In theory, the iPhone 4 was available for pre-order the morning of 6/15. Unfortunately, it didn’t work out quite as planned. As we have seen on the news, the pre-order experience continues to be plagued with timeouts, error messages, and confused employees. According to reports, the root cause appears to be AT&T’s authorization servers, which check customers for eligibility prior to approving the purchase. These servers are required whether you use the apple.com site, AT&T’s site, or even when visiting an AT&T store in person. Clearly, this kind of load is not something that most of us need to worry about, but there are still important lessons that we can take away from this experience.
Lesson #1: Monitor and test your entire ecosystem of dependencies
The most important lesson we should take away from this experience is the need to monitor and test all of your third parties. This is especially important for the third parties that are required for your key user flows to succeed. You need to test and watch these dependencies just as closely as you test and watch your own systems. The bottom line is that to your users YOU are down, and blaming your vendors will only go so far. In this case, the technically inclined might quickly surmise that AT&T is at fault, but the majority of customers have no idea that this isn’t directly Apple’s fault, and so Apple looks bad (in spite of being prepared on their end).
Lesson #2: Transparency
The second lesson is that transparency can go a long way in reducing the impact of an event like this, and potentially allowing you to come out on top. When events like this occur, a simple statement from the company or companies involved – along with some plan or timeframe – can go a long way toward alleviating confusion, frustration and hostility. Apple has done things like this before, with their real-time store-by-store inventory status for the original iPhone, which clearly helped their customers work around the limited supply and other issues Apple experienced that first time.
Lesson #3: Have a backup plan for your key vendors
The third lesson, which is always easier said than done, is to have a failover strategy if your primary vendors are down or unable to handle the load. In this case, Apple cannot switch to a different carrier, but they can potentially disable the eligibility checks and deal with the details offline. In other cases, if your primary cloud provider goes down, or your ad network becomes slow, having a backup platform in place that you can switch to when necessary gives you back the power.
The overall takeaway that I hope to leave you with is that the more interconnected the online world gets (e.g., clouds, APIs, web services, single sign-ons, etc) the more important it is going to be that you consider the health and scalability of every third party on which your service relies.
Comments are closed.