Remodelling Flipkart’s Structure, to support Multi-Marketplace Model
How I helped remodel Flipkart architecture to support multi-marketplace solutions
--
Problem Statement
Flipkart’s future looks at increasing the services we are currently able to cater to the end users. In order to support those new initiatives, the structures need to be refined & restructured in order to support that growth. The current systems were not designed to accomplish that, so it becomes necessary for us to tackle the problem right now.
What makes this solution special?
With a well established multi-marketplace model, Flipkart would be able to integrate 3rd-Party Stores & future initiatives of Flipkart (like Pharma, Insurance, Digital Store, etc.) very easily into the system. These new stores & marketplaces can utilise the new constructs to connect better with the users.
Hypothesis
- By restructuring the Search module, we can better navigate the users towards the type of service they are looking for.
- With a better awareness about the service, people would be able to make a decision on which service or initiative under the Flipkart umbrella would be able to serve them better.
- Awareness setting would provide a better communication channel between the business & the users.
- A restructuring of Cart would help in improving the conversion funnel, by removing the unnecessary prompts that act as blockers.
- The changes to the multi-marketplace model would help us scale our tech better & in a much more efficient manner.
- After the restructuring deploys the atomic elements to the Search & Cart experience, newer features & solutions can build on top of that base.
Roles & Responsibilities
- Data analysis & pattern recognition about the user behaviour on our platform.
- Finding problems with the current solves, and breaking them down to their Atomic entities.
- Restructuring the Atomic entities to form compartmentalised solutions with those atomic elements which can be reshaped based on objectives.
- Setting up guidelines & base structure for the solution.
- Walking the people from Product & Design team through the various changes & evangelise the solution’s aspects so that they can build on it.
- Getting businesses on-board to the new solution & convey the impacts that are being predicted with the new solutions.
- Re-using UI components that can support the compartmentalised User Experience design.
With the evolution of the market, we had new opportunities coming up to add and launch myriad services to the product. These new services came with their own set of challenges, like logistic changes, business sustainability MOVs (Minimum Order Value), prescription validation, etc.
With each business solve, bringing a new challenge to keep a track of, it was clear that we needed to refine & remodel certain aspects of Flipkart to be able to support a super-app that can power multi-marketplace experience with ease.
Understanding the Problems with the Current System
During my time working on Hyperlocal Marketplace (Flipkart Quick), I wandered into the above-mentioned territory of problems, and it was evident that the current models weren’t apt enough to help us get the type of traction we needed. I knew the problem set well, as I was also helping Grocery Marketplace (Flipkart Supermart) for their experience solutions.
Since I had already faced the issues, that came out of a marketplace being force-fitted into a single marketplace supporting model. I looked at the pre-existing systems to suggest changes in the way we get things done for the new marketplaces. I was hoping to make major changes to the current model in a way, which would help Flipkart’s vision of having multiple marketplaces in their ecosystem.
There were 2 main areas of concern for us,
- Search — The ways a person can discover the marketplace, via browse, merch, or search.
- Cart — The core flow of the marketplace that is being launched onto the platform.
Problems with Search Experience
Flipkart’s search was designed to showcase items and did little to no effort to provide awareness to the users about the various marketplaces, and what their value propositions were. Users who would search a query would come across the products from all the marketplaces on a single search results page. And on clicking of the product, the user would be taken to the product page from the respective marketplace.
This was a concern, as even with just grocery (the only marketplace LIVE) at the moment of discussion, the grocery business team, was getting orders that were mixed with BGM category of Flipkart. This was, in turn, reducing the total amount spent on Grocery as a marketplace and was dividing the sum between Grocery & BGM. From the user’s POV, they were then needed to forcefully cross the MOV defined by Grocery, which was quite bothersome as per the research’s feedback.
Problems with the Cart & Checkout Experience
The existing cart experience had the following problems,
- If the basket isn’t above the MOV, the person first gets an alert to remove it, which in turn adds an additional click to the conversion/checkout funnel.
- The tech constraints around the current model, and how the multi-marketplace support is going to make the solution very complex to keep stable & maintain.
Our main objective from the cart experience was to make sure that it moves out of a regular checkout if the user has not completed the MOV bracket for any other marketplace (if there exists any), to make sure that the presence of any marketplace doesn’t impact Flipkart’s checkout as a whole. Since any additional click, due to a bottleneck that a marketplace runs into would impact Flipkart’s conversion rates in the long run.
The existing system had a flow as below,
Defining the change in User-flow
The major shift in approach that I was suggesting was,
New User-flow for the Search Experience
To fix this, we sat with the funnel PM, to understand his plans on the long run, and suggested the following changes.
We would be remodelling the search flow, to focus on finding the product first, provide Flipkart’s selection and then present the users with different marketplaces which could also offer the searched product. This would help the marketplace to sell their value propositions better.
Once the user shows interest in another marketplace by tapping on the redirection card, they would be taken to the search results page of the picked marketplace, with the search query already pre-fired.
New User-flow for the Cart & Checkout Experience
While refining the structure based on the data about checkouts happening at Flipkart, we moved towards the direction of pure-checkouts as a primary checkout option but also making sure that we suggest other baskets to check out if they are ready to be bought & delivered to the same delivery address.
We started chalking out a flow plan that can reduce the additional click, and make the system open for the future marketplace integrations in the long run. We came up with the following flow during our quest,
This new flow would not only help us avoid an additional click, but in case when multiple marketplaces exist on Flipkart, the user would not be bombarded with a barrage of alert dialogs. Also, it was able to solve for the issues that we had currently come across w.r.t. our Cart & Checkout Funnel.
Just a quick comparison between the two approaches also gave us a clear winner, as we were actually making sure that the marketplaces start getting priority in their respective areas, and drive a better conversion ratio.
Deploying UI onto the new User-flows
Search Experience
We powered a set of navigation-cards that had different promotion capabilities to allow the marketplaces with a better & more reliable communication with the end-user. Based on their needs, they can use either of the 3 widgets and optimise their user’s experience.
Cart & Checkout Experience
Expected impacts of the new Model
Expected impacts of the New Search Experience
The new approach would soft lock-in the user, this brings along a varied set of advantages listed below,
Experience Level Impacts
- Ability for marketplaces to create awareness about their offerings.
- People can be locked in and make the basket of all the grocery, or similar marketplace items together.
- Reduces the mixture of items coming from Supermart & Flipkart. And supports the user to complete the basket from the same marketplace.
Business Level Impacts
- Better up-sell & deals presentation in an effective way
- Chances of achieving the MOV are increased
- The marketplace items get a priority over the items that are not in the same marketplace.
Expected impacts of the New Cart & Checkout Experience
We are expecting to drive the following impacts with the help of the above changes to the cart & checkout experience,
Experience Level Impacts
- Less complicated solves to provide support for a bigger family of multi-marketplace experiences
- A better integration of mixed checkouts to allow the users to checkout multiple marketplaces together, if they wish to do it, without causing too many experience level issues.
- A cleaner way for Flipkart to plug and play new stores/offerings with the market before having to integrate it deeply with Flipkart systems.
Business Level Impacts
- An improved conversion rate, with the prompt going away from the system
My Learnings from this Project
This project required me to be well aware & informed about the Search & Cart based functioning of the product, and that allowed me to learn on how we as Flipkart were operating. I was able to connect to some of my previous projects with my previous employers, and look at the differences in solutions.
Flipkart’s systems were built for regular eCommerce before, and were now being optimised to support the expanding nature of its services, along with making sure that the product is understood & used by the n100Mn users from the different regions of India.
I was also able to learn on how to connect business flows together, and how a minor tweak in experience can help me restructure some of the core functionalities of the experience.
Note: I would have loved to add all the behind the scenes sketches that went during the process of envisioning this product. But due to the NDA, I can’t add those notes from my office desk and rough work.
