Remarkably, this is often not considered. This is because marketing and IT are usually two separate silos. IT therefore often does not know when systems are being processed by large numbers of visitors, and marketing does not know whether the IT systems are designed for peak loads. And in most cases, no thought is given at all to the question of which technical conditions must be met in order not to get into trouble.
Can't it be done differently? Yes, but it requires close cooperation between marketing, development & operations (DevOps) and of course IT. Below I share three tips to ensure that smoking servers don't ruin your campaign ingloriously.
1. Know what to expect
If we look at the examples mentioned above, it is striking that there is always a similar pattern. Large numbers of people are mobilized to visit a website or app at approximately the same time.
An example of how it can be done is the hong kong mobile number search renewed app that was released in 2017 for the popular AVROTROS program 'Wie is de mol?', which keeps the Netherlands busy for weeks every year. The accompanying app was used by large numbers of viewers at set times, namely just before and just after a new episode. With the app, viewers can participate in pools with family, friends or (un)known people and indicate who they think is the mole. In this way, 'Wie is de mol?' became a cross-media experience for the 680,000 users of the app.
At peak times, the servers on which the app runs had to be able to process 5,000 visitors per second. In the five minutes after a broadcast, the app had to process more than 1.5 million user requests. If the servers are overloaded, the entire experience disappears. Given the large number of Twitter users among the app's users, that is a trending topic in no time . AVROTROS therefore opted for a different approach in 2017, in which the smart combination of cloud computing and DevOps played a leading role. I will come back to that later.
In any case, it is crucial to know when peak loads occur. This can be after a commercial on TV, when tickets go on sale or when your online campaign goes viral. So analyze your campaign and ask yourself if there are times when you can expect peak loads.