So in my previous post I talked about one of my favorite companies – LDC.
Next up is Yipit. But this isn’t just about how awesome Yipit is. No no – today they launched profile pages, and in the process, further convoluted their home page. So this post is about how I would look over their site, and improve their UI and SEO.
To me, SEO is about giving the user what they want. In the case of a local search site, that relates to what they want (fishnet stockings, lava lamps, Company X, etc) and where they want it (New York City, Greenwich Village [neighborhood], 10015 [zip code], or some other informal space). It is also very possible that the user has a general idea of what they want but do not know how to exactly define it.
So – off the bat, the home page for Yipit is rather confusing. We have 1 2 3, where #1 is a non-standardized (in general UI terms) breakdown, #2 is slightly more confusing (it is pre-filled as NY – which doesn’t match either of their three criteria), and #3 is … something that should not need to be defined. Under it we have a list of Top Furniture and Top Discount Furniture stores, with a crap-ton of ‘popular searches’ that somehow relate to each business (but how?) On the right is a quick description of Yipit that seems stuffed with keywords, then twitter updates, then more popular searches, and then popular business searches.
What in God’s name of stuffing is this?
On a slight tangent – the English language can be hard. What may be X to me may be Y to you. Taxonomy is not an easy business. But the complexity in taxonomy is far simpler for Yipit – by focusing on furniture (more later), the verse of words people could be possibly looking for is greatly narrowed. Instead of a drop down – just have a box saying ‘What kind of Furniture?’ Best of all – this lets them generate a list of keywords based on what users are searching for. Eventually they may learn that 75% of people use ‘fold-out bed’ while 20% use ‘sofa bed’ and 5% use ‘bed that opens up’. Sure they may miss it the first few times someone searches for it (as it would not be in their taxonomy) – but over time they would have this incredibly rich database of words people search for relating to their own internal category. Back to the topic – replace ‘Near’ with ‘Around (optional)’ and have it pre-filled with New York City. Then a simple search submit. All of them next to each other, no #s, and maybe a few examples under the two input boxes. Forcing their categorization on me was an unsatisfying UI experience.
On a related note – the ‘definitions’ link looks like it is clickable – but it sure ain’t. Confused me there.
So – now we have greatly cleaned up the main search part – no forcing a user to figure out what you think they want – let them define it, and no need for the ‘duh’ 1 2 3 #s.
Next up – we have a bunch of ‘Top Furniture Stores’ under it. First off – all the profile page links go to the generic /furniture__ny/ page – I assume that is a bug. Second of all – why do we have 900 keywords stuffed in there? Having ‘new york’ stuffed in every single result leaves me with a bad taste. You have tags for each business. You should be using those.
But never mind the tags even. You guys have two objectives here – get people to use the Yipit code, or get people to find Yipit when searching for a popular key phrase / store. At the same time, Yipit’s authority in the eyes of search engines is still weak – it is new, and does not have a few links. So – instead of 12 businesses across two categories, list 4. Two are your advertisers, and two are either the two most popular furniture stores in NYC, or two advertisers. Ideally this would let you track how effective the home page is in getting people to use the coupon, and thus the value it generates. Furthermore, just list the tags you ahve for each business. No need to put in all these extra popular searches.
Now to the right side. The first paragraph is nice, the second paragraph throws me off with all that red text for links. Is it really needed? You already have popular searches – leave it at that. As for the twitter updates – honestly how many people care about that? You have a blog link in the footer – leave it at that. Leave those popular searches and popular business searches (though really rename it to ‘Popular Businesses’) – so your right is now a quick blurb on how Yipit does furniture (but should be slightly more fleshed out), and two quick lists on Popular Searches and Popular Businesses.
And bring the nice little footer image to the top please
So – now we have a home page much nicer and cleaner.
Next up – the directory pages in the footer. They are employing a Neighborhood directory, a Zip Code directory, and a Business Directory by alphabet. The Neighborhood directory expands into a bunch of pages for every category for every neighborhood, the zip directory the same, and the business directory just straight links.
Down the road this may work. But as I said before, personally I think that is stretching it too thin. Trying to run before you can walk. If I could run things, I would nuke the zip code directory, nuke the business listings, and only use Neighborhoods. The neighborhood-specific page would then have a list of just the businesses found in that directory. They have 337 in total – no worry of having too many in one neighborhood
The profile page I cannot complain about much. I would get rid of that Search Yipit drill down they have, and replace it back with the simple What/Where search box (with the Where pre-filled with that neighborhood). And instead of Other Searches of Interest, I would simply list the 5 closest Furnituer stores. Too many keywords (at this time) seem to be spreading yourself thin.
So we finally end up with the super confusing (to me) /d/xxxxx pages. Before I get started on it specifically, I noticed some links where /d/xxxxx_ny and some were /d/xxxxxxx_NY. That is duplicate content in the eyes of some search engines.
Now – the idea is smart. Search pages or tag pages – both are the same to Yipit. But that UI is just ‘what the fuck’ to me. At 1024 res that map on teh right is so squished. And way way too much information.
For example, I am looking at Murphy Beds. I fail to see why I care what the popular searches are on that page. I want to know name, info, and yes knowing their amount is nice. But popular searches are out. They could easily wipe out that entire blue box, giving the map a lot more space. Plus – you have a map – no need to put in the intersection. I think it should be just Name, Price (clickable with a tooltip to say exactly what $$$ means), address, phone, # of pieces, and a ‘more details’ link to their profile page! (profile pages are found nowhere on the search results page).
And as I said before – please just use a free-for-all input box for what.
I know I am not the only one confused by this page – the breadcrumbs lead to nowhere.
I would also slightly tweak the
Well – I could go on more, but I am hungry for some ice cream, but this should be enough of a start. My main focus is on cleaning up the site for information not needed (or applicable at that time), and also cutting down on the # of pages they are generating. Google says they have 17k pages indexed, but whiel normally they let you see upto 1000, it tops out at 422 and deems the rest similar. And that is including the 100+ pages the blog has created.
3 Responses to Site Analysis: Improving Yipit's SEO & Usability
Andre Marcelo-Tanner
December 10th, 2008 at 1:23 am
I know posts can be split into and intro post and a full page post?
Yipit Blog » Blog Archive » More Features, Implementing Feedback
December 12th, 2008 at 6:59 pm
[...] announced our business profile pages and elicited feedback on Yipit. Well, we got it in the form of an awesome UI & SEO review from techsoapbox. Depending on your math, there were just shy of 30 discrete [...]
Vinicius Vacanti
December 12th, 2008 at 7:10 pm
Ahmed,
This was an awesome post! Jim and I created a task list based on these suggestions and have already implemented a number of them. Thank you for taking the time to write about Yipit and your insights were truly appreciated.
Best.