%)

Last month, I mentioned that I was applying for Is It Keto to join a premium ad network. The network approved me, but I backed out of the deal. The contract terms felt especially onerous, even relative to typically onerous ad deals.

My reading was that the ad partner could request unlimited site changes from me, and the contract required me to carry them out. It also required me to serve their ads exclusively for at least three months. If I even spoke to another advertiser, I’d be in breach. The ad network, on the other hand, can drop me at any time or change the terms of the contract on a whim.

So, I’m sticking with boring but reliable AdSense. My numbers are down, and it’s at least partially seasonal changes. I might also be losing visitors to a competitor that popped up in the last six months. I’m still focusing on TinyPilot instead of worrying about it.

Hit the Front Page of Hacker News 🔗︎

Metric July 2021 August 2021 Change

Unique Visitors 109 393 +284 (+261%)

Gumroad Revenue $218.09 $728.90 +$510.81 (+234%)

Blogging for Devs Revenue $27.30 $0.00 -$27.30 (-100%)

Total Revenue $245.39 $728.90 +$483.51 (+197%)

In August, I tried a pay-what-you-want promotion for Hit the Front Page of Hacker News. For a week, customers could pay any price for the course, starting at $3. It roughly doubled the total number of course customers and brought in over $700 in revenue.

I was hoping the promotion might jump-start sales even after I went back to regular pricing. If customers who bought at a discount recommended the course, it could create a self-sustaining cycle, but this hasn’t happened yet. There haven’t been any course sales whatsoever since the promotion ended, so it’s probably not going to create a viral purchasing frenzy.

Zestful 🔗︎

Metric July 2021 August 2021 Change

Unique Visitors 547 585 +38 (+7%)

Total Pageviews 1,300 1,467 +167 (+13%)

Total Revenue $620.67 $390.80 -$229.87 (-37%)

Last month, I saw an unexpected spike in Zestful revenue, and I dismissed it as likely fraud. To my surprise, it was a legitimate customer, and they paid their balance in full.

This month, another customer had a big spike in usage, and this one seems legitimate as well. Both charges seem to be one-off bulk processing, but it’s good to see more usage.

These past two months bring me to an exciting Zestful milestone: I’ve earned back my investment!

I spent $7.4k hiring a freelancer to help me launch the first version, and I had a few other smaller costs leading to a total cost of $8k. Including August, my lifetime revenue from Zestful is now $8,246, so the project has officially turned a profit. Of course, this was possible because I paid myself $0. On an hourly basis, Zestful has probably earned me about a nickel per hour.

Lastly, a company reached out in August about potentially acquiring Zestful. I was upfront with them that its earnings are small, but if they want to acquire my solution instead of rolling their own, I’m open to it. We had a preliminary call, and they requested some historical data from me. They said they’re considering an offer. I’m not getting my hopes up, but it will be a nice bonus if it happens.

Wrap up 🔗︎

What got done? 🔗︎

• Helped European distributor launch his sales site
• Revamped TinyPilot’s homepage
• Delegated QA testing to my local staff
• Booked appearances on two podcasts
• Automated license checks for TinyPilot
• Migrated TinyPilot’s mailing list from Mailchimp to EmailOctopus

Lessons learned 🔗︎

• Even though I’ve freed up enough time to write every day, it’s too early to borrow time from TinyPilot.
• I need to choose development tasks that have a high chance of impacting TinyPilot’s earnings.

Goals for next month 🔗︎

• Publish a sample chapter of Refactoring English
• Start development on a monthly service-based software complement to TinyPilot
• Finalize the design of the Voyager 2