It’s the search box, stupid – a plea to learn from the lessons of the past

No primary category set

David Brackin is an occasional guest writer on Tamebay and is a Director of Stuff U Sell, the UK’s leading eBay Trading Assistant. He used to be head of Strategy for LookSmart, Inc, an internet directory service, which provided the web directory behind the MSN portal.

At the end of the 90s, the first Internet bubble was inflating nicely thank-you very much and I found myself in San Francisco riding the roller-coaster. I was working for a small start-up trying to categorise the web. This was long before anyone outside Stanford had heard of Google: it was the days of Netscape, Alta Vista and Excite. We were producing the ultimate category-based guide to the web and this seemingly impossible task was not entrusted to vast banks of computers, but to a dedicated editorial team of humans. Each was an expert in their area and built up a category, they worked hard, barely seeing the light of day, summarising and cataloguing the best websites. After all, who better than a vet to introduce you to the best ten sites on the Web for feline health?

Websites were categorised and the library grew accordingly. At the same time, the search engines were emerging – initially searching the category databases, and then fanning out across indexed web content, they filled in gaps in the library. The newly invented portals adopted a policy of combining both together – delivering category results backed up by web searches. As the web grew it became harder and harder to categorise sites – the directories asked webmasters to self-categorise when submitting sites. The editorial oversight was lost and the advantage over the machines broke down. Google won and webmasters were left to build websites, not categorise them.

Fast forward 15 years and we have two dominant e-commerce platforms in the UK. Amazon which has grown from a catalogue, and eBay which is based on free-form listings. One of the great joys of shopping on eBay has always been the chance to find just the idiosyncratic thing that you are looking for – and as a buyer knowing that the listing was created just for you, describing to the best of the seller’s ability the very item you will receive. Perhaps there’s a little scratch on the iPhone case or a stain at the bottom of the Hermes bag – you get to see it and decide before buying.

So I’m disappointed to see such a focus on categorisation emerging again – the eBay catalogue and the proliferation of item specifics reminds me so much of our earlier doomed attempts to have webmasters categorise their websites for the benefit of web directories. It’s early days, so you wouldn’t necessarily expect the catalogue and associated tools to be very good yet, which makes their compulsory nature frustrating – heaping operational problems onto sellers who would rather be selling. At Stuff U Sell, we spent an entire weekend manually relisting nearly 3,000 of our clothing listings to satisfy the new compulsory item specifics, which mostly added nothing to the buyer experience.

The real trouble, however, is that it won’t get much better. The size of the catalogue is just too big: already any brand past the letter “L” had to be typed in manually when we did our update. You don’t read about Amazon becoming more like eBay, because it can’t: the very strength of eBay is that its seller and item distribution has a hugely long tail which it aggregates to become so big and strategically unassailable. The stronger it becomes, the worse any cataloguing attempt will be.I feel like I’ve seen this all before.

So what is the answer? Learning the lessons from the past, it’s all about fixing search. Huge teams are working at eBay on this issue, but noone has cracked it yet – both Amazon and eBay still generate pretty frustrating results – but the prize is worth getting. Google’s aim was to allow readers to find publishers. eBay’s is to allow buyers to find sellers. Their approach needs to be grounded in the same strategy, and not allow internal and external resources to be diverted from the task of selling by distractions.

21 Responses

  1. too right item specfics are the biggest and most frustrating pain in the arse invented by some nerdy tech geek who knows nowt other than how to blag ebay into self destructing

  2. A walk down memory lane indeed, and at the end of the road, Google.

    I will no longer buy anything on eBay as a matter of principle but I do add to their page view totals from time to time. My eBay expeditions are becoming rarer simply because if I am searching for collectibles it is impossible to find anything.

    Unless you are looking for something utilitarian like labels my advice to buyers would be Google it, not only Google Shopping but a web search too.

    You answered the vital question yourself.

    Neither eBay Management nor the huge teams working on the ‘finding experience’ want to hear common sense. They will continue throwing good money after bad because nobody will accept the career ending responsibility for the fact they have been barking up the wrong tree for the last 7 years. They would like to continue receiving their pay checks.

    From a business point of view, if eBay works for you then you will just have to work with them. Only you can tell if the effort involved is worth the benefit you receive. eBay is not going to change.

  3. I’m a former eBay employee, and a core member of several teams dedicated to search, classification and taxonomy.

    This article, and comments even more so, makes me pretty frustrated, because they seem deeply uninformed.

    Item Specifics are a way of tagging meta data to an listing. That meta data gives Buyers numerous additional vectors from which to ‘slice’ a set of search results – far beyond what can be done with just title keywords.

    Item Specifics are a resounding success, they are highly adopted by Sellers, heavily used by Buyers, and drive up conversion rates, average sale price and the number of bids on an item.

    I’m sorry Chris, but your analogy with the early 90’s websites is incorrect. eBay is in no way looking to be categorising the entire inventory of the site – we are just looking to create tools that allow Sellers to more accurately list their items, add more details and more information on the item. Buyers can then filter results far more accurately.

    I fail to see how this can in any way be said to be “mostly added nothing to the buyer experience”.

    I can only surmise that you’ve misunderstood what Item Specifics are for?

  4. I sell technology products. Specifics are just plain innacurate. Some features are included that don’t exist in the product. Sometimes specifics are changed without me knowing – could be for better or worse. I just don’t use them at all, especially if it only takes one or two bad stars to destoy my standing. I sell all stock anyway so tinkering with dumb specifics is not worth my time and effort.

  5. With the release of the new Google search algorithm it seems that Usability and User Experience are a key factor for Google ranking. An website that has more then 10 pages needs search box. Thanks for the article

  6. Why are we currently focusing on categories and item specifics? First and foremost, it’s because people searching and buying on eBay use this information to help refine their searches and make a decision about what to buy.

    Of course, some users prefer just to use the search box, but from the data we can see that many more users make use of the category and item specific filters. This is the model they are used to on other sites and how they expect eBay to function, too. eBay can only make this experience work well if sellers provide us with this information; there’s no way we can guess this accurately. This is why we went as far as requiring certain item specifics within the clothing category – we see that buyers use the filters in order to navigate the site and yet many listings didn’t contain the necessary information.

    Google realise that structured data about inventory is a requirement in order to do a good job of ecommerce search: take a look at the recent changes made to Google product search, you’ll see that this is the direction in which things are moving. Even within the clothing category, they are requiring that brand is submitted as an attribute. I doubt that they will stop there.

    Now, we clearly want to make it as easy as possible for sellers to provide structured data when listing and I take the points regarding consistency and not being able to ‘map’ equivalent values together (take Chris’ example with suits sizing). This specific issue is one I’d like to fix, but I’d also love to hear if there are other reasons/ problems that make it difficult to provide structured information.

  7. Whilst I accept that sellers sometimes want to use non-standard item specifics eBay really should have the recommended specifics (I.E. those that actually appear in the Product Finder) at the top of the Suggestions in listing tools.

    An example is in TurboLister for “11483 –> Clothes, Shoes & Accessories > Men’s Clothing > Jeans” for the Trouser Size (waist!) item specific. eBay’s suggestions start with XS, S, M, L, XL, XXL and that’s all you can see in the drop down without scrolling. Check eBay search or scroll down the suggestions and you’ll find the best values to use are 30″, 32″, 34″ etc.

    You’re doing two things.
    1) Encouraging sellers to list values which won’t be found using the Product Finder in the Search Left Nav Bar
    2) Compounding the problem as eBay learns which values sellers use and thus continuing to suggest them.

    Don’t offer options which are not the best values (if not just outright wrong) and certainly don’t suggest them at the top of suggestions because people will use them. The more the less desirable values used the worse eBay search gets.

    (Well that’s unless the entire world from Levi’s downwards scraps the W34/L34 way of selling jeans in favour of M/L34 but I don’t see eBay having that much influence ;-))


Why it's bad for retailers if Google are forced to scrap PLA in favour of CSS

Why it’s bad for retailers if Google are forced to scrap PLA in favour of CSS

Google PLA

CSS lobby EU for disastrous replacement of Google PLA

AI is ready to boost online sales - Not On The High Street shows us how

AI is ready to boost online sales – Not On The High Street shows us how

Shopify launches AI powered Semantic Search in Winter ‘24 Edition

Shopify launches AI powered Semantic Search in Winter ‘24 Edition

It Starts on TikTok campaign celebrates search and discovery on the platform

It Starts on TikTok campaign celebrates search and discovery on the platform

ChannelX Guide...

Featured in this article from the ChannelX Guide – companies that can help you grow and manage your business.

Register for Newsletter

Receive 5 newsletters per week

Gain access to all research

Be notified of upcoming events and webinars