Starbucks’ app issues are getting worse, and it’s costing them $44k every month

Starbucks has a very popular app. You probably know this. Maybe you used the app last week and joined in on the PSL 20th anniversary celebration. I had my one PSL for the year last Wednesday and it was as tasty as always. I’m not a hater.

‍

A sudden uptick in complaints about being unable to log in

The Starbucks app is well-rated: 4.9 stars from 978K ratings in the Play Store, and 4.8 stars from 257K ratings in the App Store. But there seems to be a consistent complaint from users that they are unable to log in to the app, especially after the frequent forced updates. And it’s getting worse. At Enterpret, we are passionate about customer feedback, so we ingested data from some of Starbucks’ public customer data sources, trained our models to identify what customers are talking about, and did some digging. Pretty quickly I found a distinct cluster of complaints touching on the interconnected topics of logging in, managing passwords, and updating the app.

‍

Have a look at the time-series chart:

‍

Starbucks customers care enough to provide feedback

Whoa. This is not a new set of issues, but the recent spike is dramatic. So what are customers saying? They are unable to log into the app, often as a result of a forced update, and often despite multiple attempts to reset their password:

‍

Something is clearly going off the rails. (As a side note, users are frustrated with these forced updates not only because of the technical shenanigans that often follow, but because they sometimes block transactions.)

‍

Enterpret can help Starbucks understand what customers are saying

Let’s take another cut of the data, by source, to see if we can figure out who these users are:

‍

This gives us an interesting perspective. The big spike in complaints in the middle of September was driven by Apple users… but then the pattern reversed, and in a big way. So… at least they’re spreading the pain evenly? As an Android guy, I suppose I find this kind of refreshing. I’m used to our app versions always being the noticeably buggier ones, or just not existing at all (I’m looking at you, New York Times Audio).

If we want to go even a layer below, we can dig into app versions. On iOS, the issues are happening most often in the three recent releases, with 6.57 driving the big spike we saw.

‍

‍

On Android, it’s the 6.58 version that’s the real killer – clear as day.

‍

The cost of ignoring the data...$44,000 in lost sales

What’s the impact of this? Here’s a little rough calculation.

• If we look at the period from September 1 onward, and focus on the app reviews (since the larger Twitter data is a bit noisy), we find that around 300 individual users mentioned one or multiple of the above issues
• The denominator: in that same time period, we ingested ~3,600 reviews from those two sources, so that’s around 8% of reviews where it came up
• Some small portion of app users actually write reviews, perhaps in the <5% range (according to a few blogs, as well as ChatGPT, who confidently gave me a similar figure but then declined to cite sources)
• Suppose, then, that around 75 thousand would-be customers tried the app for the first time in our time range, and a similar percentage faced these issues, so 6,000.
• If all of them threw up their hands and walked away from a $7.35 PSL, that’s $44k in lost sales, just in these handful of weeks.

‍

Would I swear by this calculation? No, but I’m a former consultant so I’ll say it’s close enough. If nothing else, hiring a few more developers is probably worth it, huh? We only worked with the public data, but I’m sure there lots of great information in their helpdesk tickets that would give dimension to these issues (Starbucks, holler at us).

‍

Well, that’s it for today. As always, this was a fun little deep dive exercise, and a few different interesting threads we could have pulled on. Phew, people have a LOT of thoughts about tipping 😳