PC Repair Rip-Off!

Service Test: Defining Our Approach

PC World doesn't often go undercover. But that was the only way to find out how consumers were really being treated by the big chains' repair services. Here's how we did it: First we purchased 20 identically configured, refurbished Compaq Presario 4122 PCs--using one for each of five stores at each of the four chains. (Compaq was not consulted on our test planning or in the development of this article; this make and model was chosen merely because of its near-universal acceptance by repair shops nationwide.)

Before setting up our test problems, we verified that each system was in perfect working order. We reformatted each hard drive, then reinstalled Windows 95, along with several applications and dummy files.

How We Tested

Posing as typical consumers, we presented the following tests to the stores:

Problem 1. To test phone support, we sabotaged each PC's display by renaming the video driver. A good tech should be able to diagnose the problem over the phone and guide a user to a solution: reinstalling the driver.

Problem 2. On 15 of the 20 systems, we damaged the IDE hard drive cable by cutting several of its internal wires. The damage was not visible, but the system wouldn't boot. Correct solution: Replace the IDE cable. On the remaining 5 systems we disabled the CD-ROM audio by cutting the sound cable. Correct solution: Replace the cut cable.

Problem 3. We moved one SIMM in each machine to an incorrect slot, reducing usable system memory from 16MB to 8MB. We didn't tell the technicians that the memory seemed faulty. The PCs ran so slowly, and the wrong order of the SIMMs was so easy to see, we believe a trained person should readily spot this problem.

How We Graded

We evaluated stores by measuring key factors in the consumer experience, assigning points according to relative importance. The best possible score was 100 points.

• Accuracy (20 points maximum in store, 55 phone). Did the store diagnose the problem correctly and without confusion? If so, it scored big; in the few cases where a store got close enough to identifying a problem without clearly defining it, we gave partial credit.
  
  
• Quality (30 points, in store only). Did the store fix the problem? This is the acid test of repair experiences, of course. Without a fix, no store was graded above F. We penalized stores that tried to sell unneeded parts or services.
  
  
• Cost (20 points). Was the final bill fair and consistent with estimates? If so, we gave high marks; overcharges were penalized.
  
  
• Efficiency (15 points). Did the store meet its own completion estimates? If so, we graded it high, even if the PC spent a long time in the shop. But if a store wasted our time with sluggish repairs, communication breakdowns, or long waits on hold, it lost points.
  
  
• Helpfulness (10 points). Were employees instructive and courteous? If so, we gave them their due. Stores with terse, uncommunicative, impatient, or rude techs were marked down.
  
  
• Misplaced SIMM (5 points, in store only). Did the store spot the problem and fix it?
  
  

We gave each store an overall grade based on these criteria, then averaged grades for the five stores within each chain to get a single letter grade as follows:

A = 90 to 100 points

B = 80 to 89 points

C = 70 to 79 points

D = 60 to 69 points

F = less than 60 points