Because our specific breed of bugs could come from three different places, the complexity of our fixes are not multiplied but cubed. Could Amazon be providing bad data? Constantly. Could we have just found another Shopify constraint? Definitely. Or could these two complexities conspire against our servers to bring down the whole wobbly stack of cards? Definitely.
One customer came to us and said that his listings were missing. Why are they missing?! What are you all doing?
Over Slack, and across thousands of miles, I look at Nafaa ,my head dev, and shrug. What’s wrong Nafaa? As if we are looking for King Tutankhamen himself, we start digging, what could this be? How could this happen, we’ve spent 100’s of hours testing and coding. What’s wrong?
The code looks good, the API is up. Neither one of us has any idea. Then Nafaa, while going thru the reports Amazon automatically generates, notices something. The SKUs this customer is complaining about are NOT in his Amazon reports. What?! No way, I pull the reports myself and look.
Sure enough, this customer has inventory, that we can confirm, that are not on any of his reports. My heart sinks. If we can’t count on the data in the Amazon reports, what can we count on? With this bug, I start looking at my own reports, from my sellers account. And sure enough, missing SKUs, duplicate ASINs with different SKUs, as well as ASINs that are new, used or both. Shit, what are we going to do now?
This one offhanded mention, by one of our first customers, sends us sprinting to figure out a solution to fix a quarter of the entire codebase. We realize that the reports Amazon provides can’t be trusted, instead we need to pull the data from their API, but how do we know which SKUs to pull if we don’t have a list? We have to figure that out too. The solution takes us over 3 weeks to fix, at a cost of a couple thousand dollars. This software business is looking less and less profitable by the minute.
We get through that bug, only to find more coming at us. Shopify has a limit on the number of characters in a title, in a description, and even with the number of variations. I imagine the good folks at Shopify were in a meeting and the topic of variations came up. “Well, we have to limit it to SOMETHING, what does everyone think is a good limit? A couple people shrugged and threw out 100. Sound good?” Everyone nods and they move on.
Whelp, we have a customer who on one product has 178 variations. Yes, even I had a hard time understanding that one. 178 variations. What in the world has 178 variations??!
Regardless, those 178 variations broke their push. Shopify just wouldn’t accept it. That took us a half day of troubleshooting to figure that one out. But we squashed it with a satisfying crunch.
Thankfully the bugs have slowed down, and we’ve gotten better at identifying the ones that do come up. We’ve discovered that the technology we deal with has a personality. Shopify is like your solid big brother, who sometimes wears flamboyant socks. Someone you can count on with occasional flashes of creativity. While Amazon is like a wild-haired, wealthy mother-in-law. Someone who sends you on wild goose chases, and after a couple glasses of Chardonnay will slur her words, but for whatever reason you keep trying to impress her, and continue running as fast as you can.
There are a million different data points you can track in your ecommerce store. Clicks, conversions, bounces, visitors, etc. How do you know what they all mean and which ones to focus on?
Most people spend a ton of time getting the home page just right, but it’s vital that you get the product pages just as perfect.
One of the best ways to get more shoppers to your store is with publicity. Media outlets already have the traffic, why not let them send it your way?