Converting Window Shoppers

state & condition

Illustration by 8box — *Illustration by* *8box*

2 – 3 nights to complete the task

I was in the middle of a recruitment process and got into the assessment stage.

The task actually had a longer time to complete, however, I was not informed by the recruiter when the email was sent.

I saw the email on Monday afternoon while the deadline was on Friday night; thus, I was left with less than 5 working days.

I was given 2 tasks to complete and the other task required around the same amount of efforts. Therefore, I had around 2–3 nights to complete this task while working for a full-time job during the day.

brief

Push window shoppers to make purchases

The brief was to push window shoppers who have no urgencies or needs to make purchases in a C2C e-Commerce.

interpretation

Characteristics

Window shoppers = browsers

I used the definition in Designing for 5 Types of E-Commerce Shoppers for window shoppers to understand their characteristics:

“Browsers are leisurely shoppers who go to their favorite sites or new sites for inspiration or to kill time. Browsing shoppers want to stay up on the latest trends, dream of future purchases, or be prepared for the next shopping trip, or looking online in advance of a visit to a store. They want to see what’s new, what’s popular, and what’s on sale.”

Behaviour

Adding to cart to flag items of interest

With the assumption that window shoppers will flag items of their interests for future purchases, I decided to focus on 1 common behavior of flagging items of interest, which is adding the items to the cart.

This behavior was chosen because items in the cart are only 1 step away from checkout page.

Problem

Lack of continuity from cart to checkout

The problem is the lack of continuity from cart to checkout. The window shoppers simply leave the items in the cart.

The goal is to convert the ‘abandoned items’ to purchased items while making it as targeted as possible to window shoppers.

possible solutions

Reminders to make purchases

Simply giving reminders to make a purchase to window shoppers will not work because we have an assumption that they do not have any intention to buy the items right away. They put the items in the cart for future references.

Notifications if a few stocks left

We can’t control stock numbers in a C2C marketplace — if the stocks are never drained, we won’t even get the chance to do this.

Well, we can manipulate the numbers but our trustworthiness is at stake which is not worth it.

Discounts for the abandoned items

This might be a promising solution. This certainly comes with a caveat that the discounts have a short time limit to evoke senses of urgency and not to leave the window shoppers with too much time to consider.

implementation plan

Check the rate of abandonment cart

Before even starting, the first thing to do is to check the rate of abandonment cart.

There was no point of doing this if nobody ‘abandons’ the cart.

Design the MVP 0 and conduct some usability test(s)

Design the feature with the least effort.

Then, we need to make sure users get it before releasing it.

Do a lot of a/b tests and regularly analyze the data

Many aspects need to be optimized with quantitative data such as the copywriting, notification timing, time limit, how long the product should be in the cart before notifying users, type of products, how often the discounts should be given, etc.

Beware of any online and offline events such as Harbolnas (similar with Black Friday) and public holidays (Eid al-Fitr).

Improve the feature

Increase the complexity of the feature one at a time

The aim is to create a system where the sellers will be the ones who give the discounts which will increase the unpredictability.

feature description

Make as if the discounts are specifically for them

Make it as if the discounts are given to specific users by using the users’ names as the voucher codes, just like referral code, and the code should not be able to be found anywhere else.

Focus on high-margin fashion items

Since we target the window shoppers, we can focus on the fashion items as the initial target.

There should be a lot of people in fashion industry who want to see what is new, what is popular, and what is on sale. They want to stay up on the latest trends and dream of future purchases.

Besides that, we want to make a profit, so the items should have high enough margins for discounts and fashion items have quite a big range of margin.

Higher minimum transaction amount than the item price

We need to have a discount’s term where users need to reach a minimum amount of transaction that is higher than the item’s price and much higher than the discount amount in order for users to get the discount.

We need to do this to get some profits.

Short expiration time for the discounts

A short time limit for users to use the discounts are necessary to evoke a sense of urgency, as well as not to leave window shoppers with a lot of time to think.

marketing

App notification as the channel

Channels to notify users are pretty equal in reach and cost:

App notification: low cost, low reach
Email: low cost, low reach
SMS: low cost, low reach
Google/Facebook/Social Media Ad: high cost, high reach

And, among those 4 channels, app notification channel usually reaches the power users — not all users allow app notification in their phones.

Since there is no point if this does not work on the power users, we can start with the app notification first.

Send on weekday during lunch time

Since there was no previous data of what day and what time with the highest open rate, or even any other users’ data, we need to:

Check the user persona and predict the users’ free time
Do a lot of a/b tests fo the optimal timing

User Persona

Let’s decide on the user persona for window shoppers then …

Career women in their 20s who work from Monday to Friday (9am–6pm) and commute by public transportations.

Weekday or weekend?

Predict the free time but not too much free time:

Too much free time on the weekend will give users too many considerations when they are making purchases.
Weekday either during lunch time or after office hours is a better choice.

Lunch time or after office?

Lunch time is a better choice, because:

They can switch to laptop to make purchases (some people do).
Users will be cramped in public transport after office hour.
If they forget/miss it, we can notify them after office hours.

The items should sit long enough in cart

Give long enough time to ensure users do not have any intention of making a purchase to be sure that they are window shoppers, but not so long that the interests to the items subside.

We need to do a lot of a/b tests to find the best time.

But, what’s the rule of thumb for now?

Based on Abandoned Cart Emails Strategy Timing, it’s 60 minutes after users’ last interaction with the site. However, since we aim for window shoppers who have no intention to buy right away, the timing should be way longer than 60 minutes.

high fidelity design

This is not the MVP 0

The MVP 0 should be as simple as giving promo code to users and ask them to fill in the code by themselves. However, as this is a product design exercise, I went a little further.

Note for the data & engineering team

Track the CTR of app notification, abandonment cart rate, and checkout rate
Track the funnel and drop off rate
Use name of the user for the promo name to emphasize the personalization, e.g. BELLA50.
For simplicity, discount applies to one store, not per product.

outcome

2 interview invites and a job offer

This exercise got me 2 interview invites to 2 different companies — 1 company did not want to wait for me to do another exercise so they used this for the assessment.

I turned down 1 invite though because I had accepted a job offer from the other company.