in Success Stories

Customer: Southeast Asia’s leading eCommerce company
Industry: eCommerce

The Challenge Addressed

More than 90% of shoppers invest time online to hunt for the best prices and 20% of eCommerce website traffic comes from a plethora of Price Comparison Engines. To ensure that an eCommerce website gives the best prices, it has to track the prices of different products from its competitors and reflect the best on their website. Price tracking – it’s important for eCommerce companies to map their products to that of their competitors on their websites.

A lot of companies employ an in-house team for this activity. However, burdened by high cost and low turnaround time, scaling up operations becomes slow. Quite counterproductive considering that TAT is very important to stay competitive. Another challenge they faced was scraping data from different website interfaces. In a nutshell, in-house operations are more developer effort and less results.

One of our customers, among the eCommerce giants in South East Asia, wanted to optimize their product pricing across all countries that it serves.

The technology tying all of this together to generate the best possible prices – Competitive Price Intelligence. Simply put, the process involves feeding two products with all available data points and it is to be determined whether both the products are the same or not. Once identified that the products sold on the client website and the competitor website are a perfect match, the price of the product on the competitor website is extracted.  In the wide spectrum of eCommerce operations, the checks performed to ensure a perfect match between a product pair on client and competitor(s) are:

  • Image
  • Title
  • Shipping Type
  • Product availability Pincode
  • Brand/Seller Warranty
  • Product Model
  • Condition of the product(whether New or Used or Refurbished)

Scale: Products matched per day is 5000 with 5 competitors for each of Singapore, Malaysia, Philippines
Turnaround Time: Reached 4 hours from 24 hours of current in-house process.
Expected Quality: 95% +

The Playment Advantage

At Playment, we built an automated workflow consisting of the major verifications required to classify the product as a match. The client would integrate with Playment’s API to facilitate automated data transfer in real time. Once integrated, the tasks were routed exclusively to our expert players, selected using Playment’s Player Selection Logic (PSL).

The client fed Playment with the seed product URL and its competitor product URL. After which the Playment workflow identified whether the products in the URL pair matched or not. If the products did match, the product price from the competitor website was extracted and shared. If a mismatch was identified then, the exact reason for the mismatch was shared.

In terms of quality, the single pair of product URLs is shown to the multiple number of Playment cloud workers – already trained users – which ensures that only the answers which reach 80% consensus are correct. Another notable advantage for the client was TAT – Barely 4 hours for 5000 product pairs per day. This can be seen as a result of the flexible nature of our worker base.

The cloud workers conducted quality checks against variants and other factors with the following  objectives:

    • Obtain exact match and mismatch.
    • If matched, then provide variants and price for the right variant.
    • If mismatched, provide reason code.
    • Provide specific reason for mismatch

The final output was provided back to the client in the desired format through an API.


Benefits over in-house operations

One of the key elements of Playment’s success as an effective data operations solution is our scalability. As we strive to constantly improve our accuracy and reduce TATs, we plan on improving our scale from 5000 products/day to 10000 products/day.

Schedule a Call with Playment

See how machine learning with Humans in the loop can help to get a lot of work done efficiently.

Schedule 1:1 Demo