In reflection, as a data analyst, I wonder how an analyst can do to help the company reduce operational failures and cost, and increase efficiencies and customer satisfaction. Read on if you wish.
The incident
I ordered some coffee from Amazon and selected the one-day delivery service on 2/14/2019 Thursday. The merchandise is supposed to arrive on Friday.
Update on 2/29/2019: The package is lost, I am told on Sunday. I published most of the post on Saturday. Quite hilarious.
To give some background first, I live in a big apartment complex in San Francisco Bay Area, not some remote place. We have a management office that can accept deliveries. Its hours are Mon-Thu 9:30-6:30, Fri-Sat 8:30-5:30 and Sun closed.
So the above tracking history gave me some insights into the failure to deliver on time.
1.It took 1 hour 20 min for package to go out for delivery. Is it too long? I don't know.
2.Our office closed at 5:30 on Friday. They went out for delivery at 5:28. Seems they failed to know our office hours. We are a big apartment complex. There must be a lot of deliveries from Amazon in the past and they still don't know the office hours. That's going to fail them more and more in the future.
3.At 6:35 pm most businesses must be closed. They should have known one hour ago or even earlier. No delivery should be attempted.
4.The most perplexing thing is, the apartment business is like no other business: I live inside. They could have just given me a call so that I could have come out to pick up the coffee. Just as simple as that.
5.(Updated on Monday 2/19) Where did the package get lost? Yes, it's sent from a 3rd party. Then it got lost by the delivery driver? Apparently it is already in San Leandro of the Bay Area.
How to fix the issue analytically?
1.Have knowledge of the apartment office hours
This could have helped estimate the arrival time against the office hours. Then make decision to go out or not. Apartment complex has a large number of residents. It is worth while to have that knowledge on file.
2.Make rental apartment a special business category
It's like no other business when the office is closed. The person who ordered the merchandise is living inside. When late in the day and the management office is closed, the person may be already at home. Why not just give him a call. The analytical result should have suggested this option to the delivery driver, or train them to do so in advance.
3.Take notice of repetitive failures
This is the second time the delivery failed for exactly the same reason. No wonder. If some failure happened multiple times, there must be something wrong worth notice. You don't want to fail again and again. It is easy to take out repetitive failures in the data set.
4.Be suspicious of KPIs
A KPI is an aggregated number like 98% delivery success. Is that a satisfactory ratio? What happened to the small percentage of failed cases?
If some analyst at Amazon dug into those cases, he may find the root cause of the failure. This requires him talking not only to the delivery driver but also to the customer who didn't receive the merchandise on time, in order to get the full story.
Note that a KPI can be quite artificial. We should always be suspicious of any KPI's effectiveness as a good business indicator. After all, our business goal is to satisfy not KPIs but the end customers.
Conclusion
Don't take me wrong. Amazon is a great company that does great things. Being a data analyst, I would like to draw some lessons from those incidents. This may help myself in my analytical methodology. Also, if someone from Amazon can see this, they may think about how to streamline their operations. Hope that my orders can arrive on time and are not delayed for the same excuses again.
Add a comment