For a retail store, users are frequently complaining about long queues during check out. How would you go about coming up with a strategy to improve the experience in short term and long term?
Clarifying Goals: Why should we care?
1. Users being dissatisfied with the checkout experience might stop being loyal customers.
2. What is the source of these complains? - Let us assume these are daily reports of organization employees, in-store surveys that checkout time has increased to the extent that this is a real problem
3. Do we know how much the time has increased and how does it compare with average store experience? - the time based on informal surveys and instore measurements has increased upto 15% or more. It is unclear what the impact on customer retention is.
4. Do we know if there are days of the week where we have a checkout bottleneck? - for e.g. Saturday, Sunday are usually busier? - no, the wait time seems to have been unpredictable.
5. Do we know if the number of customers we are processing through checkout has increased over normal trends? - no, the number of customers have not increased.
6. Do we know if there are any other events / product introduced recently after which the checkout times have increased? - no, nothing extraordinary has happened.
7. How has our checkout capacity has changed over the last few months? - not much, same number of people
8. Has the average order size increased that might lead to longer checkout time?
Summarizing, the checkout problem:
The number of customers have not changed much, the number of items per order has not changed. So the checkout load has not changed much. The number of checkout gates provided has not changed either, so the processing capacity has not changed much. Therefore, something in the process of checkout has changed - either in the type of items being processed, how they are being processed.
Tier 1 Metrics: Customers processed / hour, Customer Average Wait Time, Average Order Size per customer
Tier 2 Metrics: Order mix (types of items)
Investigating root causes, we might potentially find a few indications to the problem such as:
1. Customers are purchasing more vegetables --> implies that manual checkout is used by customers for these items
2. Register is finding defective items such as items missing price tag or bar code not working --> implies manual look up
3. The number of returns per customer has increased substantially --> registers have not traditionally handled returns and have to get help from managers to process RMAs
The problems fall into the following types:
1. A processing function has a defect
2. A processing function is becoming a bottlenect due to change in the mix of items
3. A processing function has had to process things that it is poorly equipped to handle
Solutions could generally be:
1. Add more capacity for processing
2. Optimize current processing
3. Add new processing capacity
4. Set/reset customer expectations
Depending on the type of problem and how long it takes to fix, we can recommend different short term and long term solutions:
For e.g. root cause #1, customers are purchasing more vegetables due to better pricing. Is a good thing for business. This is not going to change shortly, at least the business hopes not. Short term - Add additional store clerks to help with increased load, Long Term - create new packaged vegetables tagged with bar code / pricing to ease auto checkout, Improve self-checkout counters to add vegetable catalog, weighing and checkout process.
For e.g. root cause #2 - register is finding defective items - this is unexpected defect in the process. Short term - identify items with higher than average defect rates, sweep through store inventory, arrange for new suppliers to supplement stock. Long term - strengthen in-house quality control and supplier requirements
For e.g. root cause #3 - returns per customer has increased substantially --> short term - create dedicated temporary returns registers to ease load on main checkout, long term - monitor return rates to automatically prepare stores for increase in returns
Finally - problem agnostic long term,
1. customers standing and waiting in queue is a poor experience. Customers are much happier if they are prepared mentally for the queue time and staging. A simple token based experience can be implemented where people are called up to a checkout based on their token number.
2. create online pickup experience for commonly ordered items so that most customers can avoid waiting in queue.
3. create online concierges who can fulfill an order and deliver it
4. Implement metrics to measure checkout experience as above, and leading metrics to predict checkout problems