While plenty of people just search, plenty of people browse or search then browse/refine.
If you’re not including attributes, you’re not showing up for those people who browse or browse/refine.
Here’s your proof. There are 15,610 watches available on PriceGrabber.
However, if you add up the total number of products under ‘Fastener Type’, you only get 4954 products:
In other words, 60% of the watches on PriceGrabber will not be found through browsing the ‘Fastener Type’ refinement attribute. Now I’m not quite sure how PriceGrabber is getting it’s refinement attributes (partly from the manufacturers, partly from data feeds, I’d assume), but if it’s at all from data feeds, then it’s worth adding this type of attribute information to your feed (in the title, description, etc.).
And this rule doesn’t just apply to PriceGrabber. All of the shopping engines pride themselves on being able to pull out relevant attribute information. If there’s no attribute information in your data feed, your products will not be found by anyone refining their browse/search results.
The moral of the story: Include as much data as possible in your feed. And if there isn’t a field defined in a data feed spec which you think should be there, make them create one (heard of Google Base’s custom attributes???). You’re the expert on your products, not the shopping search engines.
You want an example?
My friends at Evogear are experts at playing in the snow (skiing, boarding, staying warm, etc.). Their refinement options for skis include brand, ability level, size range, waist width, and price. I’m a skier (and beginner boarder). The first thing I look at when purchasing skis is size range. Evogear provides a snazzy little pop with the following information:
Do any of the shopping engines provide ski refinement by size. Don’t think so. Because of this Evogear’s conversion rate probably takes a hit as consumers will search for skis, click through on a product, and potentially find the perfect pair of skis in the not so perfect size. Evogear pays for the click, but gets no conversion. ROI goes down. Evogear pulls it’s listings from the shopping engines, and everyone suffers.
Yes, I’m generalizing and over-simplifying a bit, but not much.
Disclaimer: All optimization strategies are suggestions and do not guarantee success (although I wouldn’t be writing these tips if I didn’t think they mattered). These are data feed optimization tactics I have used or others have suggested which I think everyone should at least think about, if not test (just test, please). Use at your own risk (you can always go back to the old, boring, pedestrian way of doing things). Or don’t use the tips and write a comment telling me I’m insane.
Brian, no doubt this is a great post and I really like your idea but; extracting and formatting the Product Attributes data is one of the most challenging task these days. I have implemented guided navigation for some of the prominent sites and honestly speaking; getting the Product Attributes data was the biggest challenge for me. Most of the retailing sites have good amount of Product Attributes data in their PRODUCT DESCRIPTION, TITLE and CATEGORY PATH; and some companies use CSS to present them effectively on their sites but; the challenge is to filter out the marketable attributes with their values from these descriptions and put them under different buckets. This is atleast 3-4 months exercise based on your catalog and not every company in this retailing sector have budget and time to carry out this exercise and even if they do it; they don’t follow it consistently. I would love to see CSE Engines modifying their DATA PARSING algorithms and come up with some great data filtering and formatting techniques and present the data in relevant attributes. I have seen some of the CSE sites who have started showing filterable attributes but; only those stores can take advantage of this who have guided navigation on their site or, those who have a smaller catalog but; the mid-size big catalog companies got affected by this more.
Comment by Sushant Ajmani — December 3, 2006 @ 7:01 pm
Well Said Sushant. FYI - we’ll be there in 7 weeks. We have a plan my man.
Comment by Michael Golden — December 7, 2006 @ 9:36 pm