In my previous two blog posts, I discussed the challenges that the domain industry is undergoing based on the increase in domain extension options and the growing percentage of domain extensions with a specific “TLD Focus.” Now it’s time to talk about how registrars and registries should be responding to these changes. I’ll start by focusing on registrars in this post and cover registry response in my next post.
The approaches and strategies registrars will need to employ in response to this change boil down to one thing: merchandising. This term may be unfamiliar to a number of domain industry participants so here’s the definition from Webster’s:
Merchandising: sales promotion as a comprehensive function including market research, development of new products, coordination of manufacture and marketing, and effective advertising and selling.
When there are only a handful of domain extensions that apply to most customers, you don’t have to think about how you merchandise these domain extensions much at all. However, the increase in domain extension options is a huge shift that requires registrars to make merchandising a primary focus rather than an afterthought if they want to make the most of this opportunity.
There are very few industries that have undergone a change of this magnitude, but I’d like to present an example of a business that forced a change like this on itself: Amazon. Most of us can still remember Amazon when it was a simple purveyor of books and music. Over time, Amazon has added what feels like an infinite number of products across an equally impressive number of categories. To successfully navigate this transformation, Amazon was forced to become one of the best merchandisers in the world. There are lots of areas we could discuss related to this–upselling, cross-selling, bundling–but we’re going to focus on discovery since that’s the most fundamental aspect of merchandising. Let’s take a look at how Amazon’s interface had to change over time to accommodate all of the new products and categories.
You can see the steady progression in the top-level navigation as Amazon worked to surface their primary product categories in order to focus the user’s search. This approach would ultimately be abandoned due to the exponential growth in the categories they supported. So what did Amazon do instead? They became experts at product discovery–the art/science of combining searching, browsing, and product placement to get potential customers to the right product.
When you search for “taco salad” on Amazon the results look like this:
On the far left, there are product category listings that allow me to browse the search results, in the center are suggested products based on relevancy, and on the far right are ads. At this point, very few, if any, registrars are able to merchandise with this level of sophistication.
At Name.com, the standard search experience has none of the elements you’ll find on Amazon; instead, search results mostly consist of a set of domain extensions that are pre-selected and not based on relevancy to the search query. A search for “taco salad” produces results that look like this:
What do .NEWS, .NINJA, .SALE, and .CO have to do with taco salads? Not much. The domain extensions shown in these results are essentially billboards, and just like the ones you see on the road, they aren’t personalized for you in any way. There are two relevancy-based results towards the top of the page, and they’re the best suggestions on the page. With this in mind, we’ve been testing relevancy-based results, essentially replicating the center section of the Amazon results. In our test, the results look like this:
In these results you can see we’re doing a better job of suggesting relevant domain extensions, and we’ve started to offer different options to the left of the dot (with mixed results). Are the results awesome? No, but the important thing is that we’re testing and learning. Why? Because by figuring this out, we think we’ll drive higher search to registration conversion rates and higher revenue per search session.
This type of incremental learning and improvement is how Amazon transitioned into the successful company they’ve become today. (Did you hear that they recently became more valuable than Walmart?) The metamorphosis that Amazon went through is the same one that registrars are faced with. So, how do you get started?
We’re gonna need a bigger boat
The first step is a merchandising self-assessment. Do you have merchandising expertise on your staff? Does your current technology support your merchandising needs? At Name.com, we did this and realized we came up lacking in both departments. In response to that we’ve opened a merchandising role on the team, and we’re building our own domain discovery platform, which powers the test mentioned above.
As we continue to improve our merchandising, I’ll share details of the steps we’ve taken and results from our tests, both successes and failures. In the meantime, I’ll leave you with a collection of things we’ve learned so far:
- The Bare Minimum: Make sure your domain discovery solution handles new domain extensions reasonably well. At minimum, it should return them as results (I’m looking at you Verisign Suggestions).
- Exact Match is Great: Users like it when your search results exactly matches their query (and with the availability of new domains, that’s more possible than ever). So spin selectively, your focus should be on matching the right domain extension with the customer’s search. One exception to this is when the customer’s search matches a domain extension with specific “domain extension focus.” (Example: If the search is “Austin divorce” then there’s a strong chance that .ATTORNEY or .LAWYER will appeal to the user. Besides returning www.austindivorce.attorney, you should spin the search terms to present other .ATTORNEY domains, like www.texasdivorce.attorney.)
- Price Sensitivity: Price sensitivity is an important consideration. New domain extensions bring with them a large range of standard and premium price points. There is a lot for us to learn here, but here are two learnings we’ve had already:
- Customers searching for more general terms are far more price sensitive than those searching for more business or commercially-related terms. Using the previous example, the customer searching for “austin divorce” is more likely to pay a higher amount for the right TLD than someone searching for “water balloons.”
- Using a pure relevancy approach can result in presenting too many expensive options due to the number of premium names in market. In our next test, we’ll be trying an approach where we have a set of relevant non-premium results and a separate listing of relevant premium results. Per my note above, we’ll let you know how this works.
Thanks for taking the time to read this post and I’d love to hear from you on Twitter, or via email (firstname.lastname@example.org). You can also follow me on LinkedIn. The LinkedIn url is a bit cumbersome so I use www.productmanagement.ninja as a shortcut because it’s more memorable and helps brand my professional web identity.