Since the last time I sent out this newsletter less than a week ago, the number of people who subscribe to my mailing list has grown from 48 to 165! I feel humbled, inspired, and in disbelief at this number. Thank you for being here. Onward, dear friends.
After I submitted my last post "How to Mail an SD card with Gummy Glue" to Hacker News, a comment by the user tjohns caught my attention.
[…] Don’t try to play games with customs in the mail. It’s a terrible idea. […]
Needless to say, drawing the ire of any customs agency in any country would make for a bad start. This shrouded warning from the gods prompted me to double-check a few things.
A standard mailing envelope, like the one I plan to use, is intended for correspondence and nonnegotiable documents only. It does not require a customs declaration form to send it internationally, which may not even be possible to complete for this purpose. I’ve certainly never seen it done. Upgrading from an envelope to a package is always an option but would add to the cost.
As for customs duties, take the United States, for example. According to the Harmonized Tariff Schedule of the United States (HTS), the SD card and its contents would fall under heading 8523, and may be imported duty-free. That is unless you are mailing it from Cuba or North Korea, which I am not, in which case 80% rate of duty would apply.
Banks such as Wise, based in the United Kingdom, and Wells Fargo, based in America, routinely send replacement client cards internationally. The client card, in that case, is actually a "smart card." It stores data on its integrated circuit chip and even has an antenna to energize itself. In the context of standard post, I see no practical difference between the integrated circuit in my SD card and the ones in the "smart cards" mailed by the banks.
Ultimately, is my distribution strategy completely beyond reproach? It is hard to say. In reality, though, is anyone going to care about a few SD cards sent in the mail? Unlikely.
In sales, one would do well to generate some urgency. Scarcity helps too. I was long wondering how many SD cards I should prepare for sale.
An empty card costs 7 EUR. Stockpiling 10 of them to start seems reasonable to me. I decided to limit the release to 100 cards.
With a background in SaaS, I am used to the industry standard 2% sales conversion rate. This means that for every 100 people who come to my store, just 2 will buy something. Somewhat worryingly, I read that the conversion rate for physical goods, especially in consumer electronics, is much lower. Correspondingly, to sell 100 cards, I would need, at a minimum, 5000 subscribers across all of my marketing channels. That is my goal for Christmas.
While the number of subscribers to my mailing list has been growing exponentially, I suspect that the hacker and maker community is somewhat over-represented in it, which may translate to fewer sales. So, I also started a YouTube channel and uploaded a practice video which I expected no one to see. Shockingly, it received 68 views, 732 impressions, and 4 very supportive comments in just 3 days since it was published. Waking up to notifications about new YouTube comments was a new and interesting feeling.
When applying to Y Combinator, you are asked to submit a video in which you talk about yourself and what you’re making. It is suggested to speak without a script and to resist the urge to edit the resulting video. This may work well for the Y Combinator interview process, but not when you’re selling a product. I did use a script, as this was just for practice.
As human beings, it is in our nature to love storytelling. I wanted to tell the story of why computers are personal to me dating back to my childhood. Judging by the comments, it seems to have worked.
As for the newsletter, I decided to transition to a weekly publishing schedule, delivered to your inbox every Thursday at 21:00 London time.
It may come as no surprise to some of my readers that selling physical goods is vastly different from selling their digital counterparts.
Take chargeback fraud, for example. You create a good, a customer orders it, you ship it, they receive it, and both of you go on your merry ways. A few days later, though, your payment processor informs you that the customer disputed the payment. The product never arrived, the customer claims. Or it does not work as advertised. Or their card was stolen and the payment was made fraudulently by miscreants on the prowl for experimental computers.
This trope, growing so common that there are now Internet forums for perpetrators to share strategies for getting free stuff, has become endemic to online commerce. I experienced my fair share of it while selling SaaS and have learned a few strategies to deal with it over the years. The physical world, though, is much more precarious. When a chargeback occurs, not only is the payment held up for months while the banks examine the case, but the chances of you ever seeing the product returned to you are close to zero.
So, I decided to reexamine what I am actually selling to the customer. The value, after all, is not the SD card itself, but the OS and software that it contains. I loathe to admit it, but in this particular case, digital access rights may have a part to play.
In that approach, the customer would purchase the digital rights to access the software and then be shown the download link at the end of the checkout flow, thereby completing the business transaction. Immediately after, the customer would be presented with the option of having the SD card mailed to them. In this scenario, the digital goods purchase would be harder to dispute. The reviewing bank is more likely than not to side with the seller if I can furnish the relevant server logs and screenshots of the checkout process.
At the same time, I am deeply opposed to Digital Rights Management (DRM). If something were to happen to me or to this project, then there may be people out there who bought the compatible hardware but are unable to use it because they can no longer get access to the software. Therefore, I will also provide a BitTorrent link. Software piracy is not a concern for me at this stage. A potentially tarnished reputation for vaporware, however, is.
Unfortunately, Shopify, the e-commerce solution that I planned to use, doesn’t seem to be optimized for purchases of digital goods. As the Fates would have it, SaaS is my bread and butter, and it would not be difficult for me to build this checkout flow myself.
Another benefit of this approach is that the customers who are capable of flashing the SD card on their own would be able to do so. This means that even if I run out of cards, customers would still be able to get the software. Then again, that is a double-edged sword: those customers would be buying compatible development boards instead of my future products. But that is a blog post for another week.
Until next time!
I greatly appreciate your insightful comments on Hacker News. They help steer me clear of danger as I travail the wine-dark seas of bootstrapping a hardware business.
This newsletter was published on my engineering blog, which can be found here.
My significant other has urged me to remind the reader that none of this is professional or legal advice. I’m not recommending you do anything I mentioned above. If you’re planning your own SD card empire, get professional legal advice.