The Doctor looked at the bump on my neck and said "it's nothing, but we'll do a biopsy just to be safe." And the result was – cancer. I had stage IV squamous cancer that had started in the tonsil and spread through part of my neck. My odds were good, but… what matters is where the data point named David Thielen falls on those odds. It's very disquieting to realize that I can do everything, absolutely everything, and still end up dead.
So how do they treat squamous cancer? Basically they almost kill you at the cellular level. Every cell is severely wounded. And the cancerous cells are killed in the process because they are killed by treatment that only (interesting word only) seriously hurts the healthy cells. And they also apply radiation to the throat area. That so injured my throat that for 6 weeks I was fed via a tube permanently tied into a vein in my chest. My wife and youngest daughter mixed up the feedbag twice a day and plugged it in to me – good times.
Now here's the part they don't tell you (and it's a good thing they don't). The day the treatment is over, a week after treatment is over, a month after treatment is over – you are still half dead. But slowly, ever so slowly, you start to regain energy. And that led to a major problem. I wanted to work again. I was going nuts not working. But what company is going to hire someone who says "I now have enough energy to put in a good 3½ hours/day of effort?" Not gonna to happen.
So I decided I needed to write a program to keep busy as I regained my strength. So I thought back to an idea I had had several years earlier…
The CFO at a startup where I was VP of technology was describing the reports he needed our system to produce for their tracking. We were struggling over it for various reasons and so I asked him to create what the final reports should look like for me in Word and/or Excel. He would make up the actual numbers but it would give me layout, formatting, and the data we needed to record.
Important lesson I learned here – there was data he needed in the reports that we had not put in our schema because we did not realize it would be needed. Always design the data part of your reports before you lock down your database schema. Always!
Ok, so he put that together and emailed them to me and what struck me was – why can't I now just click a button and turn those into report templates. It was all there, and yet we then had to start from scratch using Crystal Reports (which my developers detested) to build up what was those examples. So that became my guiding principal – to take sample reports created by the end user of the reports and just add tags for the data.
The Product Design
Ok, the way you're supposed to do this is you look at the market, review what is out there, talk to potential customers to see what they want. But I did it the way most software start-ups operate, I had an idea that I believed people would want and had a clear idea in my head of what it should be. I also had an advantage that I wanted this at several companies I had worked at.
The program was written in Java because 7 years ago Java was the technology for enterprise software. CGI programming under C++ was in some use and .NET was getting started, but Java was the big boy. I also went with XML instead of SQL for the datasource because you can push SQL to XML but putting XML into SQL is difficult to impossible in many cases.
So I spent a couple of months, got a complete system, and almost released it as open source. If I had it would have been a very different story, it probably would have been left as is. Why? Because it was so easy to use and rock solid – so there would have been no reason for any company using it to purchase support. But I did decide to put it up as a commercial product and…
In that first month a couple of companies bought it, including a bank in Japan and a state agency. It was a real product and enterprises found it useful. Not only useful, they found it to be a significant time saver. And this was version 1.0 – XML data only, Word only, no AutoTag, no charts, etc. But that core idea, simply implemented – people found it compelling.
At first it was a pretty minimal operation. I had a student working part-time covering support as needed (we would get 1 – 2 questions/week), another person part-time doing marketing (mostly Google ads and some white papers), and a part time person handling operations, finance, etc. And as much as possible we were focused on eliminating the need for a person to answer questions – making the program so easy to use there was no need to ask us questions and setting up the online store to handle sales.
I'm very proud of the fact that we had customers who evaluated the program, purchased it, and used it without ever needing to ask a question. I think that speaks louder to the simplicity and how solid the code is than anything else. On the flip side, as we've grown and grown, for a long time there was no significant income from training or professional services – because it's rarely needed (this has changed as it is now being incorporated into very complex systems).
Ok, so I started selling it and people bought it. Not a lot, but enough. So we were a very rare thing, a start-up that was profitable from the beginning. Because of that, we have never had to raise venture capital. This was invaluable during the recession because we were not in the position of many companies who needed to raise money to stay in business. Having control of the future in your own hands is very scary at times, but it also has significant advantages.
One funny note on retaining control, we do have an advisory board with some very experienced people who's advice I greatly value. A year ago they told me that I needed to bring in a CEO and move myself from CEO/CTO to just CTO. And it was absolutely the right decision – under our new CEO we've grown at 50% during the recession – she's really good. But for those that figure if you hold a majority of the stock you can't be demoted – yes you can (and most often should once you get to a certain size).
With all that said, we did hit a point about 2 years ago where we started losing money month after month. We cut where we needed to and my wife and I put some of our own money in to get us through that trough. Running off cash flow with no investors is great, but make sure you have reserves you can call on, if not your own then an investor in some form.
Ok, so I had something and the company needed more than me. So the first people I got were a full time salesperson, a part time marketing person, and a part time support person (we had so few questions a full time person was not needed). And with that team we grew to almost a million/year – 2 full time people and 2 part time. We've grown a lot since then but I want to emphasize this for those looking to start your own company – you don't need a ton of people. You're left having to do many things yourself but that's why it's called sweat equity.
Granted, we figured out how best to use Google (both PPC and SEO) as well as social media. The things we did can't be repeated today. But there are new approaches today where you can get a lot of attention for a small amount of effort. And Google remains the great equalizer. Competing with the big guys using the same marketing approach they use is close to impossible, so you have to find a better way to get the word out – and that rarely requires a lot of money.
For example, we were asked a lot if we would be around in a year. Windward was first created to develop the game Enemy Nations and so the company was over 10 years old. So I would point this out and say we existed before Enron and are still around after Enron. People would laugh, and then accept that as we will be around. The trick is to find some distinct advantage or difference that you can use.
- We ported the engine to .NET by using Microsoft J#. So we have a single set of source code (about 2% is platform specific), but it is then complied to native Java and .NET code. It's a really cool way to do this where we keep a single code base, but both engines are native code. The J# group at Microsoft did a superb job.
- Moving to our new offices a couple of years ago. The larger offices and the nicer facilities was to me a sign that we had become a substantial company.
- Fully incorporating AutoTag into the Office ribbon and making chart & bitmap tags actual charts & bitmaps in Word. It took us to a level of natural use for designing reports in Office that was an order of magnitude over our previous effort. And that previous effort was way beyond anything else out there. I think we now have a report designer that makes everything else look prehistoric – and that's super cool.
- The emails I get from customers. That's the validation that we have delivered something to them that is of great advantage, and that we do so with a very responsive support team.
Sometimes I miss the early days. The program was 100% my code, a high priced Google keyword was 28 cents, and the overhead was minimal. But mostly I really enjoy where we are. I've built a development team that is as good as I've worked with anywhere. We have a killer product that saves people a lot of time and money. And I have a company that provides good jobs to a number of people.
So to those of you thinking of starting your own company, if you think you have something good – go for it. And remember step 1 does not have to be raising money.