The Perfect Shopping Mall: Inside Google

I may have been drinking a bit too much Kool Aid recently, but this post reviews why I think Google has a better business model than eBay... Although I want to, I have yet to read the book called The Perfect Store: Inside eBay. eBay has seen slowing growth, has limited feedback mechanisms, and the announcement of Google's payment processing service is a shot across the bow at eBay.

Buying Cycle:
Sure people buy a large amount of stuff on eBay, but many people only go there AFTER they are in the buying mood. Many people use the web to research, and eBay clearly misses out on that opportunity. Recently they bought, but that brand and service is still targeted at the later ends of the buying cycle.

Google's business model allows them to profit many times along the entire buying cycle, and their new payment processing will allow them to collect a percentage of the sale when people finally buy.

An Ad Everywhere:
Want to search for something? Here are some ads. Reading something interesting? Here are some ads. Checking your email? Here are some ads. Want to buy something? Buy it through Google Wallet.

Trackable, Precise & Easiest Ads to Buy:
There are a few oolies, but generally those ads from Google are the most precise and some of the easiest ads in the world to buy. They also allow people to buy using different mechanisms. You can create search only ads, also place those ads on content, target them to a region, and target them to a language.

Google also offers free tracking so they can collect more market research data.

If you are worried about click fraud on content sites you can buy site targeted ads on a CPM basis. Eventually they will probably add features which allow you to limit the number of times any one person sees your content ads.

Could I afford to run branding ads on Not until Google site targeting came about. By making the ads available in quantities small and large they get more participants in their ad auction, higher ad prices, and more ad revenue.

Limited Value of eBay Feedback:
The numbers and text eBay feedback comments are limited in value. In the hotel lounge at SES London one of my friends was looking at an item and said the guy selling it had a 97% possitive feedback. Immediately the guy sitting next to him started explaining how that is not a good number. Is it? I don't know, but a 3% return rate is not uncommon for many products sold over the web. eBay has a ton of community feedback, but most of the members still remain faceless.

Trust, Value, & Price:
We tend to be willing to pay more to people we trust. Trust is much more of an emotional issue than a mathematical one. Over the weekend MasterCard announced a breech of 40 million credit card accounts, which happened through what consumers likely feel is a nameless faceless middleman. Until I read articles about it I did not know it happened or who CardSystems Solutions was, and now all I know is that they suck. If people trust you at the other end of the purchase they are more inclined to trust the steps in between.

When items come from a faceless vendor, with mostly textual sales copy, and a one off relationship, it is hard to build a brand and tell a story. That is the flaw of eBay, currently it is limiting how well it can allow it's members to tell stories. There will always be some disconnect between true value and price. Why not host merchant blogs or value added story building tools like that on eBay?

Saving Money & Paying for Certainty:
There is nothing on eBay which deeply motivates me to build a relationship with a merchant instead of bidding a few dollars cheaper to buy the same item off someone else.

The whole frame of mind at an auction is "let me save a few dollars". People pay a premium for the story that goes with something. They also pay a premium for certainty. For some reason I think the Fletch DVD is out of print in the US. I will probably end up paying $20 more for it to know I am getting a certain price instead of going through the auction process at eBay.

Hollow Shops Not Building Trust:
I searched for eBay shops and was brought to They have featured merchants on the home page. Look at how ugly these shops are:

  • Golf Madness - nothing sells golf like bright ugly red?

  • Candle City LTD - I am not sure why, but it FireFox there home page has huge flashing text. what is that?
  • Look Cool for Less - the name stresses save money. not pay extra because you are worth it.

Those were the first three shops I looked at. The individual product pages look so ugly and modularly built that they remind me of Geocities (or Todd's favorite site). The fact that those sites were the featured ones tells me that eBay, merchants, or consumers must not take that idea too seriously.

Brand Extention:
Recently Forbes did a piece on brand extension, stating that consumers believe many brands can be stretched. Some have suggested that Google might want to buy a merchant with a recommending service like (see Epic 2014), but Google has access to far more information, and could probably create a better recommending program.

Pure Data:
Google bought Urchin, and also offers free tracking with their AdWords product. They could afford to give away Urchin, but they charge for it to keep the data more pure. Essentially they are providing a poor tax to filter out anomalies and smaller poorer sites.

Urchin not only allows them to track their own search service, but it also allows them to see shifts in market share, as well as how well competitors monitize their traffic and merchant ROI from competing services.

While it is likely Google will profit from their payment system, they may also want to create that service to have access to more raw data.

Ad Recommending Engine:
When you buy site targeted ads Google asks you to enter a few sites and a few keywords to recommend where else you might want to advertise.

Adgooroo lets people see some of the best terms their competitors bid on which they are not yet bidding on. Based on competing site ad buying habbits and conversion details Google could know where you should be buying ads. Of course, there will be limits to what advertisers define as acceptable sharing of information, but after the information becomes widely available elsewhere it might be considered acceptible for Google to share more data.

Google could even recommend bid prices and use competing advertiser details and web browsing patterns to tell you what parts of your sales cycle they believe are weak and how you may be able to improve them.

Thought Recommending Engine:
Search personalization, semantic web applications, and social filtering help guide people to certain channels and new ideas, helping people find what they love, and helping the ideas spread (causing the ideas to be refined and creating more content to place ads on).

Owning the Stock Market:
Having the most pure data means you have better data than anyone else in the world. Company insiders have some data tha Google does not know, but Google does get a deep snapshot of many businesses, buying trends, and better data about your competitors than you do.

Google could also look at link and news citation data or query volume for business related issues, and perhaps even predict when buyouts or mergers will occur. Eventually Google will probably buy or create something similar to Technorati just for the market research data. Perhaps they could do that with Google Sitemaps (I am not sure, but I do not think Google Sitemaps has a ping feature yet).

Google roughly knows how much value there is in nearly any market at any time. More importantly they can spot market shifts and buy or sell shares in near real time.

Investors are somewhat attached to their money and trade with emotions. Google has access to a large investor userbase which they feed news to and recommend thoughts to. It is not uncommon for a small cap stock to gain 10% or more from being mentioned on certain sites.

Google could create self training genetic algorithms to make bets for or against companies based on internal data. The program could teach itself how to make snap judgements based on self training criteria. This data could be used as one data point, or entirely automated - without any human interaction.

So long as Google was right 51% of the time they would make a ton of money, and surely it would be easy to create something that was right more frequently than that with all the data they have.

Imagine how Google could leverage their reach, their data, and their market capitalization in the stock market. Even if Google did not directly use the data imagine how much money investors would pay to access it.

Employment & Scalability:
Google has made scaled computing cheap, can recruit the top talent in the world, pays millions of dollars to outside sources to create content for it's engine, and scaled their internal system to be able to employ cheap remote workers. Their search data will soon be accessible almost anywhere, especially as the cost of communication continues to drop.

The Perfect Product:
I don't think Google will end up selling physical inventory directly because:

Filtering data is what search is all about, but Google could end up collapsing under its own weight if their hunger for data causes them to lose focus on what originally made them successful & the space is constantly evolving. If Google doesn't screw that up and do not lose their trust factor by collecting more data than people want to give they are poised to be the most powerful and richest company in the world within a decade IMHO.

Published: June 20, 2005 by Aaron Wall in google


June 20, 2005 - 12:04pm

Phew! A ton of information there, Aaron! I like the fact that you're prepared to bat ideas around, which on most forums is a no-no since industry-credibility is at a premium and 'not looking stupid' means that no one takes risks.

Personally, I think we're headed for Artificial Intelligence in the SEs within a decade, since they have the huge resources, the commercial competition and the know-how to get there. As soon as they start buying small AI research firms and advertising AI jobs we'll know the competition's hotting up.

The implications are fascinating but it's still a concept that most forum-plodders resist so there's no point in going into them. Semantic Indexing and Personalisation seems to be as far as most forums have got, and even that's stretching the bounds of most minds.

Back to your post. Ebay feedback sucks because there's now an unwritten law (sometimes it's not unwritten but blatantly stated) that the seller will give good feedback only when they receive good feedback from the buyer. There's also some sort of altuistic effect where feedback tends towards the good rather than towards the bad. Feedback of 100% is commonplace. You also have to check out what market you're in. A seller of dodgy software will have a street-wise clientele who know how the system works.

June 20, 2005 - 12:34pm

Hi Richard
thanks for the comment :)

Ebay feedback sucks because there's now an unwritten law (sometimes it's not unwritten but blatantly stated) that the seller will give good feedback only when they receive good feedback from the buyer.

I think that is a huge part of it right there, but another huge problem is that I jumped on the web when I was 19 (no thats not the problem hehehe) and now almost 7 years later eBay has the exact same feedback mechanism that it had back then.

They know their feedback system is gamed to make ALL sellers look good. Eventually though someone buys bunk products from a 5.0 merchant and their opinion of eBay is diminished and the feedback means less.

How smartly you are using the data is just as important as what type of data you are collecting. Why not try to challenge the boundaries of the feedback system? Even with the small data they have they can find the common review words and Statistically Improbable Phrases.

How do you make that feedback better?

  • Give people free PayPal cash for writing long and useful reviews. They are probably going to end up spending that money back at eBay anyway, plus in giving them that cash bonus you are reminding them to come back to eBay.
  • Try to give people a greater sense of ownership. Your feedback is what makes eBay great...
  • Maybe give people free credits for posting pictures of their products and their smile when they get them.
  • Maybe require merchant blogs (or even just automate most of the postings...just share the data in ways that are easy to surf through) that track each item and the history of the seller. Automatically connect rated noded and feedback from people who bought the same product from the same merchant or other merchants, as well as items sold for around the same price by the same merchant.
  • Many times getting some of my PayPal details are a bit of a nightmare. Nowhere near as easy as getting my AdWords history. As a company I think eBay is undervaluing the value of their feedback and what feedback people are willing to give them.
  • Instead of a 5 star rating system let people enter a 1-100 number on a couple different categories. Product as promised, timely shipping, no hidden fees, etc etc etc

My ideas might be complete garbage, but somehow eBay needs to challenge the limitations of the 5 star system. I was bored the other day and looked through some Amazon feedback and this perfectly sums up the limitations of the 5 star rating system:

It doesn't even merit a star
I can't for the life of me understand why people listen to this band. No talent whatsoever. Avoid this.

Amazon seems to get far richer feedback because most of the feedback is about the product and not the merchant, and it works with or against the marketing messages associated with those products.

Humans value the experiences of other humans, but you need to find a way to get them to want to share that data, and then you got to make it easily organizable in ways that make sense. Sorta like search :)

Add new comment

(If you're a human, don't change the following field)
Your first name.
(If you're a human, don't change the following field)
Your first name.
(If you're a human, don't change the following field)
Your first name.