From time to time rumors surface about Apple being poised to introduce some sort of cheap iPhone to fill a hole in their lineup, or we have so-called experts pontificate on why Apple needs to introduce such a device in order not to leave that market entirely to Android. The ways this gets discussed gets me wondering how these people could possibly not have noticed that Apple still sells the iPhone 4 and 4S, now as lower cost iPhones; I don’t know, maybe these don’t count because they have cooties? In reality, while they were introduced some time ago indeed, they can stand the comparison with “base” Android models as far as I can tell, and buyers do not seem to be snubbing them.
If at least the conversation was about whether the iPhone 4 and 4S were appropriate base or low-cost models for Apple to sell, as then there would be things to say on the matter (but no, it is always framed as if these two did not exist). Indeed, I think this strategy was justified when the iPhone 3G kept being sold along the new 3GS, it might have been tolerable to keep the iPhone 3GS along the new iPhone 4, but by now Apple should have long changed strategy. For a number of reasons, as the iPhone product lineup, or rather hardware platform, matures it should instead have low-cost models meant for that purpose.
The first reason goes back to the dark days of the Mac in the nineties, where former top of the line Macs would be discounted and sold as base models when another Mac model replaced them as the top of the line Mac; as a result, people would be hesitant to buy the (more) affordable Mac which they knew was not really up to date, and did not want to shell out for the current model, so they ended up just waiting for it to be discounted. It was hard for Apple to have a clear message on which Mac someone in a given market should be buying: so yesterday that model was not appropriate for consumers, but today it is? The heck? Fortunately, Steve Jobs put an end to that when he introduced the Mac product matrix (with its two dimensions: consumer-professional, and portable-desktop, and four models: iMac, PowerMac, iBook, and PowerBook).
Which brings me to the second reason, which is that not all technologies make sense as being introduced for professionals exclusively, even at first; USB, introduced in fact with the first iMac before it was on any PowerMac, is a prime example. Today, we have for instance SSDs, at least as an option or in the form of an hybrid drive.
But I think the most important reason is that having dedicated base models would allow Apple to sell devices where the hardware design flaws of yesterday (visible or invisible) are fixed, instead of having to keep having to take them into account in the next X years of software updates. While some new features in iOS releases have not been made available on older devices on which they run, the base OS has to run nevertheless, and even with this feature segmentation performance regressions have been observed. The other side of having specifically developed low-cost iPhone models is that Apple would seed with these devices current technologies to better be able to introduce new and interesting things in future versions of iOS (think, say, OpenCL), for instance because third-party developers are more likely to adopt a technology if they know every device sold in the last year supports it; this goes doubly if the technology cannot serve as an optional enhancement, but instead is meant to have apps depend on it.
The example the iPhone should follow, where Apple itself does it right, is with the iPad and in particular the iPad mini. I joke that now with the 128GB iPad there are 48 (count them) iPad SKUs, but that’s in fact OK, as the lineup is very orthogonal: from the consumer’s viewpoint, color and connectivity are similar to build-to-order options over capacity variations of the 3 base models; it must be somewhat challenging to manage supply and resupply, but apparently the operations at Apple is managing it; and sometimes the one combination you want is out of stock at the store, so you end up getting a slightly different one, but that’s minor in the grand scheme of things. On the other hand, the introduction of the iPad mini was indispensable to diversify the iPad presence and make the iPad feel like not just a single product, but a real hardware platform.
The one thing in particular that was done best in that respect with the iPad mini is that internally, the iPad mini is pretty much the iPad 2 guts with the 4th generation iPad connectivity: Lightning port, Bluetooth 4.0, and LTE as an option. I/O is one of these things where it often does not make sense to introduce technologies only at the high end at first, because of network effects: they apply whether the I/O is to connect devices with each other, in which case you need to clear a penetration threshold before people can use it in practice, or if the I/O is for accessories, in which case hardware accessory makers are more likely to follow the better you seed the technology to everyone.
Now I’m not saying it is going to be easy to do the same for the iPhone. It is clear that each iPhone hardware model is very integrated, requiring for each model a lot of investment not only in hardware design, but also in the supply chain and the assembling infrastructure. Designing each model to be sold for only one year would make it harder to amortize these investments, but planning for some hardware reuse during the design process could compensate to an extent, and I think the outcomes in having a clearer and stronger product lineup, better technology introduction, and decreased iOS maintenance costs, would make it worth it.