Last Updated Jan 18, 2011 1:48 PM EST
I started managing Tumbleweed Tiny House Company in April of 2007. We build tiny houses that range in size from 65 to 172 square feet. For the first five months, our sales averaged more than $30,000 a month.
Then the real estate and stock markets crashed. November of 2007 was our worst month, with just $2,300 in revenue. We hit rock bottom and had to lay off two of our four employees.
We knew that the construction market was in real trouble, and that we needed to find revenue elsewhere. That's when Jay Shafer, the company founder, and I decided to change our business model to cater to the do-it-yourself builder.
As part of that shift, we also decided to take our website, which had always been mostly an online brochure, and turn it into a serious revenue producer.
Our first task: Test the heck out of the website's design to figure out what works. Ultimately the changes we made were small but they had a huge impact. Our sales for the month of December 2010 reached $123,000 -- all of which was generated online. Here's how we did it.
The Tumbleweed brand was already established when we started the transformation. People knew about us, and the website got a lot of traffic. But the online sales for our "Small House Book" and house plans, as well as sign-ups for our building/design workshops were pretty minimal.
So we increased the site's interactivity, giving visitors the opportunity to post comments and exchange ideas. This encouraged people to come back more than once. We also worked to get even more traffic by implementing basic search engine optimization practices. That got us up to 2.4 million visits for the year.
Then I started to think that it wasn't about getting visitors to the site -- they were already there. The problem was in understanding how the visitors engage with the site. What they do when they get there? What pages do they go to? What are they buying?
This is where split testing, also called A/B testing, came in.
The way it works is that you make two versions of a Web page -- sometimes as many as five versions -- and you see which version gets the most engagement -- the most clicks. You have a specific goal -- whether to get visitors to click to the next page, buy something, or sign up for a newsletter.
For example, I may put up two different pictures of a house with a button for ordering our book. "Visitor Group A" sees one picture and button on its page, and "Visitor Group B" sees a different picture and button on its page. Whichever test page gets the most purchase clicks is the one I'm keeping.
Small changes, big results
We found that simple adjustments to the navigation bar at the top of our homepage had a huge impact.
Originally, our navigation bar showed "home," "houses," "plans," and "book" listed from left to right. We eliminated "home," and then went with "houses," "Small House Book," and "plans," moving them all the way to the left in that order.
Just making those small changes doubled the sales of the book. Twice as many people were going to the book's Web page because the link buttons were now located in a higher-visibility area. Through our testing, we found that the further you move something to the top left of the page, the more clicks you get. Very small improvements can make a difference, and we've learned how to test the impact of those changes: switching colors, pictures, and where elements are positioned.
And we've discovered that word choice matters. For example, buttons in software's default mode often use the word "submit." But we've found that any other word works better than that. For example, just changing "submit" to "download," "join" or "sign up" can double the number of responses.
Before the economic crash in September of 2007, our success rate for visitors who bought our book, plans, and workshops was 1 in 600. After the crash, it was 1 in 1,000. Then I started doing the split testing; now it's about 1 in 200.
We have also used testing to increase sign-ups for our e-newsletter. The first year we had about 4,000 people sign up, the second year we had about 6,000. This year we've added 65,000 new readers. The conversion rate for signing up improved drastically from 1 in 148 to 1 in 28.
In 2009 we did $220,000 in sales. In 2010 with seven employees, we did $585,000.
-- As told to Peter Weed
Steve Weissmann's passion is building and developing a Web-based business with the hope that one day he can travel the world and work from anywhere.