There’s no shortage of competition between home improvement giants Lowe’s and The Home Depot, that’s for sure. Today, we thought we’d check out their websites and see how the two retailers stack up in the e-commerce arena. Since there’s just too much to sift through on both websites, we’re only going to explore a couple of elements: the search function, and the sorting and filtering options. I decided to search Lowe’s and The Home Depot for an item any homeowner needs in their arsenal: a ladder.
Let’s start with Lowe’s. As you type into the search field, it starts generating suggestions. It’s really cool how it suggests both broad categories, as well as specific products below that. The suggestions even include small product images and brief descriptions. I’m curious as to why all of these suggested products are the same brand, and I wonder if they are paid sponsors or best sellers or what… That aside, it’s a great way to start a search!
I clicked on the first category suggestion of ladders. There are some good sorting options along the top of the page, a comparison feature, and shoppers can even choose to view the results either as grid or a list. These things are all great. However, along the left navigation, the only filtering option is to choose a department: Building Supplies, or Outdoor Living & Recreation.
At first, I was a little puzzled as to why I didn’t have more filtering options to choose from. It turns out that they’re hidden, and you have to first choose a department before the site will display the next set of refinement options. Once I clicked on the Building Supplies option, I got a whole new list of helpful filters.
These refinements, like rating, price and brand, are definitely good, but they’re pretty generic and not specific to ladders. For example, ladder height would be a good option to add. Also, I wish these filters weren’t hidden on the initial results page. It makes shoppers have to take an extra step to refine for what they need, and some people might miss it altogether. I would suggest making all the filters available right from the beginning.
Now let’s see what The Home Depot has to offer. This search function is similar to Lowe’s in the way that it begins making suggestions as you type into the field. The drop-down gives you the option to shop by category or brand, or even click to to project guide videos and a ladder buying guide.
That’s a lot of options, but I just clicked on the Ladders category, like I did at Lowe’s. It brought me to a page showing all the different categories of ladders, complete with generic product images and the option to choose the material. I like this because if a shopper isn’t quite sure about what type of ladder they need, they can refer to the pictures as a guide. Even though it does hide many of the other refinement options until the next page, I think this is a great way to point customers in the right direction.
Since I had to choose a category to move forward, I selected Step Ladders. When I got to the category page, I found sorting and filtering galore! The left hand navigation had tons of refinement options specific to ladders, including a few that Lowe’s doesn’t offer, like height, ladder material and the ability to set your own custom price range. I also really like how the tabs along the top allow shoppers to choose to view all the products, the ones that are available online only, or the ones available in their nearest store.
This e-commerce face-off is a pretty close call, especially since the two search functions are very similar. However, The Home Depot comes out ahead for their refinement options. Sorry Lowe’s, but it’s like you say: Never stop improving. That goes for your website, too!
3 Comments
The reason Home Depot has that super-cool megamenu is they’re trying to get you to category-based results, not searches. Searches are always potentially noisier, and subject to polysemy ambiguity. Why does this matter? Well, you know more about context and which facets to show when there’s no ambiguity. So you can show the detailed facets in those.
For example, if you show all facets for search, and the word “ladder” appeared in a book title “The Ladder to Success” or “How to Make Lots of Money While Sleeping by Robert Ladder” you could get facets for publishing date, author, etc. Whether this is a pervasive issue for Lowes, I’m not sure. It’s not in this particular case, but I’m sure these cases exist.
Ultimately, had you drilled _twice_ to the category ladders, you’d have seen the facet “ladder length.”
Actually, you didn’t, I think. Lowes looks like it’s running a search, which will _always_ result in potential polysemy ambiguity. Uh oh.
This all comes down to signal vs. noise. They might be choosing only to show basic facets on the search page to _encourage_ you to drill, thusly making your intent very clear. The problem is, that doesn’t really work 100% here because not everyone will realize they must drill twice. It will crack it much better on a shallower taxonomy. That a user will drill once (indeed you _did_ drill once) is a much more likely proposition.
But Home Depot avoids this problem with that fancy filterable menu that encourages you to get there in 1 step!
So they could have done much better with Home Depot’s filterable megamenu thing, or even by sending you to Category:Ladders instead of Search:”Ladders.” Home Depot’s approach is the better of the two, since it won’t try to preclude a search if the category really is not what you want — just because it’s a category doesn’t mean the problem of polysemy goes away. So it may be misleading to display “ladders” and assume that they want the category. That would be wrong on Amazon for sure.
Make sense? Win for Home Depot.
The other variance to your comparison, is that on Lowes you did a SEARCH, and on HD you were essentially browsing the actual categories.
Although not always the case, some sites have a different filtering option when viewing results from SEARCH and BROWSE.
If we dig deeper into the two different technologies each result set used…
Lowes leveraged Google Commerce Search and the Search As You Type to generate the results.
Where as HD leveraged Endeca Page Builder to generate the result set on the Category Landing and Product List Pages.
Just a thought.
@RD
Hat tip. There are a few reasons (maybe more) for this:
1. Short key phrases (that’s most of them outside of a legal/medical research setting) are subject to ambiguity. You don’t necessarily want to show “ladder height” for a query that might actually be for a book. It can be confusing or seem like clutter. This is especially relevant for 1 word queries, and especially relevant for department stores. Big box stores a little less so — they are more focused and “ladder” means only one thing because they do not sell books.
Note that one could argue that sometimes it’s OK to show because it might implicitly disambiguate when selected via correlation. But it’s a potential issue.
2. Even then, if you have 4000 detail-oriented filters, it can be expensive to count them regardless of the algorithm you use to compute said counts. Google doesn’t show exact counts of documents for a reason. Counting facet values is much worse than just a document count subject to 1 query because you then have to count for various filter states on said query. It’s said computers are good at counting, but that only goes so far. There are a few ways to count, and the best way can depend on selectivity, number of values, sparseness of application, etc. However, at the end of the day, it’s still expensive.
Anyway — driving users to the relevant category will always yield the best result. Signal to noise ratio is always best there. That’s why Home Depot tries to drive you there I would surmise.
Lowe’s could probably publish some of the detail-oriented filters on the 2nd level of their taxonomy rather than wait for total disambiguation at the 3rd. That would be a compromise and mitigate the ambiguity. I doubt performance is an issue, but it would also address that.