My startup story #2: Inmar, Inc.

posted: September 18, 2018

tl;dr: B2C takes money, and most product ideas have a finite lifetime...

The second startup company I was involved in, Inmar, Inc., was where I got my business education. You can learn an awful lot about how a business works just by being in a small company, keeping your eyes open, and pitching in wherever help is needed. Inmar was close to the smallest of all possible companies: just four employees at the peak.

Inmar exposed me to just about all aspects of a business: not only product development (both software and hardware design) but also manufacturing, vendor relations, quality control, fulfillment, customer support, marketing (advertising, collateral, trade shows), sales, and accounting. I got to watch closely during important business negotiations including the sale of a line of business. Years later I would attend an executive education program at the Harvard Business School, but that was, for me, more like a formal finishing school which gave me a certificate for material that I had already learned on the job. Much of that learning was at Inmar.

Like the first startup company I joined, Xciton, Inmar was also founded by my father, Paul Shaver. I joined after college, and another family member worked for Inmar during his college years. We were cheap labor 😀. Inmar was a very different kind of startup than Xciton, which was VC funded and had other outside investors. Inmar was the bootstrapped startup model, living off the savings of the founding team (in this case just my father) and the cash flow that the company could scrounge together, hopefully as early in the company’s life as possible. So one of the key lessons I learned was the importance of cash flow management.

To get some cash flowing, Inmar had a sideline business assembling and selling “clone PCs”, i.e. knockoffs of IBM’s original PC design. In the late 1980s there were many companies around the world making components (motherboards, power supplies, disk drives, cases) that could either be put into an original IBM PC or used to assemble a PC clone. You could follow and participate in this market by buying a copy of The Computer Shopper, a large-format newsprint magazine with hundreds of pages of ads, which functioned as the industry’s price sheet and supply chain directory. One entrepreneur, Michael Dell, who started selling clone PCs in his dorm room at the University of Texas, eventually rose up from this bazaar of thousands of companies to become the leader of the preeminent clone PC company, Dell Computer. Selling clone PCs was never the main focus of Inmar; it was just a way to bring some cash into the company.

The echoBOX, to the right of an original IBM PC keyboard

Download page one and page two of the original echoBOX datasheet

The focus of the company was an invention which is going to seem incredibly silly, trivial, and useless today, but was exciting enough at the time to convince multiple people besides myself to embark on a risky startup journey. The invention was the echoBOX, a programmable macro keypad for PCs that sat inline in between the detached PC keyboard and the main CPU enclosure. The echoBOX had 12 keys plus a shift key. It provided what were, in effect, completely programmable function keys that, when touched, would playback whatever sequence of keystrokes they had been programmed with.

Younger present-day computer users will have to realize that the world was a very different place in the late 1980s. Yes the Apple Macintosh was released in 1984, but the IBM PC and its clones owned the business market, and it was still years before Microsoft would release a usable version of Windows. When PC users booted up their PCs, after a few minutes they were presented with a command line prompt, typically:


and were then expected to start tying in operating system commands. Yes, we developers even today rely upon the command line heavily, but a command line waiting for a user to type a cryptic, non-English-word command from memory is very intimidating to non-technical users. That’s one of the key problems that graphical user interfaces such as the Macintosh and Microsoft Windows solve. The paradigm has shifted completely away from the command line: today most computers boot to the graphical user interface, and technical users are forced to hunt around to find out how to launch the command line.

In a pre-Windows world, the echoBOX was one solution to make it easier for non-technical users to use a PC. The keys could be programmed with the commands to launch application software, or to perform certain common tasks within the applications software, which also relied on the users typing on the keyboard. The echoBOX had several advantages over other similar solutions. It was a completely self-contained device with its own processor and software, and didn’t need any driver software to be installed on the PC. It was very easy to program and use. It simply sat inline in the keyboard cable path, remembered keystrokes when put into program mode, and played them back when in normal operation.

On the technical side, my major contribution was to redesign the product to produce the second generation device. I joke now that I’m a full stack engineer from the time when being full stack meant designing your own hardware and then writing the software which ran on it (in assembly language, in this case), which is exactly what I did. The major advance in the second generation was switching the technology of the non-volatile memory which stored the keystroke sequences from battery backed-up Static Random Access Memory (SRAM) to solid state technology using a new Electrically Erasable Programmable Read Only Memory (EEPROM) chip, basically a precursor to what is today called flash memory. I’ve always tried to avoid batteries and moving parts whenever possible; solid state technology is just so much more, well, solid (and reliable). I also took the opportunity to migrate the microcontroller from an Intel 8048 to a Motorola 68HC11; like Steve Jobs, I appreciated the more elegant processor designs and instruction sets made by Motorola back in the day.

The main challenge we had, which we couldn’t ultimately solve, was that the product was fundamentally a Business To Consumer (B2C) sale. We were a tiny bootstrapped startup company, with very little money to spend on marketing and advertising, trying to get consumers to change their behavior and adopt a product that they had never heard of before. We made the mistake of falling in love with the market size, thinking “there are 10 million PCs out there, if we only capture 1% of the market then we’ll sell 100,000 of these things and we’ll all be rich”. To capture even that 1% of the market takes more resources than we had available to us. B2C firms typically have to spend a lot to capture their initial customers and marketshare; there's something called "Customer Acquisition Cost", and it's usually not zero. We learned this lesson the hard way.

There were many other lessons learned. Our closest competitor was another keypad company that had an inferior, harder-to-use technical solution that required software to be installed on the PC. However they had a nicer-looking industrial design: their keypad attached to the top of an IBM PC keyboard with the exact same coloring scheme, so it looked like it belonged there. More importantly they had a bigger advertising budget. They sold more than we did, just like VHS outsold Betamax. I also learned the challenges of designing a fully interoperable solution that could work with the hundreds of keyboard designs on the market, some of which had strange or poor implementations of the PC’s keyboard interface, which wasn’t actually an official standard. Getting something that works 100% of the time, in production, under all conditions is one of the main challenges in product development.

What ultimately happened? We were getting a little bit nervous about the rise of other technologies such as graphical user interfaces and the mouse. Fortunately we had attracted the interest of a wealthy businessman who fell in love with the product and offered to buy it from us. After pitching it to several other interested parties we sold the invention, the designs, the intellectual property, the customer base, and the in-process inventory to the wealthy businessman, who ran it a while longer (this was the first asset sale in my career). However, he also lacked the financial wherewithal and marketing acumen to garner the attention of the typical PC consumer.

After the sale, which netted me a little money, I moved to Portland, Oregon, and ended up joining my third startup company...