Random misbehaviour could be a viable way to expire a closed source/proprietary software/hardware product without inflicting too much damage to the own brands name.
If products expire in an abrupt way it is obvious that negative resentments will rapidly accumulate in product forums, comment sections and by word of mouth. But what if the expiration happens in a gradual way? And what if the onset and the severity of the presented misbehaviour are also determined by a function of increasing probabilty?
I would argue that the consumer, because he can not pinpoint the cause of the misbehaviour due to the randomness, will first come up with theories in which he views himself as beeing able to temprarily fix or bypass the malfunctions. The genesis of such beneficial theories could also be nurtured by intelligent algorithms. Because the consumer finds himself in a situation which is, compared that of sudden expiration, harder to describe, he will, I hold, not so readily spread information that would damage the brand. Done right the consumer could ideally be given the impression that the malfunctions stem from other, somehow connected/related devices. I would further argue that the consumers willingness to talk about the malfunction of a certain product decreases over time. This is why it should be beneficial to nurture the genesis of theories in which the other devices are the cause of the malfunction. That way the initial frustration and most of the negative communications during the "hot phase" of bad reputation generation are focused on multiple brands and/or technologies. When the frustration regarding the own brands device eventually reaches the replacement threshold, the willingness to further generate negative communications or dicuss malfunctions has already decreased and the own brand takes lesser damage.
Also it is valuable to note that the tech press would would find themselves in the same situation. Depending on the legal framework of the respective country they would face the additional obstacle of possible lawsuits when publishing articles based on guesswork.
To give a concrete example of this concept:
Let's say a company produces a bluetooth headset. The bluetooth headset needs to be initially "paired" with other devices before it can relay audio for them. When switching a device on and off, the headset needs connect anew but must not be paired again. The proposed planned obsolescence would work in the following way:
The first day the device gets paired with another device an interal counter starts. The counter will, after a one or two year period has passed (depending on the guaranteed warranty) activate the random obsolescence function (ROF). Good malfunctions provided by the ROF could be to require the consumer to start the connection procedure multiple times when switching on the device or even to require a new pairing procedure to make it work again. Malfunctions like disconnects could be introduced when several bluetooth devices are connected at the same time. Unexpected behaviour like connecting successfully but still not relaying audio could also be introduced. The ROF could then, from week to week, increase the probability of such malfunctions in order to continually frustrate the consumer and to incentivice the product replacement process. Since the ROF acts only on a few connected devices the consumer will suspect other devices as possible sources of the malfunctioning and thus will also talk about them when discussing possibles solutions or giving product recommendations. When the consumer is eventually frustrated enough and replaces the headset he will no longer so willingly talk and theorize about the cause of the missbehaviours. The consumer would ideally be left with the impression that the malfuntions did stem from the technology standards used (eg. "Bluetooth 3.0") or that they had to do with the combination of multiple brands.