For those who like Bootstrapping Case Studies, here is mine.
2012 was the year I moved on from a Day Job and started doing my Bootstrapped Company CNCCookbook full-time. I’m not the first to do so, and certainly not the last, but I thought I’d provide a historical background and then some data on CNCCookbook in 2012 along with lessons learned so far to help those that follow. There are a lot of other stories along these lines that range from extreme bootstrapping successes like 37Signals to smaller successes like patio11′s Bingo Card business. This is not my first entrepreneurial endeavor, but rather my 7th. However, the others were all venture funded while this has been the first that I bootstrapped from scratch without any outside investment.
CNCCookbook was incorporated 2 years ago in late 2010. We sold our first software in November 2010. The product was called “G-Wizard Calculator“, and it is a very specialized calculator for CNC machinists. I started the company largely because the Silicon Valley VC Startup scene had veered a long ways away from my vision of what startups ought to be and always had been for me. I wanted to have fun building awesome software that solved real problems my customers had and gave them good value. I wasn’t much interested in Riding the Bubble, lucrative as that can be. I was more interested in doing the sorts of thing I read about from people like Seth Godin than in building the next Consumer Internet app. In short, I wanted to build value the Old Fashioned way, or at least that was the delusional fantasy that has sustained me on this enjoyable journey.
I quickly discovered there’s quite a lot to be learned in the course of bootstrapping, but fortunately, I found all of it to be interesting. Some of the key problems you will have to solve to build a successful one man software company include:
- How to take orders and charge for the orders.
- How to control access to your software so that you can have free trials, paid subscriptions and renewals.
- How to get people to try your software.
- How to provide customer service by way of answering questions and dealing with all the minutiae that comes of helping customers complete their purchases and learn how to use the software well enough to realize the value they’ve purchased.
- How to keep the business moving forward and growing at an acceptible rate.
- Perhaps most of all, how to juggle all the balls while getting everthing done and holding down a Day Job using only 10 hours a week. If you can’t manage your time effectively, get fired up to do more after a long day at work, and know what not to waste time on, bootstrapping may not be for you.
Back in 2010, I had a reasonably thriving web site/blog that I’d started in order to chronicle the machine tool hobby I’d started 10 years before. 2007 was the first full year the web site was in operation. Prior to the end of 2010, I did little to the site except write blog posts that were interesting to me on an irregular schedule. People tell me I’m a pretty good writer, and I had a substantial readership with nearly 600,000 visits in 2010.
Here’s some data on site traffic from 2007 through 2012:
From 2010 to 2011 growth flattened out because I was splitting my attention too many different ways and not producing enough content. At the end of 2011 I instituted a rule that I would not work on marketing unless it was after 7 pm in the day. 2012 grew like crazy because I had figured out the content marketing formula and I was cranking on good content rather than wasting time on things that didn’t matter. We did very well ending 2012 with over 1.2 million visits. Is that a lot or a little traffic?
Look at it this way. The market leading CNC software company is Mastercam, and SEMRush says their organic traffic is 10,810 while CNCCookbook’s is 4,032. If a one man shop can get 40% of the traffic of the market leader, I’d say that’s pretty good. In fairness, they count on being sales driven and using a reseller channel, but even so, that’s not too shabby. Not only does content marketing work, but it works well.
We attract our visitors entirely via the content, whether through referrals from other sites or via organic search traffic. There’s no paid advertising except when I am experimenting off and on with it. As a solo entrepreneur, I don’t have a tremendous amount of time to spend on site traffic. In fact, I have a rule that I only do marketing work after 7 pm in the evenings. The rest of the time is devoted either to customer service or building product. Not only does content marketing work well, it is extremely time and cost efficient, at least if you’re good at producing content that works.
Despite my marketing “curfew”, quite a lot of marketing gets done. To give an idea, I logged 126 specific events in my marketing event log that I use to try to understand what was going on when the various analytics changed. Events are typically things like running an A/B test of some landing page. There are another 100 or so items from my Agile Backlog for marketing marked as completed that didn’t deserve logging as events to compare analytics to. As another metric, I wrote 239 blog posts–about 5 a week on average, during 2012. In addition, there was more content produced, such as detailed tutorials on the two areas of industry practices that my two software products are designed to help out with:
- Feeds and Speeds: This term originated with manufacturing and machine tools, though High Tech embraced it too.
- G-Code Programming: G-Code is the “assembly language” of computer controlled machine tools.
Content marketing aficionados will recognize these latter two as the sort of “Evergreen” content that is extremely valuable in terms of SEO and generating a solid foundation of site traffic.
Early in the process of starting CNCCookbook, I decided Content Marketing was the only way to go. As a bootstrapper, I didn’t want a business that required lots of expensive advertising because I didn’t want to invest the capital. I had the opportunity to try my Content Marketing theories back before they were as fashionable as they are today. I used both CNCCookbook and my day jobs at the time, Helpstream and Hotchalk, as the test vehicles for the strategies. They worked extremely well for all 3 companies, which is a testament given how different they are: CNC Manufacturing Software both B2B and B2C, SaaS Social CRM/Enterprise Software, and Online Education.
I had a real leg up with the CNCCookbook web site, because I knew I had already achieved a Content-Audience Fit and that I only had to learn how to optimize the content machine further. My Marketing philosophy was that I would give away tons of valuable content (or at least content that is as valuable as I could make it, but the numbers show my audience felt it was valuable too) with little or no marketing “spam”. I do not spend a lot of time asking for the sale and the content on my web site is designed to attract any machinist, not just machinists who are ready to buy my products right away. As a result, I get to build a relationship and some reciprocity debt with my audience before I go for the sale.
The sales funnel works like this:
1. Attract people to http://www.cnccookbook.com by offering valuable content that scores well in organic search.
2. Priority #1 is to get them signed up to the mailing list so they can continue to receive the valuable content in weekly blog post digest. At the end of 2012, we had about 16,000 names on our mailing list. I get very decent open rates and site visits each Wednesday when I send out the weekly blog digest. About 1/3 to nearly 1/2 of the list will visit as a result.
3. Visitors in this active “content community” become aware of our products through content. We report all product progress via the blog. When solving any problem, mention relevant related capabilities in our software. Report periodic special offers in the blog. They also become aware through various “ads” that appear at the top of the screen (Hellobar), bottom of screen (homegrown slideup banner), and in the navigation panel on the left as well as the “Software” menu at the top of every page and the site’s Home Page. All of these vehicles direct users to the various home pages for our software where they can sign up for a free 30-day trial.
4. During the trial, we use an email drip campaign to try to make sure users are coming up to speed. The email sent is intentially kept friendly and conversational and is directly from me. I try to avoid making it too “slick”, and I try to make sure everyone gets what they need to come up to speed and be as productive as possible during the trial.
5. The product itself also has a number of features designed to enroll folks into the community and help them along. I’ve discussed these in a prior blog post.
6. The product and email drip campaign will countdown at the end and present the trial user with links to purchase.
7. Whether or not they purchase immediately, we work hard to keep them on our mailing list. We run sales periodically, about every 6-8 weeks, that substantially increase our close rate (some folks will only buy at a discount), but we don’t want to run them so often that people only buy at the sale prices. I think this timing, which was worked out via testing, has served us well.
To date, about 16,000 have participated in the product trial. 13% will end up buying, a number that’s pretty high for software free trials. The software is sold via SaaS-style subscription at $69/year, and renewal rates are pretty decent, although we haven’t been selling long enough to see very much of that yet. Sales in 2012 were up about 120% versus 2011.
The software side is interesting too, but I’ll save the technology details for another post sometime.
Going Full Time on CNCCookbook: Lessons Learned
I went full-time on CNCCookbook July 1, 2012, so I have 6 months into it. The company throws off sufficient revenue that I can pay myself what I would make in salary at a startup. Typically, my position would either be VP of Engineering/CTO or CEO, and I’d be paid at the upper end of startup salary scales, so I’m pretty happy with CNCCookbook. Our revenue grew 150% plus in 2012, and if it keeps on like that, it’ll be a very happy experience indeed. I own 100% of the company, and have taken no outside investments. I started actively trying to make it a business in mid-2010, at the 37Signals recommended pace of 10 hours a week. I went full time after 2 years, and reached salary-parity with a Day Job, or Escape Velocity as I like to call it, 2 1/2 years into the exercise. I love the business, products, and customers and I’m very pleased with the path I have chosen, so much so, I wonder what took me so long.
I’ve just started working part-time with some developers who I hope will ultimately join CNCCookbook when we’re large enough to be able to hire them. Meanwhile, I have a 10-hour a week plan for them to participate, and that’s worked well.
In no particular order, here are some lessons and thoughts I have for others who want to try something similar.
Things That Worked
- I had the luxury of starting this company having already achieved Content-Audience Fit. I knew quite a lot about my audience, I knew what content they were interested in and hence what problems they wanted to solve, and I even had quite a lot of empirical web analytics data to go on. In hindsight, I wouldn’t call all that a luxury. I can’t understand how to do a decent job bootstrapping a company without it. Anything else and you’re just building a product that you’ll throw over the fence and hope someone will consume.
- I focus relentlessly on spending my time doing things that nobody else can do for my business. If I can buy a solution, preferably SaaS, I will do that rather than build it even if I know I could do a better job. That means I bought my e-commerce solution (1ShoppingCart), customer service (UserVoice), hosted wordpress blog (Page.ly), and a lot of other things.
- I’ve taken a radically different approach to the back-end of my SaaS software architecture, which I call “Fat SaaS”. In essence, I put as much intelligence as I possibly can into the client and keep the back-end very simple. It all runs on Amazon Web Services, is highly scalable, very reliable, and didn’t take much effort to build. Eventually I’ll write about it. The thing is, focus your efforts on what your customers can see and be delighted by. The rest is just plumbing.
- I am firmly of the opinion that our high close rates are due to not only the quality of the software, but also because we don’t use the hard sell. The trouble with a hard sell is that while it may increase close rates for those ready to buy, it can so turn off those who are much earlier in the funnel that they won’t return when they are ready to buy. Having seen the process at a lot of companies, frequently they’re not even measuring the impact on the later tail of prospects. They’re only interested in what’s closing in the next 3-6 months. I still see significant sales even 1 year after the end of the free trial.
- We appeal both to hobbyists (B2C) and businesses (B2B). This sounds odd and unworkable, but for the community of machinists, it works fine. Based on various surveys I have done, we seem to have about 60% professionals and 40% hobbysists. The professionals will pay more and require less support, but the hobbyists get the word out in the various online communities. The reason I can serve both markets is the participants are passionate about machining. I believe passion is an essential litmus for content marketing success (and likely bootstrapping success). If you can’t detect passion in the audience, how can you get them interested in your content? How can you get them to tell others about it?
- Only having 10 hours a week for the first 2 years was beneficial. When you’re learning what to do, having a lot more hours available doesn’t necessarily help. It takes time for the ideas to percolate, for the right influences to be felt, and for the results to be measured. Heck, it takes time to think of all the ideas. You don’t have them all up front so that all you have to do is implement them as rapidly as you can. And of course, limited resources maximizes your focus.
- Using an Agile Methodology to run everything including the business and marketing side has been hugely beneficial.
- I have never owned or touched a server for this business. It’s all been in the Cloud from Day One. It started at LunarPages, my original El Cheapo ISV, and fairly quickly migrated to Amazon. Today, the site is entirely on Amazon. Static pages are hosted via S3 and we use the Page.ly service to host WordPress. My order of preference in implementing a service is:
- Defer it. Do you really need it now?
- Use SaaS. They do it for a living. Are you really going to do a better job or is it a distraction?
- Outsource it.
- Build it. Least desirable. If I build it, it better be highly differentiated and valuable to my audience.
- I farmed out the simple back end we use for authentication so I could focus on the machinist-specific client.
- There is a tremendous art to using Sales (by that I mean special prices for a limited time) to enhance your revenue and close rates without trashing your average selling price. Learn to do that as early as you can through experimentation.
- Measure everything. My primary measurement tool is Google Analytics. But, I get lots of data from SEMRush, SEOMoz, SurveyMonkey, and the various telemetry I’ve built into my product. All the key data goes into a weekly marketing report as a row in a spreadsheet. I focus entirely on the trends and let the absolute values work themselves out over time. If the trends are right, you can’t lose. There’s no more powerful force than compound interest.
- Drive the business with the metrics. If you can’t measure it, it isn’t real. If you don’t have an Analytics + Metrics + Decision Making Machine, you don’t really have a scalable strategy, you have a bunch of gut ideas that may or may not work.
- Taking on 3rd party software to resell. See my note below about the Road Ahead. This has worked extremely well for us and added about 30% to our top line revenue for very little effort.
Things I Wish I Had Done Differently
- Our first product was easier to build, but was too cheap. At $69 a year, you need a lot of customers before you can quit your Day Job. Being able to charge even a little more would’ve accelerated the process, assuming we could still get it sold. One of the problems is that being a solo bootstrapper, there were real limits to what I could produce, especially while having to learn all the other non-technical jobs. A bootstrap team of 2 or 3 could attempt a much more ambitious first product and probably get to full-time viability sooner.
- The scalable formula for growing a Content Marketing effort is simpler than I thought. Over time, Google has made it harder and harder to play SEO tricks. The more I wanted to find the Tricky Secrets, the more time I wasted. This is one reason why I eventually limited myself to working on marketing after 7pm at night. In mosts cases, only three things have mattered: Doing good keyword research. Doing enough A/B testing. Producing great content. Figure out what formula works for you, and as soon as you have a scalable formula, focus on scaling it. Save time for a few experiments, but not too many. You don’t have enough attention to split too many ways. There is such a thing as Minimum Viable Marketing and it is just as important as Minimum Viable Product.
- I made some bad choices on various services I signed up for. This was mostly a function of being largely review-driven and not having a clear enough idea of my long-term needs. The heart of the commerce platform for CNCCookbook is email. Initially, I chose a service called 1ShoppingCart because it integrated everything I thought I would need in one package. 1SC turned out to have a very poor set of API’s and was better suited to sales of physical goods than SaaS software. I’ve made it work, but at this stage, mostly as a shopping cart. After having looked at a lot of email platforms, and gotten a pretty deep false start with Constant Contact, I am finally happy with MailChimp. I would’ve saved myself a lot of trouble if I had started everything from MailChimp and built out around it. Spend time reviewing the API’s of any service you sign up for. You never know when you’ll want to dig into them.
- I started with a static web site for historical reasons. CNCCookbook was originally formatted much differently than a blog so it didn’t occur to me to use blogging software. Today, I have integrated WordPress and love it. If I had it to do over again, I would probably have simply built everything on top of WordPress.
- Social Media are helpful, but deliver a lot less marketing value than you would think. Automate your use of them and let it ride.
- I got started focusing on my mailing list late. There was no way to sign up for the list or even get a real RSS feed until the beginning of 2012. This was a huge mistake. A good mailing list is as much an asset as anything else your business will have. Treat the people on it well by not wasting their time and you won’t be disappointed.
- Our products are not a pure SaaS model. When the subscription expires, the product continues to deliver considerable value, though far less than if the subscription is paid up. The jury is still out on whether I gave away the cow instead of selling the milk on that deal. I am happy with renewals, and it may be that without this feature I would’ve had far fewer sales to hobbyists, so I can’t truthfully say I would do this differently. But, it is something I think about a lot.
The Road Ahead
CNCCookbook’s biggest problem is lack of product. We have a fantastic readership, traffic is growing steadily, and the word of mouth around our first product and the web site is great. It’s become the de facto standard and we’ve even signed up a bunch of big name manufacturers who use it. Our problem is not having enough software to sell into this channel. As soon as I realized this, I went in search of 3rd party software I could resell to this audience and was successful with a couple of packages during 2012. I would guess they added about 30% to our revenues. 2013 will be the year we ship our second product, and hopefully also our third and our fourth products. My suspicion is that this will radically scale the business along every dimension and I’m really looking forward to it.
I was pleasantly pleased to discover that once you have a one man SaaS business, you can run it from virtually anywhere in the world that has a decent Internet connection. All my servers are in the Cloud, and all my tools will fit on a laptop. Consequently, I’ve had the pleasure of being able to run the business from the cabin of my Alaskan Cruise Ship, from a rented condo on a dive trip to Cozumel, Mexico, and from my hotel room in Waikiki. It’s been a real blast and I’m eagerly looking forward to what comes next.
(Cross-posted @ SmoothSpan Blog)