We Raided Egghead After Midnight and Found Product Market Fit
The story of MenuWorks
It was New Year’s Eve, 1986. I was on a ski trip with friends in Big Bear, California, and it was suddenly my turn to announce a New Year’s resolution. Not having given it much forethought, I simply blurted out the first thing that came to mind. I’m going to quit my job — tomorrow!
Yeah, I did that. I was as surprised as everyone else. I was 26.
Back then I was VP of Engineering at Vault Corporation. The company had done well with a unique solution for diskette copy protection I helped create, but sales had leveled off and I didn’t have enough stock to keep from having a wondering eye.
So without a plan and only about $5,000 in the bank, I put in my two weeks notice and struck out on my own. Again. You see, I had been a founder before. A few years earlier I started Genie Computer and had a good run with a line of hard disk drives for Apple and IBM PCs until IBM eventually introduced its own lineup. I needed to be a founder.
I made a few calls and landed some consulting gigs to pay the bills, but what I really wanted was a product; and more specifically, a software product that I could code and package on my own and sell in computer stores.
I named my company PC Dynamics, Inc., and got down to business.
Something got me to thinking about menu software. The product choice seemed within my economic means and I had watched people become frustrated and confused using competitive products at Vault Corp. I was always telling myself I could make something better.
The premise of the product was simple. It was still a command-line world and people wanted an easy way to run their favorite programs without fumbling with arcane syntax. Windows was still several years out into the future.
I spent the next six months or so doing contract programming by day and working on MenuWorks at night using an early version of Microsoft’s C++.
For context, color monitors were still fairly new, and most people still had legacy green-screen CRTs. WordStar was the de facto word processor of the day and mice were just around the corner.
My biggest competitor was a product named Direct Access which sold for $89, so I priced MenuWorks at $59 with the packaging shown above, which, as ugly as it seems now, was very much in style at the time. My cost of goods was around $8.
MenuWorks had a slow start. Real slow. As much as people liked the product, I simply didn’t have enough money to promote it. My small magazine ads in PC Magazine weren’t working. I wasn’t happy.
Do Things that Don’t Scale
Paul Graham at Y Combinator famously teaches to do things that don’t scale. Do whatever you need to do to cultivate a small group of people who absolutely love your product and let them become the ambassadors to build widespread traction. If only Paul had been around to advise me in 1986.
Being totally determined to make this work, I hired an outgoing salesman named Jim to go door-to-door visiting mom-and-pop computer stores throughout Southern California. Jim was a natural-born salesman and I had no doubt he’d get us on the right track. I couldn’t do this myself because I was still working contract gigs to pay for it all. And besides, Jim was working on commission, so unless things started to sell, this plan really wasn’t going to cost me all that much.
The strategy was to install MenuWorks on the stores’ machines so their customers would interact with it as they looked at the computers, and hopefully buy a copy for themselves. It seemed like a great idea.
Time’s Up, Get Lost
The feedback was pretty instant. Jim told me the store owners would give him about five minutes of their time and then send him packing. That simply wasn’t enough time to even get the product installed for a demo. However, the few who waited for the demo really loved the product.
Jim was frustrated. He told me that unless I got the install time down to a minute or two, including having all the menus built up to launch the applications already installed on the stores’ machines, he’d never be able to get through a demo before being shown the door.
Seriously, Jim — learn how to type faster.
But still, we knew we had something because many dealers loved the product and we were a hit at the local swap meets — yes, we did swap meets!
Man, we were so close, but the traction to survive just wasn’t there.
The bottom line — we didn’t have product market fit. Of course, this fancy phrase didn’t exist back then. I seem to remember we used some other less-polite words to describe our situation.
Automatic Menu Building
Our dealers were literally telling us what we needed to change — we just needed to listen more carefully to the clues.
The “desired” solution they were leading us toward was to rework the installer to magically find, identify and categorize all the installed applications on users’ PCs and configure them into MenuWorks from the get go. But how? There were thousands of popular applications.
Writing the code to do this wasn’t the real problem. The logic was pretty easy to work out. The real problem was that the data needed to make the magic happen simply didn’t exist. I needed access to over a thousand packaged products so I could note the file names and file sizes of their core executables and identify them during setup by scanning users’ hard drives.
To get that kind of data I’d pretty much need to break into a big computer store at night and tear open all their shrink-wrapped products and look at the diskettes to get the data I needed.
Hoping to avoid a felony conviction, I instead cooked up an ingenious scheme with a bright young man named Dave; the manager of a nearby Egghead software store. Egghead had everything.
I bankrolled a plan for him and a few trusted accomplices to come back to the store after hours for many days on end and open up all the boxes, record the file information I needed from the diskettes in a spreadsheet, and then re-shrink-wrap the boxes — like nobody was ever there. It was perfect!
I soon had everything I needed to build a killer automatic menu builder that recognized over 2,500 applications (after including shareware/freeware apps) and created categorized menus for them in under a minute.
For me, finding product market fit was simply about working through those first two critical minutes during setup, and making it so customers never even needed to look at the manual because there was nothing else they needed to know. Nothing else in the software needed to be changed.
New Packaging, New Price & Counter Displays
Preparing to relaunch MenuWorks gave us an opportunity to review another nagging problem — getting lost on the shelf.
Admittedly, we hadn’t actually sold enough of the first version to experience this problem, but it was sure easy enough to imagine given that we literally just opened up hundreds of packages at Egghead to build our automated installer. We surely didn’t go through all that pain just to become invisible in a sea of software.
Our proposed solution was to reduce the product packaging to the size of a thick floppy diskette and create small acrylic counter displays. Better yet, we’d change the retail price to $24.99 and create a standard wholesale SKU that included a display and a dozen copies of MenuWorks for $144.
Granted, these days, getting a counter display placed in thousands of stores would be either impossible or insanely expensive due to coop fees, but back in 1987, stores were smaller and the counters were clear. Surprisingly, all we had to do was ask. We ultimately succeeded in placing thousands of displays without ever paying a dime for placement.
And there was another unexpected benefit of the counter displays — dealers were seeing them start to empty out and would order refills ahead of time to keep them looking full. After all, who wants an empty display next to the cash register— that just feels wrong.
Traction at Last
This revamped version of MenuWorks with a lower price point and cool counter displays definitely had product market fit. And further, the packaging change brought the cost of goods down to around $1.25.
Our automated installer literally blew people away and we were getting lots of great press; including having the honor of being the first product to be reviewed by Walt Mossberg in his inaugural Wall Street Journal column. I can honestly say, we were on fire!
And best of all, we finally landed several key distributors and Jim from sales no longer had to personally visit stores or do swap meets on weekends.
In the end, we sold nearly ten million dollars worth of MenuWorks in various flavors over the next few years; after which point, Windows started gaining momentum and we took that as our cue to move into a different space.
Maybe someday I’ll have the pleasure of meeting Paul Graham at YC and relating my tale. I think he’d be amused at our tenacity and just how little needed to be changed to go from zero to sixty.
And Dave from the Egghead caper — he loved the product so much he joined our team and headed up customer support. He was awesome.
Product market fit is amazing once you find it. We had a great run.