Gaining the Wisdom of Crowds in a Bootstrapped SaaS Company
Posted by Bob Warfield on November 19, 2012
When you’re bootstrapping a small company, sometimes it’s hard to do the things larger organizations take for granted, like making sure you’re listening well enough to your customers. On the other hand, you can take advantage of your nimble nature and the availability of some great technology to do some things that even a lot of larger organizations don’t manage to pull off. At CNCCookbook, my small Manufacturing Software company, I’ve had to think long and hard about how to register the wisdom of my Crowds to make sure the company is on the right track with its products. Lest you think small companies with fewer employees than you can count on one hand don’t have Crowds to learn from, CNCCookbook gets over 1 million visitors to its site every year and we’ve had over 15,000 machinists use the software to date. We count some of the world’s largest manufacturers on our Customer List as well. In short, there’s plenty of Wisdom to be had from our Crowds, it’s a matter of finding the right ways to capture it and put it to use.
Having come from a Social CRM background at Helpstream, the value of harnessing the Wisdom was not lost on me. It was something that had worked well for me throughout my career and something I very much wanted to do well with at CNCCookbook. Here is a brief history of how I went about it and which tools, techniques, and technologies were put to work to do so.
Phase 1: Forums and Web Analytics
Right from the very start I deployed a set of User Forums which I called the “G-Wizard User Club” (CNCCookbook is our company and web site, G-Wizard is the software brand that labels our products). Much as I miss the sophisticated capabilities we had at Helpstream (they haven’t been rivaled by any product since), I had to make do with what was available and what fit my budget. I knew I wanted a SaaS-based service. However easy it might be for me to install and administer phpBB or some other Open Source bulletin board, it would be one more thing for me to do. As the sole person working in the company at this time, I made the decision to focus as much of my time as possible on things that were uniquely differentiated for our company. Deploying phpBB wouldn’t even come close, so I went with an alternative that was both a SaaS service and ad-supported called ProBoards.
It has worked reasonably well, and served its purpose. I moderated membership and got a lot of mileage out of the boards. They continue to be popular to this day, and we have not quite 2000 members there today. To make sure every User was aware, I also instituted an in-app button to take open the browser and take them to the User’s Club.
You can see there’s more than just the User’s Club there on that Login Bar, but it started with just the User’s Club and grew to encompass a number of resources every User needs to be aware of. While our app doesn’t run in a browser (it’s an Adobe AIR app as disconnected running is often important to our audience), it behaves in every other way like a browser-based SaaS app and we have embraced a lot of the design concepts for such apps, such as seamless access to the important parts of our web presence and incorporating that presence as a first class citizen of our navigation structure.
Another critical source of the Wisdom of Crowds is your Web Analytics. We use Google Analytics, and there is a wealth of information to be gleaned. For example, our User Guides are entirely online and we can see from the Web Analytics which parts of the product are more interesting than others just by watching the traffic patterns. As we do each new release we write a blog post that discusses the new functionality in the release and again this provides a framework for using Web Analytics to understand what’s going on with the product.
In app access to Getting Started Resources, our Support Portal, and the User’s Club Forums…
Phase 2: Blog Comments, Social, and Surveys on the Web Site
CNCCookbook started as a plain old web site and went for quite a while like that. We had an area where articles were presented in a quasi-blog format, but it wasn’t really a blog. It didn’t take long before we’d outgrown that format and it was time to add a real blog based on WordPress. If I had it to do over again, I would recommend that every company simply start with WordPress and eschew the plain old web site phase. It’s a fantastic content management system that has a rich ecosystem supporting it. In keeping with my SaaS philosophy (why would I spend my scarce time maintaining a commodity like WordPress instead of focusing on what makes our company different?), we signed up with page.ly to host WordPress for us. We spliced the blog into our plain old web site using DNS Made Easy, a SaaS DNS service that’s been excellent.
This transition marked a big step up for us in a whole lot of ways. There were obvious SEO advantages that were very visible in the Google Analytics reports. It became much easier to manage our content and we did a major upgrade to the site’s look and feel (it’s getting close to time to do another, I think). Best of all, we now had comments on every post and could deploy a host of social widgets to help harvest as much feedback from our audience as possible. One of the first things I did once WordPress was up and running was to go out and survey key sites to see what sorts of plugins they were using with WordPress. My approach was to use a variation of a Blackjack card counting strategy I had perfected to decide my Social Widget strategy. I’ll say more about the Blackjack in a future post, but suffice to say I analyzed the widgets used by a number of top marketing blogs on the theory that these people should know. I went to companies that clearly had lots of experience with conversion and A/B testing like Unbounce. I went to specific marketing gurus like Neil Patel’s Quick Sprout blog. It was an excellent way to focus my efforts and populate the CNCCookbook blog with what I think are an excellent set of Social plug-ins to maximize engagement.
Having done that, I turned to Surveys. While it was kind of an expensive luxury, I bought two different tools. I wanted a survey tool that would be innocuous and unobtrusive. I hate visiting a site and getting hammered with a full stop “please answer our survey” ten seconds after I get there. At that point, I have formed no opinion but a negative one about the damned survey. At the time, KissMetrics had an awesome tool called KissInsights that would slide up from the bottom of screen in a very low key way. That tool is now sold by Qualaroo and works great. It’s biggest issue, and the reason I don’t use it for all my surveys, is it is limited to simple surveys. So, I also subscribed to SurveyMonkey.
I use the Qualaroo tool to derive a Net Promoter-style feedback score on the overall product (ours is very high) and I use the Survey Monkey to do more detailed surveys aimed at understand the details of my audience. For example, I have done surveys of which CAM software they use or which CNC control is on their machines. Not only is this invaluable data (sort of like surveying which PC, OS, or browser a PC software audience uses), but it makes great content to publish on the blog. Some of my all-time best traffic articles are just the results of such surveys. Apparently others also want to understand the Wisdom of Crowds.
Phase 3: Ideation and CRM
For Phase 3, I wanted some Social and Conventional CRM. It was time to get a Trouble Ticketing system going. I chose a vendor called UserVoice for several reasons. First, it comes with a very nice Ideation App. Ideation gives my audience the ability to suggest new features and vote on them, like Dell’s Ideastorm. This is an extremely powerful capability for a small organization to use to focus scarce development resources. The results will often surprise you. Ideation is one aspect of what we had at Helpstream, so it was nice to get some of that back. Second, it’s SaaS. And third, I got a great deal on it via AppSumo. BTW, AppSumo has yielded several good deals for my bootstrap venture. I’ll warn you in advance, they’re very spammy in their email and you really have to know what you’re looking at when you consider the products they push, but if you are patient about wading through some spam and have a clear idea what your business needs, you’ll find some great deals to keep the overhead down.
One of our products, G-Wizard Calculator, is much more mature than our later products because it has a 2 year head start on them. While I still have a lot of ideas about where I want to take that product, it has a solid conceptual foundation. What I mean by that is that it is ready to be steered to a much greater extent by customers. Ideation tools are a great way to do this as they force customers to ration their votes. On our site, they get to use 10 votes, and can vote no more than 3 votes on any given idea. Submitting a new idea uses up a vote. Once the votes are used, they have to wait until the fait of an idea is decided, they are either implemented or rejected, at which time they get the votes back, or they can redeploy the votes. This scarcity of votes gives a clearer signal of what really matters to your tribe. Any time I am preparing to do a new release of the Calculator, I always start with our Customer Support Portal and look over the Ideation results.
Phase 4: In-app Feedback and ET Phone Home Telemetry
This brings me to our current stage of evolution–In-application Feedback and Telemetry. In keeping with our theme of making the product behave like a web application, we added a Beta Survey popup such as you see above. This has been a very useful way to monitor our progress from Beta to release-ready. After spending 10 days focusing development entirely on issues raised in the Beta Survey, we’ve been able to move to 81% of respondents scoring the app during the last week as either “3 – I could use this” or “4 – GWE rocks!” For the period older than 1 week, the score was only 47%. Clearly, users were able to tell us what they needed that was missing from the app. We intend to continue for a while longer until we see a point of diminishing returns and then we’ll declare the Beta done.
In addition to the Beta Survey, we also receive what I call, “ET Phone Home Telemetry.” This is basic telemetry on which parts of the app are actually being used and how well they perform. For example, the centerpiece of the application is a complex 3D graphics simulation that shows how the machine tool cutter will move as it executes the g-code program loaded into GW Editor. We monitor and report back the longest runs so we can get an idea of how the system is performing and whether we need to do more work on performance. We also track usage information like how many times the user has logged into the app.
The technology that makes the in-app telemetry and Beta Survey easy is something called “Mandrill” that is offered by the MailChimp people. Rather than having to build back-end server infrastructure that loads all this data into some form of database using an API, the app simply emails it back to us with Mandrill. The volumes are such that it is very straightforward to collate the information in Excel for analysis. Building a full-on database application for a 2000 person Beta test would have been needless complexity and time taken away from our focus on doing what differentiates our software. Mandrill is what MailChimp calls “transactional email”. I take that to mean email generated by machines, rather than by people, and that’s exactly what we’re doing here. MailChimp has a Freemium model, and at our level, Mandrill is essentially free. Not only was it very easy to implement, but it doesn’t cost us anything. For bootstrappers, that’s a hard combination to ignore.
Just because you’re bootstrapping and have minimal budget and resources is no reason to ignore the Wisdom of Crowds. In fact, I’d argue that having the Wisdom of Crowds helps you to allocate your scarce resources where they will really matter. Towards that end, what we do differently at CNCCookbook as bootstrappers is build as little software as possible. We want to focus every line of code written on problems that you simply can’t get solutions for elsewhere. Problems that are unique to our audience of CNC machinists. The more of those problems we can solve, the more value we bring to our customers. Everything else is just overhead. Towards that end, we have relied heavily on SaaS, on the Amazon Cloud, and on our ingenuity to lash together the available off-the-shelf technologies to give us the ability to deliver an overall User Experience that is arguably better than almost everywhere I’ve ever worked. This despite every where else having vastly more budget and resources at their disposal.
I’ll give one last plug to SaaS and the Wisdom of Crowds. We do as much testing as possible, but again, as a bootstrapped organization, we don’t have large numbers of testers. Our software quality is therefore a focus of three things. First, unit testing is important. Whenever complex new subsystems are added to the software, we make sure there are unit tests. I personally believe in single stepping the debugger until I’ve seen all the lines of code executed and verified the intermediate results are good. Unit Tests not only help tee up the execution of all the paths, they also ensure that down the road we can validate intermediate results as changes are made. Second, we release often. I don’t like to change too many things without doing a release. This means that the amount of testing per release is relatively contained to new functionality and our scarce testing capabilities can be focused.
Lastly, we use what I call a “feathered” release methodology. Each time we release, there is a 7 day cycle. On each day, we expose an additional 1/7 of the user base to the availability of the release. Customers that insist on having the latest and greatest can change a setting so they see every release immediately, but most stick to the default. This ensures that if anything is too badly broken, we’ll hear about it before a very large fraction of the installed base is exposed to it. In this way, we’re also using the Wisdom of Crowds to help safeguard the quality of our software, and it has worked extremely well to date.
So, whether you’re a bootstrapper or a big company, think about how you could take advantage of the Wisdom of Crowds. Not only will it make a big difference for your software, but it’ll show your audience that you care and that they have a voice.