It all started with a realization Jeff Bezos had in 1994:
Retail giants like Walmart had up until then only focused on optimizing shelf space as they were working with physical real estate.
With the advent of the internet, Bezos could see a new possibility opening up: online shopping meant every retailer who sold online would have infinite shelf space. An online retailer also wouldn’t be limited by the local market but could expand to the national and global scale pretty easily. Shelf space was not a real constraint anymore.
Whereas a traditional retailer had to deal with tradeoffs as they had finite shelf space, an online retailer could display page after page of products with almost zero marginal cost for more products. The online retailer didn't even have to care about what to stock. Amazon could let its customers add all sorts of items to its catalogue, measure web traffic for each item, track what items customers had added in their carts, and only stock items in its warehouses that seemed most likely to sell.
Bezos, in short, wanted to make a limitless Walmart on the internet.
He removed geographical and space constraints, let vendors add as many SKUs as they wanted and implemented a search functionality on the platform, along with a system to capture customer reviews. In this new online model, the quality or selection of products on display wasn't Amazon's concern anymore, it was just pure quantity.
But while infinite shelf space worked out very well for Amazon over the years, quantity comes with a problem:
How do customers really discover new products?
You might say that the search functionality is enough. But is it, really?
The search is a simple categorization of products along with some filters like avg. reviews, featured, price, and newest arrivals.
How does Amazon curate products like an offline retailer like Walmart does?
How does it decide which products are in the front of the aisle and which products towards the back?
How does it decide which products to place on the top shelf vs. the bottom shelf?
To solve this problem, Amazon has tried to create a ranking algorithm that weighs product reviews and sales velocity. The more reviews a product has and the more units it sells, the higher it climbs in rankings. This creates a positive feedback loop: the more a product is exposed to customers, the more it sells; the more it sells, the more reviews it gets, and the higher it climbs in rankings, and the more it gets exposed to customers.
This is a problem for Amazon as it promotes tested products and vendors over newer entrants. It leads to dull and boring search results and puts off a sizeable chunk of customers who are often early adopters and like trying new products.
Amazon has tried to solve this problem with Amazon advertising, something I've covered in my #100DaysofMBA series.
Amazon advertising allows sellers to boost the visibility of their products by paying for ads that show all across the search results page — marked in really tiny font as "Sponsored."
In a typical search page, the first row of results is almost entirely filled with sponsored ads. But this is not a good experience for customers in the long run, because the ability to buy the top slot in search results favours sellers with deep pockets and products with high gross margins — not the products that are the best deal for customers. Also, if the seller is paying for ads, they're naturally marking up the prices of their products to make up for the advertising spend.
Moreover, it's very hard to notice if a listing is sponsored. They don't just occupy the top portion of the search page like in Google Search, but are sprinkled intermittently across the page, making it very hard to notice.
Now you might say, "But Aditya, don't poor products get rated poorly and hence get weeded out quickly?"
Yes, that happens. But hold on. There's a problem even there.
Product reviews on Amazon live in perpetuity. If a seller was initially selling a high-quality product that amassed thousands of 5-star reviews, those reviews stay even after the seller has replaced the product with an inferior one. So the bad reviews that the product is now getting don't show up in the average review ratings, fooling the customers into thinking that they are buying a highly trusted product.
The solution:
Remove product reviews after periodic intervals.
This solves two problems.
Firstly, it ensures that the product reviews apply to the most recent version of the product that is being shipped.
Secondly, it evens the playing field in some sense for new entrants. They would now only need to compete with the quality of products being sold in the past few months, and not with vendors who have amassed positive ratings over the years and are now simply selling shoddy products based on those ratings.
The advertising business might be lucrative for Amazon, but in the long run, I can't help but see it as degrading the customers' discovery and purchase experience. Remember that search is the only way to discovery on Amazon; they drive the entire e-commerce business. And if their ranking algorithm isn't nuanced enough and seeks easy and profitable cop-outs like advertising, they might be putting a lot of customers away in the coming years.
Organic, customer-centric product rankings that earned Amazon much of its success and revenue are now at the mercy of ads. And people who don't have foresight think this is a good thing for business and profits.
