How would you design a feature to reduce Amazon's order return rate.

Clarify 

Are we focusing on Apps or Web ? - Both

Any particular platform like iOS or Android ? - Both

Any Particular Geographical region ? - All 

Anything else I need to know ? - No

Obj - To reduce Amazon’s return rate

Customer Segment

1. Customers those who buy product and return immediately

2. Customer who buy and return product after 1-2 use(1-2 weeks)

3. Customers who buy and return products after multi usage(just before the 1 month expiry).

I will use data to evaluate which segments cost us more but for now I will pick segment 2. 

Pain Points 

1. Certain products are expensive but customers need the product for 1-2 occasions such as events like a party, weddings, graduations etc.

2. Certain products are hard to choose without trying them out or physically evaluating them (ex-Shoes).

3. At times customers buy because of recommendations and offers but later lose interest and return the products immediately.

Again I will use data to evaluate which of the pain point is most prevalent or important, but for now I can choose either #1 or #2 since I feel both of these cases are pretty common. 

Solutions

1. Amazon Rental Service - For certain products that are used once in a year or a few yrs like Wedding Tuxedo, Graduation Gowns etc, you can rent those items from Amazon. Can be operated the same way we do delivery and return, except we will charge customers for renting items 

2. Buy and Sell or Refurbished - For Certain Products you can buy from Amazon and you are eligible to sell it on Amazon within a certain time frame and a discounted price. The more the number of times the product is sold, the selling price reduces further.

3. Multiple Price Options - For certain products that are susceptible for returns we will have different price tiers such as reduced price with no returns allowed(Flights do so with no cancellation)

Tradeoff and Priortization

Solution 1 - High Effort and Medium Impact. Harder to Implement, execute and operate at scale for a company like Amazon. Plus there are already rental companies and options available for customers and we will be competing directly with them. So I will not consider this at this point.

Solution 2 - There are used items already being sold on Amazon by our 3rd party suppliers. I don't think this feature will have much impact on our customers. Low Effort and Low Impact.

Solution 3 - This is an innovative solution and I haven't seen this feature or concept being used significantly in the retail/ecommerce industry. This is a Low effort and high impact feature. I will launch it for certain expensive items on Amazon where Margin is high. Launch it in a particular Geo location and test it.

Success Metrics 

1. Reduction in return orders for that particular category(Both no of items and amount)

2. CSAT Scores or NPS Scores(to measure how happy the customers are with this feature)

3. Increase in revenue in this category.

4. No. of items purchased(at a reduced price) with no return policy

#1 is my north star metric

Clarifying Questions:

Are we focusing on increasing the revenue of Amazon by reducing the cost of return or improve customer satisfaction? - Both

Are we focusing on Apps or Web? - Both

Any category of products that has higher return rate? – Spread across, no specific category.

Any specific surge in return or a regular spread across the period – Across the period of time, no specific surge in return

Any specific vendor/service provider where return is high - No.

Any Particular Geographical region? - All 

Any specific time period after the purchase that the customers are returning? – No, equally spread across, 1 to Last day of return eligibility.

Any specific age category that is found common in return category - No.

Anything else I need to know? - No

Goal - To reduce Amazon’s product return rate and improve customer engagement/satisfaction.

Customer Segment

1.    Regular retail Customers who buy product and return immediately.

2.    B2B Customers who buy in bulk and return it.

3.    Retail Customers who buy and return product after 1-2 use within next 1-2 weeks of purchase.

4.    Customers who buy and return products just before the return eligibility expiry (after multi usage).

As we are focusing on more general outreach, I will focus on segment -3 as cost involved in return pick up is higher considering the number of returns are high even though the number of return items from each pick up can be low. 

User Journey:

1: User login to Amazon app and selects a product.

2: Checks out and pays for the product.

3: Delivery personnel delivers at customer doorstep.

4: Customer uses it for 1 time and checks it does not suit the criteria and raises a request to return it.

5: Or Customer uses it multiple times and returns it just before the return policy expire date.

6: Delivery personnel picks up from customer doorstep.

Pain Points: 

1.    User selected a product in Amazon, but the product received is not as per the original selection criteria including colour, size, design etc. Service provider might have provided a different design/size/colour/number.

2.    Customer bought the product and used it once or twice but looks like, he does not need it anymore. The customer thinks of returning it back as he does not see any further value.

3.    Certain products are hard to choose without trying them out or physically evaluating them (ex - Garments).

4.    At times customers choose as an impulsive buy and after realizing, return the products immediately even without using it once.

5.    Damaged product and hence a genuine return.

As focus is to reduce the cost of return and improve customer satisfaction, I will choose either #1 or #2 since I feel both cases are common. 

Solutions

For #1 pain point,

1.    Feature to share the actual product being shipped-

So that user can see the product being shipped if not exactly same as shown in picture. The service providers like cloudtail should have ability to upload the exact picture only if the design/colour/size differs than shown in original picture which user selected. This comparison can help the user to cancel/select a different product/ continue with current selection. This will save the cost of later cancellation.

2.    Call out with a Quote - Feature should call out that original product might differ a bit (which they call out now) and there could be additional return charges if user returns after 3 business days after delivery. So ideally user should return it with in 3 days if they don’t like the new design/colour/size of the product.

For pain point #2

3.    Rental Service - For certain products that are used only for a very limited time, users can rent those items from Amazon. Can be operated the same way we do delivery and return; except we will charge customers for renting items for an agreed upon duration.

4.    Re-sell or Refurbish - For Certain Products users can sell it on Amazon after usage for a certain period at a discounted price.

5.    AI modelled Multi-Tier Price Options – For certain products, create AI models that will suggest the multiple price options. Guaranteed no return (lower rate), possible return in next 3 days (5% higher cost) and possible return within expiry date (10% higher cost).

Evaluation and Prioritization

 So based on above evaluation criteria, I will invest in 2 solutions for called out pain points. 

Call out with a Quote and the other one AI modelled multi tiered price options. Both these options will have low-medium implementation effort and will help in mitigating the pain points.

Success Metrics 

1.    Reduction in return orders for a particular category

2.    Increase in revenue in this category.

3.    No. of items purchased (at a reduced price) with no return policy.

Launch strategy should be staggard approach starting with a specific geo location and later expanding it further in iterative manner.

 Summary :

We started with the goal of reducing the number of returns for Amazon orders impacting the overall revenue and customer satisfaction. We selected the user painpoint of not receiveing the selected product as one of the pain points. The other pain point that we realised was because the user does not want to continue with teh product in teh long run. We identified with simple solution like call out that the product could differ and incase user still proceeds with purchase, he may be charged a nominal percentage on return.  the other solution was powered by AI model, where items with most return rate will be offered with multi tiered price options. We closed the discussion with teh success metrics and high level plan of lauching the feature.