in AI Alerts

This tweet cracked us up, but it does capture a very real problem E-commerce customers face. Indian consumers returned products 12% of the time last year leading to a loss of $300 million for E-commerce biggies like Amazon, Flipkart and Snapdeal. In an intensely competitive landscape, there is pressure on every player to increase customer confidence and convert new users to buy online.

In this post, we will focus on three things companies can do to make that happen.

1. Improve the Accuracy of Cataloging

Indian E-commerce companies add 300,000+ new products to online catalogs every single day. This means 1.5 million+ content tags are generated for products daily. However, Indian sellers struggle to source or create quality catalog data, and perform poorly – providing correct information, only 70% of the time.

This is worrying because customers rely heavily on accurate and comprehensive information to make informed buying decisions. And receiving a product that does not match the online description dents customer confidence greatly. Which is why it is really important to have a scalable, reliable, and fast quality check process to quickly flag cataloging errors and rectify them.

2. Moderate customer reviews

E-commerce is a ‘review-first retail-second business.’ Statistics show that 65 % of consumers spend 16+ minutes, and 36% of consumers spend a whopping 30+ minutes comparison shopping before proceeding to buy. Shoppers rely on reviews to gauge if the product they are thinking of buying is fake, whether it works as advertised, and if the actual product matches descriptions provided in the catalog. However, reviews can turn away potential customers when they are irrelevant. For instance, the review below doesn’t serve a purpose because it is promoting a loan provider and has nothing to do with the product advertised.

Such examples abound. Some review threads are hijacked by disgruntled customers, others use profanity, and still, others post spam (such as coupon codes or links to pirated movies). Which means that E-commerce players need to invest considerable resources, to filter, flag and escalate inappropriate reviews, in a time-bound manner.

3. Improve Search Relevance

Lastly, every buyer browses listings with a specific need in mind. They expect spot-on results, whether they are looking for a ‘dress shirt’ or an ‘automated cold glue gun’. Showing relevant results directly impact discoverability and conversions. Conversely, irrelevant results (like in the example below) leads to drop off and a direct loss in sales. Firms need to improve search results by ensuring the availability of deep, enriched catalogs. They also need to optimize search algorithms by constantly, training and testing them.

Which brings us to the question – how can E-commerce firms QC catalogs, tag and flag reviews, test and train algorithms, reliably and efficiently? Currently, firms build large in-house teams (incurring large overheads) or they outsource these functions to BPOs (who struggle to keep up with the demands of these rapidly growing firms.)

But what if there was a smarter, scalable, timely and cost-efficient solution?

Say hello to Crowdsourcing. Currently, there are a handful of firms seeking to solve data operation issues via crowdsourcing such as Crowdflower, Amazon Mechanical Turk and Playment (that is us!). Since inception, Playment has helped top Indian e-commerce companies like Flipkart, Myntra, Paytm, Roposo and Limeroad build customer trust and increase customer conversions.

> Sriram K, Product Lead, User Generated Content, Flipkart vouches for our contribution saying, “Playment has been instrumental in scaling up the content platforms at Flipkart. This crowdsourced assistance in moderating content for Flipkart solves a real world problem at scale with quality guardrails in place. I would recommend their platform for any human intelligence task at scale.”

We leverage a network of over 50,000 ‘players’, sitting in different locations, to automate data operations. Playment’s crowdsourcing model trumps in-house operations because it is cheaper and easier to get off the ground. Its pay-per-use model makes it scalable and flexible in a way traditional outsourcing is not. And thousands of people working on a single problem, means you are assured of results – fast.

Schedule 1:1 Demo