Understanding the Challenges of eBay Search
Sovoia founder Obaidullah Khawaja is an expert in ecommerce search, having worked at Amazon.com and Microsoft and now working closely with eBay. Last year, EcommerceBytes wrote about his company's service for eBay sellers called the Sovoia Search Widget, a faceted search utility used by retailers on eBay UK's Fashion Outlet. EcommerceBytes Editor Ina Steiner sat down with Obaid recently to learn more about the challenges sellers face in getting exposure in search results and some important details about eBay's search engine. You can visit his company Sovoia on this page.
First, tell us a little bit about your background and the companies you've worked for and how you determined there was an opportunity in the eBay retail space.
Obaid Khawaja: I have a computer science background and have previously worked at Microsoft and Amazon. My last job was at Microsoft, I was working within their Display Ads Network called DrivePM (now Microsoft Media Network) which at that time had a 30% share of the global display ads market. In this role I learnt the importance of optimizing advertising spend and tracking real time performance. I had already worked at Amazon at this point and realized that optimizing sales for online sellers and reporting on interesting data could be a good problem to solve.
At Amazon I was in the Retail Customer Experience team where my role involved looking after "Guided Shopping" tools like product comparison, faceted search widgets and also ensuring data quality (to power these experiences). In this role I learnt the value of having tools that help buyers find exactly what they're looking for in the simplest way possible.
For example a TV Finder we built and placed on the TVs category page performed really well because it helped buyers find TVs by Brand, Display Size and Price in a few simple clicks. On the flip side this meant having really good data, which was quite a challenge. Not only do sellers have to provide the data in the first place, it must conform to certain platform specific standards.
How did that lead you to found your new company, Sovoia?
Obaid Khawaja: I used my unique experience to hone in on two challenges in the online retail space. First, we started building great shopping experiences through our Sovoia Search Widgets. We have now expanded our line up to include more advanced forms of search widgets that are placed on store search pages (for refining results). Deals Widgets are another new product which allow stores to create custom deals on the fly based on existing inventory
Secondly, we started to tackle the issue of poor listing data quality. Without good data you can't build great shopping experiences. For example, if I want to buy a laptop and am trying to compare RAM Size or processor speed, then the user experience will be broken if even one of the laptops has missing attribute data.
In a marketplace, the problem is compounded because you have thousands of sellers with their own data sources, thousands of categories (each with their own meta data), and millions of items. The Sovoia Data Analyzer is a brand new product in this space which tackles this issue.
The Sovoia Search Widget aids product discovery for online retail customers, explain what that means.
Obaid Khawaja: First we have to understand what product discovery is. We like to think of it as matching buyer intent (buy side) with a seller's inventory (sell side). The marketplace facilitates this matching and offers essential tools like search, images, attributes, descriptions, videos, reviews, etc., to a buyer. The seller must provide the required information like an SEO friendly title, high quality images, standardized attributes, etc., to optimize sales. Then the platform must use this data and make it accessible through organic search, product discovery tools or product detail pages to the buyer
Online marketplaces like Amazon and eBay have millions of live listings, and buyers need to perform a search and drill down into specific categories before they get to relevant attributes. For example, if I'm looking for a 32" LCD TV, I'll type "LCD 32 TV" on eBay which will return wall brackets, accessories and some TVs. Next I'll explicitly have to say pick the TV category which loads just TVs. If I'm looking for a specific brand within a specific price range then that's 2 more clicks (on the left). This comes to at least 3 page loads and 3 clicks.
Our widgets remove all this waiting and offer instant gratification. Think of them as very useful display ads. We package relevant product attributes and surface them to a potential buyer on any highly trafficked page (eBay home page, store page, category page, etc).
So, for example, in the case of Shoes we show Shoe Type, Color, Size, Brand, Price, etc., as options for the buyer to pick. If a seller has thousands of items then such a widget can really help the buyer easily pick the attributes, which matter the most in one go.
Next, our widgets offer an "elastic" experience where picking one attribute will automatically remove invalid choices for the other attributes. For example if you picked Size 9 first and there are no red shoes available then "Red" will not be shown in the Color selection.
Finally the number of products that match your selections will also be displayed before the search results page displays. If you're open to Brown Shoes (200 results) as well as Red (2 results) then knowing this critical piece of information upfront can help with a smooth shopping experience as well as a higher chance of conversion
Can anyone use the Search Widget, or is it restricted to large merchants?
Obaid Khawaja: Any eBay seller can use the Search Widget but it usually works best for stores with at least 500 items. We feel that its overkill to have the widget if selection is low. Also data quality needs to be good otherwise the experience becomes very confusing for buyers. They might see hundreds of items in the store when the visit it but very few show up in search results (due to bad data)
We have helped some of eBay Europe's top sellers like SuperDry and Office Shoes to improve their data quality drastically as part of the widget go-live process. This is where the Sovoia Data Analyzer comes in handy.
What is Sovoia Data Analyzer?
Obaid Khawaja: The Sovoia Data Analyzer is a QA tool which helps any eBay seller optimize sales by analyzing and improving the listing quality. It does 2 main things in its current form. First it lets any eBay seller scan through all their live listings and identify issues. For example sellers can see how many items in a category like Clothing are missing key attributes like Brand, Size, Color. We can provide more advanced diagnosis where attributes have been provided but are not optimized for SEO. So the second part is providing a fix for the problems identified. If values are missing, then we let sellers know which ones are missing and what the correct values are. If they're already there, then we provide SEO friendly suggestions as per eBay's BestMatch specs.
How are your clients using Sovoia Data Analyzer?
Obaid Khawaja: We find that many of our clients or the people we talk to aren't even aware of the problem and its implications. For new clients, we hold special sessions and educate them on the importance of good data. For example, one large, enterprise client was doing a great job of entering shoe sizes but they just weren't in the correct format. This had a huge impact on SEO - first, their shoes were not showing up when a shopper typed in the size in a related query. Second, their items weren't covered through the faceted search filter for "Size" on the left in a more generic query.
Another client found out that their own brand was not recognized by eBay as a recommended value. We worked with eBay to resolve the problem and improve SEO scores for this client.
So in summary, clients use the tool to improve sales by identifying and prioritizing problems. Next they download the problematic data set and fix it as per our recommendations.