Since the beginning of Meesho's journey, we have looked for new ways to serve buyers and sellers. We strive to offer our customers their favourite products for the lowest price possible while ensuring that our sellers — from small to large, branded to unbranded — are able to grow their businesses.
Many of our vendors have experience selling goods offline, but are first-time online sellers. Meesho helps these sellers expand their potential customer base from their local community to almost the entire country.
When selling offline, sellers understand the local market price for similar products sold by other shops. However, the online landscape is different — multiple sellers offer the same or similar products at different prices. As a result, they often struggle to price their products for an online platform.
For example, if the average selling price of a particular style of t-shirt is, say, INR 200, pricing it at INR 240 will make you lose sales to competitors. Conversely, pricing it at INR 140 might increase sales but will lower the margins.
We need a balance.
At Meesho, one of our mantras is “User First”. Therefore, we set out to solve the pricing problem in a way that benefits everyone. However, our platform hosts over 50 million products, so we cannot depend on a manual process. Therefore, we used our data science prowess to create an automated solution.
We like to call it Price Recommendation.
How we generate recommendations
We use our data science models to pick which products to offer recommendations for. We base them on one of the following two criteria:
- Using our data science models to identify identical products. For this, we again use one of two methods:
- Image-based matching analyses a product’s picture and compares it to other products’ pictures in Meesho’s catalogue to find the same or similar items.
- Product features and text-based matching analyses product attributes. For example, in a t-shirt, it looks for features like the colour, collar type (round vs v-neck), and sleeve length, among others. If two or more products have the same attributes, they’re marked as identical.
- Using the price elasticity of demand, i.e., how the demand for a product fluctuates (in terms of views and orders) with changes in price.
Finally, we compare the products’ prices among various sellers and display them in the Price Recommendation Tool.
How the Price Recommendation Tool works
This tool lets suppliers compare their prices with other vendors’ prices for the same or similar products. In terms of sales growth, it displays the impact that accepting price recommendations has had at both the product and the overall level.
They can filter products based on item category and/or sales numbers (“Top Selling”, “Low Selling” or “Non Selling”). To help them make an informed decision, we also display the transfer price, i.e., the amount that they will receive after accounting for logistics:
Sellers can then accept or reject the recommendation, or offer their own price:
Closing words
Our Price Recommendation Tool has seen immense adoption as it helps sellers improve sales and increase their profits. In July 2022 alone, nearly 125k sellers updated their product prices based on these price recommendations and saw an orders growth of 54% on average. Moreover, as we promise in "Sabse sasta to Meesho pe hi milega!", buyers get the lowest prices while not compromising on quality.
Of course, we won’t stop here. With our goal to empower 100 million small businesses in Bharat to succeed online, we’ll be adding some much-requested features in future versions such as better insight into product return prices. As always, we’ll document our journey here, so stay tuned!
Like what we’re doing and want to join us? You’re in luck as Meesho is hiring! Join us and help us in our quest to help millions of entrepreneurs in Bharat thrive and bring e-commerce to the next billion users.
Credits
Author: Nitish Duggal
Guided by: Nikita Kodnani and Chetan Kalyan
Edited by: Shivam Raj and Rochak Singh
Cover image: Rahul Prakash (LinkedIn, Behance)
