With the goal of boosting checkout conversion for the social commerce web-app, I identified users pain points through interviews and usability tests, and revamped the product listing page in 3 months. Below are the highlighted results:

+148.85% product listing visit

2x checkout conversion

+471.1% personalised cart shared with customers

1. Define users: who are we designing for?

As a newcomer to the team without any prior documentation of who our current users are, I facilitated an 1-hour proto-persona workshop with the sales, ops, and management to extract their first hand experience with our users, in order to narrow down from our database who should we focus on for the interviews.

The team came up with 6 personas, and I organized them based on their motivation, and translated them into 2 Jobs-to-be-Done:

1. Side hustler: to establish additional source of income beside their day job
2. Entrepreneur: to build their own successful business

With the JTBD cleared, we reached out to 10 in total users, 5 from each JTBC, to conduct interviews and usability tests with them. The result is surprisingly similar that both JTBCs are indeed the same motivation but at different stages only. They started as side hustlers, and eventually grew into entrepreneurs.

2. Define problems: what are the pain points worth solving?

After conducting user interviews, we laid down all the notes, and regrouped them into an affinity map. I identified the significant insights, and turned them into a user journey map below. It showcased their entire journey from side hustlers into entrepreneurs, with pain points they faced at each step.

To summarise, there are 3 main pain points:

1. Cost & margin: The social sellers have grown their business where workload seems overwhelming. They want to make enough to cover their hustles. The lower of stock prices enable more rooms for setting competitive price. Hence, more flexibility in profit margin. But where can they find such authentic stocks at a lower stock price in bulk quantities from reliable suppliers?
2. Reputation: Rapport is the greatest asset to social seller. With the increasing amount of customers, it’s become more difficult to provide personalised service, like finding the right products, which accumulates loyal customers. Less engagement might lead to a downfall to the sellers reputation.
3. Last mile: Packing, logistics, and arranging delivery have taken up most of their valuable time. Orders exceeding their handling capacity has caused delays and refund requests.

3. Reviewing current product through usability test

With the social sellers’ journey understood, I conducted a usability test for them to perform their sales activities with our web-app. Participants are asked to pick a skincare product for a customer with specific skin conditions. Then, proceed to close the deal. I observed how the sellers interact with our current app, and found 2 major usability issues:

a. Unusable search and filters:

• Search: When users type with keywords, suggested results do not match with the keyword. Also, the keywords have to be in exact order and spelling in order to show the particular products. There is no error tolerance at all. So users couldn't find the products they want.
• Filters: After unsuccessful search, they try to narrow down with filters. However, only brand filter is available. Customers usually have preferred brands and price range. When the social sellers try find suitable products for them, it's impossible to even locate the specific products.

‍

b. Mixed stock price and selling price:

• Stock prices are displayed in the same manner as selling prices. The profit margin is indicated with the confusing “Share $100”. Not only none of the prices add up, it has also misled sellers to perceive the SquareArk’s stock prices as more expensive than other suppliers. But in fact the price in black has the profit margin included. Sellers will only be able to know the stock price by deducting the profit margin (”Share $100”) from the selling price ("HK$530"). Users can’t tell stock prices and selling prices apart. As a result, users often find stock prices expensive, and would proceed with other suppliers.
• The toggle is confusing. What does it mean by switching a product on and off? It's a control for listing the product in the seller's store or not. Yet no one can tell what to expect when the toggle is turned on or off.
• Misleading copy: What does it mean by “Share $100”? Would I earn $100 when I share the product with my customers? Or when I sell? In fact, it meant by the profit margin.

4. Overlapping sweet spot

When we compare the user pain points with usability issues, highlighting our competitive stock offers are the most critical problem to solve first because stock prices directly tie with the sellers’ profit margin, hence their willingness to use our service. Low stock price means more profit margin, and more competitive selling price to offer to end customers. In reality, our stock offers were among the cheapest in the market. It is not a procurement issue, but how we convey and present the offers on our app that need a fix.

Personalised recommendations on products for specific customers is how social sellers gain their reputation. Customer engagement still heavily relies on the social seller, but helping sellers to navigate and locate specific products can be improved from our end, which allows them to get back to their customers precisely, building up the credibility.

In terms of the pain point of last mile, SquareArk business model tackles it inherently by providing stocks, warehousing, and parcels delivery services by default. The convenience can be emphasised through marketing messages and sales engagement.

Conveying stock prices clearly and facilitating users to find the right products are the prioritised problems to solve.

As I identified the problems to solve first, I began ideation with wireframes, and tested it internally. Then I prepared user stories with wireframes, and liaised with the tech team to deliver features on product listing page and navigation within 6 sprints (~3 months). Below are some of the key improvements:

Product listing page & product cards

There are 2 sets of information for the social sellers: stock prices offered by SquareArk to sellers and selling prices decided by the sellers to offer to their end customers. Stock prices is important to show how deep the discount is, and reflect how much room for margins. While selling prices matters in terms of how much profit the sellers could earn by selling the product.

Therefore, the product listing page is revamped with 2 tabs: product list and my store. The product list tab shows stock prices along with the discount rate and the primary action button, “Add to Store”. Users will always land on the product list tab, reviewing stock prices first. My Store tab is the overview of the seller’s store where selling prices and profit margin for every product added will be listed out. Sellers can also add products into specific customer’s bag to share with their customers to checkout directly. Hence, the product cards has been designed distinctively on each tab with its CTA button.

The flows goes from when the sellers identified a product that they want to sell, they can input the selling price and add the product to their store.

‍

Filters

The original non-collapsible filters only provided brand filters. It would be more useful to provide price filter and category filter, so that sellers can narrow down the scope according to the customers’ preferences. The filters has been redesigned with sufficient tap space, and easy-to-read format. Number of search results is added to inform users if there’s any product under their selected filters.

Search

The previous search algorithm didn’t work because it used to only recognise keywords in exact wordings and order with zero tolerance on error like typos. As beauty product names are usually lengthy, users often give nicknames or simplified names to these products. As a result, users can’t find what they want from the old search bar.

So I worked with the development team to modify the algorithm to recognizing partial product names, brand names, and categories. In addition, autocomplete and typo prediction is installed to give relevant results. I also liaised with the ops team to include product nicknames in their product upload process, so that the search algorithm can pick up product nicknames as well.