The Impact of “Lean Startup” on Go Daddy

At Go Daddy, we’ve long believed in Agile values:

Individuals and interactions over processes and tools.

Working software over comprehensive documentation.

Customer collaboration over contract negotiation.

Responding to change over following a plan.

As the Agile Manifesto will tell you, it’s not to say that we don’t value the latter, just that we value the former more.

We have a wide product line at Go Daddy and are constantly evolving all of them. We have feedback loops from customers and rolling roadmaps that change to adapt to what our customers want. The bottom line is, we listen to our customers. This has been a way of life since the company’s inception.

“Lean Startup” is a term recently popularized by Eric Ries in his book, “Lean Startup.” This term and the book have certainly made the rounds here at Go Daddy. The Lean Startup is about testing your assumptions as early as possible, using customer feedback, and iterating quickly through many prototypes to learn what the market wants.

An adaptation that we’ve recently implemented goes a little bit like this (in the Agile Manifesto’s “___” over “___” format):

Testing assumptions over listening to your customers.

Now you’re probably saying to yourself, “…over listening to your customers!? Are you kidding me?” But, before you dismiss this value, let me explain. As I said earlier, it’s not to say that we don’t value the latter, just that we value the former more.

Let me give you a real life example. Have you heard the story about Walmart listening to its customers? Walmart customers were asked in a survey if they’d like to see the aisles less cluttered, and they said yes. So, Walmart stores cleaned up their aisles and reduced inventory, spending millions of dollars in the process, and their customer satisfaction increased. But, a strange thing happened; their sales plummeted by nearly 2 billion dollars. As it turns out, what customers said they wanted didn’t match their buying patterns.

For Go Daddy, “Lean Startup” inspired us to think differently about new product development. We’ve adapted this philosophy, influenced by research through a 3^rd party consulting firm, and came up with a philosophy that says get “customer feedback” by measuring their behaviors, not just by asking them what they want. Get your product to market faster. The faster your customers are using your product, the sooner you can observe their behaviors and make modifications. And, test your way to correctness.

If we build instrumentation to measure customer behaviors, then we can roll out new features and monitor their actions, rather than relying solely on what customers say they like or want.

This sounds great for new web-based application products. Build the “measure what they do” functionality from the start. Get it to market fast. Don’t worry about it being perfect, early adopters will show you, by their behaviors, what they like and don’t like. It’s okay to listen to them too, as long as you can test your assumptions first.

Here are a few real-life examples when Go Daddy employed this technique:

Express Email Marketing Billing

With our email marketing product, we theorized that charging on a per subscriber basis instead of per email basis would be better received by our customers, lead to more sales, and prompt more renewals. We changed our sales landing page and are measuring the difference in conversion rate. We’re also testing our assumption by changing the billing model on incremental sets of existing customers and measuring renewal rate.

Search Engine Visibility
When we rolled-out our latest spider for Search Engine Visibility, we used the modulus of 10 on the website ID to identify 10% of our customers to start using it. After a few days with no problems, we switched it to mod 5 to get 20% and then mod 2 to get 50% and then mod 1 to get them all. This incremental roll-out allowed us to monitor customer reaction and make sure that our new spider didn’t introduce any unforeseen consequences.

User Event Logger
Customers who activate products are far more likely to renew. So, any product with a lower than normal activation rate needs attention. One thing we can do is message customers in different ways to encourage them to use or activate the product they purchased.

To properly discover activation rate, we needed to define what activation means. It’s different from product to product. For example, with an email hosting plan with 10 addresses, is it activated when one email is setup, or all ten? For Website Tonight, a website builder, is it activated when they agree to the EULA, or when they publish their site? Our first step was to define, for each application, what activation means.

Next, we built a centralized internal service to log any customer event statistic that we want. It’s easy for our application developers to send events via an API. They may then correlate this data with customer persona data to determine what types of customers perform what types of actions.

Website Tonight now logs “signed the EULA” and “published site” events, among others. If we were to find, for example, that more people cancel after signing the EULA, but before publishing their site, what would that tell us? If we ask customers what made them cancel, only a small percentage of them will answer and we’d be making decisions based on a limited amount of information. Instead, we observe, measure, and log what all customers do, correlate that with data from our cancellation surveys, combine with direct feedback from customers, and then we can make very informed, data-driven decisions.

With our User Event Logger, we’ve created an internal service that enables all of our applications to easily capture what customers are doing. This will lead to data-driven decisions on features and functionality, which will improve customer satisfaction, which will lead to more renewals and sales. Now, isn’t that much better than asking a few customers and assuming this sample set is representative of the whole?

Product Name Split Testing
We’ve also been doing Split Testing on our main www.godaddy.com site for years. We’ve recently improved this system to rapidly test product names. Our Marketing team can set up a split A/B/C test on a product sales page and test the click-through rate to the shopping cart, on up to 3 product names at a time. Our system determines, for each visitor, whether to use product name A, B, or C.

This has led to the following name changes:

• Online File Folder is now called Online Storage
• Virtual DataCenter is now called Go Daddy Cloud Servers

Conclusion

We listen to our customers, and always have. That’s been a cornerstone of our success. Taking care of our customers has always come first and will continue to come first. ”Lean Startup,” by Eric Ries, was the catalyst to our new philosophy. We adapted it and realized we get as much from observing our customers as we do from listening to them. So, we do both. We’ll continue to listen and improve our systems, and the result will be even better value, further enabling individuals and small businesses to establish, maintain, and evolve their online presence.