November 11 is celebrated as ‘Single’s Day’ in China. It’s as good as an unofficial holiday 😂 Nearly a billion customers (yes, Billion!) flock to e-commerce major Alibaba to buy a range of products. Nearly 300,000 brands vie for real estate on the app, and major discounts are doled out during this day. It’s THE largest shopping extravaganza on the planet.
Back home, our version of this extravaganza is called the Maha Indian Shopping League (MISL). This year's MISL lasted 5 days. It pales in comparison to our larger than life peers.
But hey, we’re just getting started 😎
In our endeavour to make it lucrative for customers, teams rally together to make this event as friction-free as possible. There are a hundred stories among our teams on how we make this possible.
From logistics, engineering, product, design, every little feature, tweak, ‘optimisation’ matters. This is one such story.
How a small feature helped increase our bottomline 💪
Tl;dr: In 3 days, we ideated, built and launched a feature to use WhatsApp to improve sales. This was after our sale day went live, and how multiple teams came together to iterate and move fast ☺️
Strap-in, let’s take you on a journey 🏎💨
Oct 6, 2021 — SALE DAY 🛒
We already broke multiple records on the opening day of our sale. As the day progressed, and we saw our numbers rise on mixpanel. The product retention team (my kick-ass team 😎) wanted more. What could we do to increase our numbers…?
We decided to explore WhatsApp.
WhatsApp has more than 500 million users in India. Unlike most apps, it has high engagement and once you have lock-ins, users check it multiple times a day. I use WhatsApp at least 20 times a day responding to messages. In our research, this was a pattern amongst our core demographic too.
We wanted to do something on WhatsApp to give our numbers a bump. Instead of a usual run-of-the-mill promotional message push notification, we decided to build a simple bot - one that would not look and feel like a boring monotonous ad 😏
The idea was simple:
Can we have a range of product categories that’ll showcase the discounts in each category. This way, we spur interest from within WhatsApp, and lead users to our app to make the final purchase. It’s a simple nudge to our users.
But, can we get bandwidth during a hectic, peak sale day to do something fun, something exciting, something we were not so sure about, but wanted to experiment with?
Well, in Meesho, the answer is…
YES!
We rallied multiple teams - Category, Growth, Marketing, to make things happen.
Day 1 had already ended. We had to move fast.
Oct 7, 2021 — Things are heating up 🔥
Speed over perfection - this is a value enshrined in Meesho’s mantras.
The first thing we had to lock in for the chatbot was the flow of the bot. Once a user sends us the opening message, what are the options we want to reply with?
We had to plan out the logical flow of the responses we wanted to present to the user. Once this was locked, we had to quickly decide which are the categories and products we’re moving ahead with.
Quick calls to resolve this.
Call 1 - Category team. The team loved it ✅
Call 2 - Engineering team — Go ahead received 😍
In just a couple of hours, a prelim version of the bot was built.
Still half baked, I did the user acceptance testing (UAT) for the first version, testing the flow and noting down the improvements required.
We encountered issues with enabling the chatbot, and getting the ‘interactive’ button (for example, in the screenshot given above, the ‘Top Deals’ button) to work properly, both of which took a few hours to solve.
One of the first problems we encountered — the interactive button itself wouldn’t work when sent out to a user.
Despite repeated clicks on the interactive button, the reply message wouldn’t be sent from the user’s side. Think of a faulty ‘close door’ elevator button!
All of these kinks were ironed out with a quick fix — we were good to go on the front end.
By the end of the day, we had it all figured out.
Oct 8th 2021 — Proof is in the pudding 🍮
9 am, we launched. The notification was initially sent to around 10% of our MAUs. We wanted to test user behaviour before a full fledged-launch. Within minutes we saw an immediate boost in ‘app opens’ and ‘bot replies’.
With the introduction of the chatbot, we observed an immediate lift of 50% in the clicks/user coming from WhatsApp that lead to app opens.
Overall, we saw a 1.3x uplift in clicks to our app because of our chatbot.
All in a weekend’s work, eh?
The chatbot gave us promising impact during MISL and Maha Indian Diwali Sale, and we look to continue to deploy the chatbot with more content and offers for upcoming sales.
Here’s how our chatbot functions:
We also reduced the number of users flagging our WhatsApp business account as spam by using a conversational, two-way approach to the chatbot.
Interacting with the bot increased trust, and we observed a reduction in red-flagging our number.
Oct 9 2021 — Sit back, watch the fun
Oh, didn’t do much hereon. The project was on auto-pilot mode and we were just observing the growth, and admiring our own work 😉
But we got some more ideas from user behaviours during this period.
What’s that, you ask?
Well, that’s for another story on fun, cool things we work at while in Meesho. Stay tuned on the blog for some fascinating insights from the hinterland 😍
The wrap-up
This year, we saw a 750% growth compared to last year. But that was expected given how we’ve scaled. 🔥
Here are two points that stuck out for me:
- 40% of our new users were first-time e-commerce shoppers.
- 80% of our orders were from Tier-2 markets. (60% of this was from Tier 4)
Yes, we’re teaching Bharat how to shop online.
Want to experience the chatbot for yourself? You can do so by clicking here!
A big thanks to the folks that made this chatbot a reality in 72 hours:
Engineering: Honey Duhar, Chetan Kalyan
Category Marketing: Ayushi Malik
User growth: Monmita Kalita, Aditya Chandratre
Product: Deepam Prithyani, and me 😉
Want to join me in executing such projects in record-breaking time? Head to meesho.io!
