Live Video Shopping

Background

Bridging the gap between traditional and online shopping

Traditional online shopping often lacks the engagement and emotional connection that in-store experiences provide. Customers can feel uncertain about their purchases, leading to higher drop offs and lower customer satisfaction.

Live video shopping addresses this problem by bringing the in-store experience to the digital realm, allowing customers to see products in action, ask questions in real-time, get super-high discounts, and receive personalized recommendations, resulting in more engagement, informed decisions and increased trust in the shopping process.

When I joined this team, we had already launched first version of Live video shopping for users, but based on ~18Mn user base for which the feature was scaled when I looked at the stats it was directing otherwise.

Early designs

A known habit, a new place

The assumption was that Meesho's value-conscious shopper, often tier 2 or a smaller town, scanning for the best price on something they already want, had no appetite for live shopping. The evidence said otherwise: live commerce was already working for value-seeking audience for SEA apps and Indian audience from Tier3/4 routinely watched live selling on Facebook and bought from it.

Entry point to this feature were Whatsapp nudges, homepage banners (updates every 3rd day to ensure freshness otherwise CTR drops by 15-20%).

During my UX Audit, I realised among other banners available on the page, it was extremely difficult to differentiate this banner over scroll if someone is just skimming through content that too when we have a TG who doesn't read lot of text and rely heavily on image.

Learnings

Static banner experiment

We had ~160 unique streams/day with 10-12 streams at a time and seller would invest around 90 mins talking to users, showing their products and resolving their queries. Here are few insights I got from the Analytic team to understand what's working for users currently.

Minimalist design (focus on communicating only one message) gives best CTRs. Priority order of key attributes- discount > FOMO (time/ surprise) > PID > Live Video. Product is very important in channelising the intent, sets some sort of expectations in user's mind.

Brainstorming

Navigating through ambiguity

I organised a brainstorming session among all the collaborators including “D-level” employees to ensure I have ideas from entire team. The session was highly interactive and collaborative with more than 100+ innovative ideas across the entire live commerce journey. I collated all the idea buckets and further refined it to scope the problem and solution as per short and long term vision in collaboration with PM and Business Analyst. This activity also ensured we have a strong hypothesis to conduct a user research with diverse groups to find the right solutions for them.

Voice of users

I conducted structured research along with Product Manager across different regions of Delhi and Lucknow, these 27 users were based on their interaction on app, covering Live commerce cohorts- Interacted, activated and dropped off at different stages.

Based on all the insights we collated from everyone, these were the few problem buckets that were figured out.

Design iterations

Low fidelity

I explored different approaches like including cross-stream videos, adding "Shop a look" instead of one product shown, but we couldn't proceed with it since our audience was not ready for it yet, rather closed one video at a time view along with product variations. I presented the below low-fi ideas to internal team for broader alignment and feedback, and we decided to proceed with third option that is relevant from video format POV, and we also ideated the edge cases handling in case video doesn't load for users.

Option A

For deal seeker users, both discount and time-sensitivity are critical

Option B

Brings variety upfront for user's concern if there will be just 1 product shown in video

Option C, chosen

Provides clear visibility of video and consolidates the FOMO factors well

Wireframe

At this stage, the intent was to provide maximum space for video. I tested different layouts and how a seller would use their device to go live, and the vertical layout stood out from a feasibility and placement of product variants perspective.

Shipped

Final outcome

The home page stopped advertising the livestream and started being one. Where a static banner used to sit, the live video now autoplays with a strip of the actual products and prices underneath it.

Before

After

Making it load mattered as much as the widget itself. The video preloads from app open, and when the connection cannot keep up it falls back to a seller curated static image rather than an empty box.

Impact

What the redesign delivered

Awareness was solved by putting the live, autoplaying widget on the home page instead of an ad for it. Interest came from the product strip beneath the video, which showed the actual products and their prices and drove 40% extra traffic from the first-time experience. Recall followed from both, since a user who has lived through a live-only price drop has a reason to come back for the next stream.

It ran as a controlled experiment against a holdout before it scaled, and the first read came back positive on every metric we had set.

Phase 2

Point-in-time discount

With people now arriving at the stream, the next question was what converts them once they are there. A time-boxed deal drops mid-stream on a countdown, with prices falling in real time. The feeling designed for was not "there is a discount", it was "this is ending."

Designing the seller's half of the moment

As a discount that fires mid-stream has two sides, I worked on the seller end of flow as well. The seller is performer and operator at once, and cannot work a timer while selling live. The key account managers inform and train the sellers beforehand, and in fact a seller can decide how many such point-in-time discounts they want to provide for a particular stream from the seller dashboard. So the discount fires automatically, and a time-aware sequence keeps the seller ahead of each moment without requiring an action:

Before live

The seller knows which deals are included and when.

Window nears

"Starts in 3 mins" then "30 sec", lead time to prepare.

Once live

The buyer sees "ends in 10 mins". The seller narrates it.

What I'd carry forward

The highest-leverage call on this project was to show rather than tell, putting the live video on the home page instead of an ad for it. The product strip worked for the same reason: our user is visual rather than content-led, and needs to see the products before giving a video their time. Getting the whole team in early mattered just as much, because agreeing on shared goals upfront meant there was no misalignment left to resolve at handoff.

What I would do differently is map the failure cases before the pitch. Live autoplay was the hardest thing to sell internally, and that was partly self-inflicted, because I brought the concept to the room without having worked through what happens on a low-connectivity connection first.

Shop-a-Look was the bigger concept that didn't ship. The room couldn't be aligned on it, so I chose to disagree and commit to the focused solve instead, and the signal that came later showed users weren't ready for it either. The instinct wasn't wrong so much as early, which taught me that reading when a user is ready is as much of the judgment as the idea itself.