• Gediminas Rickevičius, VP of Global Partnerships at Oxylabs


• 26.07.2022 10:00 am
#eCommerce

Pricing intelligence has existed for the longest time in various forms. Web scraping, however, has completely changed how it's done. Previously, lots of manual labour had to be involved to gather enough data for any sort of pricing intelligence. Now, the same volume can be collected in minutes or hours.

A lot has been written on how pricing intelligence and correlated strategies (e.g., dynamic pricing) influence global markets, especially for e-commerce businesses. Some have proposed that dynamic and personalized pricing might be a net negative for consumers. On the other hand, proper implementation of such strategies will increase competition between businesses.

As the free market thrives on competition, pricing intelligence is likely to become a net benefit to all consumers, if it hasn't already. Enabled by the near-instantaneous data delivery of web scraping, businesses will have to compete more harshly over margins.

Product matching builds the groundwork

Businesses utilize web scraping and other automated data acquisition methods to extract various information from competitor websites. Most e-commerce and retail companies use web scraping to collect product and pricing information.

Pricing intelligence rests on proper data matching between sets of product details and associated value. Some of it can be quite challenging as advanced applications of automated pricing strategies require tools to identify product identities.

It may seem deceivingly simple at first. Most e-commerce and retail websites will do their best to draft product titles that state all the qualities of a product. Unfortunately, all specifications will never fit in titles and a lot of important details will rest in descriptions and other fields.

These issues will start to compound themselves when we take e-commerce marketplaces and large platforms into account. User-generated content will never be standardized to the level of a high-functioning business. As a result, there will be discrepancies all over the place.

Building a product matching engine, therefore, becomes a necessity. Small errors in the process could lead to incorrect price matching, which could throw off entire strategies and lead to incorrect evaluations.

In the end, all pricing intelligence and matching strategies will rest on proper product matching. It may seem superfluous at first, but such a foundation creates a ripple effect across e-commerce and retail businesses due to the way it changes supply factors.

Demand and (over)supply

In e-commerce and retail, pricing strategies will rest upon the product matching system. We can assume that most of these will be highly successful with little room for error. In other words, competitors will be able to match specific inventory across websites.

While there are various implementations of pricing intelligence strategies, in a majority of cases, they will try to match or undercut competitors. There are more complicated implementations which might include stock availability or even potential algorithm collusion, but, especially in e-commerce and retail, prices will tend to be drawn downwards as customers place high importance on costs.

As a result of increased competition and other pressures, it is likely that the pricing of all products that are being matched converges towards some arbitrary value. There is some noise and variance in the convergence as some companies might have higher acquisition costs for the same product (due to shipping or other reasons), but it otherwise closes into that value.

Essentially, such implementations create a gravitational pull around product pricing possibilities. At the same time, no changes in demand happen as consumers aren’t even aware of the inner workings of the pricing algorithms.

Due to pricing intelligence, an interesting side effect becomes apparent as the product supply theoretically converges. In the consumer’s eyes, it’s almost as if companies start “sharing” the supply they have between themselves. Customers can get the same product for an extremely similar price from many different vendors.

While it does rest on the assumption that a significant portion of e-commerce and retail businesses are using some form of pricing intelligence, such an expectation isn’t completely out there. Deloitte surveyed e-commerce and retail businesses on their usage of AI back in 2017. Pricing automation and intelligence came up as the most frequent use.

As a result, the adoption of various methods for acquiring and working with pricing intelligence data has likely surged further since 2017. We’re much more likely to see the aforementioned price pull towards some means than ever before.

New avenues of competition

One emerging issue, for businesses at least, is that price competition has finite scalability. Some, such as SaaS companies, might have significantly more leeway, but e-commerce and retail have long been the sector with some of the lowest margins. SAP predicts that they have shrunk further, down to 4.5%, making continued price competition increasingly difficult.

As the sea of margins continues to dwindle, other avenues of competition have to be discovered. Some e-commerce giants have turned towards customer centricity, others look for improving user experience as they have already optimized their dynamic pricing strategies, changing valuations every 10 minutes. Amazon has even attempted to further their delivery practices through anticipatory shipping, a patented solution that uses AI to distribute products across warehouses depending on visitor habits.

In any case, pricing intelligence, in the end, benefits the consumer while changing the competition between businesses. It may soon become a necessity to implement product matching and pricing systems so that online commerce and retail would be possible.

There might be some bumps along the way, such as short-term impractical implementations or even accidentally harmful AI collusion, but in the end, pricing intelligence will meaningfully benefit consumers in the long run.