The Cheap-But-Often PC Buying Strategy

It all started on Thursday when I saw an offer for a free-after-rebate Soyo motherboard on DealNews (via the daily DealNews newsletter). Not that I was particularly in the market for new PC parts — I was trying to hold out for an Athlon 64 — but I was intrigued by this unusually good deal. As a first step, I looked up the retailer (Mwave) on ResellerRatings and found that they have a 8.73 rating along with a Gold consumer-excellence award.

From there, I mentioned it to Mike to get his thoughts on it (since he’s a professional IT Guy, I trust his opinion on this kind of thing). He confirmed that Soyo is a decent brand and he went on to say that his previous motherboard was even a Soyo (though he has an Asus now). Mike was even considering buying one for himself, until he read the offer details and discovered that the deal required a CPU purchase from Mwave as well.

Being the skeptic that I am, I figured that Mwave was overcharging on the processors to cover the costs of the motherboard. So, I cross-checked their prices against NewEgg’s. As it turns out, Mwave’s processor prices are actually lower than NewEgg’s (!). Mwave’s Athlon 2800 is $108 (compared to $116 at NewEgg) and their Athlon 3000 is $144 (compared to $166) at NewEgg.

At this point, both of us had all but discounted the idea of going for this deal. In Mike’s case, he already had a spare AMD 1700 and so he could have just about built a computer from spare parts and this “free” motherboard. But, with the required processor purchase, it would be far from a free PC for Mike.

And, in any case, I was only really considering this as a stop-gap measure to hold me over until Athlon 64s (Socket 939, specifically) became widely available. And, under that scenario, I would have bought a cheap Athlon 2200 or something to go with it (which would still work out to a decent speed upgrade from my current Athlon 700 for not much dough).

For those wondering why I’m even bothering with another PC upgrade when I already have a new PowerBook ;), I have my reasons. In the short-run, I still haven’t found an editor that I like on Mac OS and, in the long-run, I still need a PC for games (Doom III, Half-Life 2, and so on).

Anyhow, I got to talking with Mike about how we missed this great deal “by that much” (since it required a processor purchase). And, at that point, Mike shared with me his newly formed buying strategy which I like to call the “cheap but often” approach.

Summarizing my thoughts on the matter up until that point, it seemed reasonable to me that it would be cheaper to buy higher-end PC parts since they last longer between upgrades (as opposed to buying cheap parts which have to be upgraded much sooner). However, Mike expressed that, with the price difference between high-end and low-end parts, it's actually cheaper in the long run to buy cheap or mid-range parts, even if it means buying them more often.

Logically, his point of view made sense to me, though it was a little jarring at the time since it wasn't the way I was used to thinking. So, to prove it to myself (one way or the other), I decided to whip up a spreadsheet to calculate PC upgrade costs over time (using both strategies). As I made the spreadsheet, I held these assumptions:

• A “cheap PC” will last three years and, after that time, neither the motherboard nor CPU could be migrated to the new box. That is, it’s assumed that the CPU is maxing out the cheap motherboard’s capabilities already and that an upgraded CPU wouldn’t be possible.

• An “expensive PC” will also last three years; however, a refresher CPU could be bought at that time which would extend the life of the box an additional two years.

• And, lastly, each “new” box would built by scrounging parts from the “old” box whenever possible — I wouldn’t be building a completely new machine each time.

And, for the spreadsheet to make sense, here’s what I have in my current PC with notations of which parts can’t be reused in the new box:

• Asus Slot A motherboard [nope]
• Athlon 700 [nope]
• Western Digital 120 GB hard drive
• Maxtor 10 GB hard drive [probably not]
• Plextor CD-R drive
• Pioneer SCSI DVD drive
• GeForce4 Ti4600
• Floppy drive
• D-Link NIC
• 1 GB PC3000 RAM (bought earlier)

With that out of the way, the spreadsheet is available in OpenOffice.org Calc format (7k) and PDF (62k). You may notice that the “now” section for the “Expensive Route” includes both an SATA drive and a DVD+R/DVD-R recorder (as both of those parts would allow the expensive PC to laster longer between upgrades). Not to worry, the “Cheap Route” gets those parts also, though only as needed (as part of the 3-year “Cheap Route” upgrade where it’s assumed that the motherboard of that time will predominately feature SATA controllers).

So, as you can see from looking over the spreadsheet, the cheap route is cheap not only in the short-run (as expected) but also in the long-run (which I didn’t initially expect). With the exception of Year 3, the cheap route remains cheaper than the expensive route right through the 8-year projections (even though it requires replacement parts sooner).

So, with the numbers in-hand, I ordered the Soyo motherboard and an Athlon 3000 to go with it. With FedEx Saver shipping, they should be arriving within the next few days :).

Update 03/31: The original spreadsheet had an error in one of the formulas, but that’s been corrected. The end result (the “cheaper route” being cheaper in the long run) is the same.