I will add my suggestions:
1) Yes, realign the scale at the upper end of the quant score. The scale needs to be able to distinguish between the high scorers. It is slowly becoming like GRE, where in the old test 4% of the top scorers would all get 800. To do this they will need to introduce harder questions at the upper end to distinguish between the high scorers. The quant scores have been creeping up. The GRE redid their scales with the revised GRE which now differentiates better at the upper end.
2) In GMATPrep tests they need to display the time spent per question. Perhaps a graph that shows where the average timing should be and how it compares to the actual time spent to that point.
3) Add two more official computer adaptive tests, say as part of Exam Pack 2. That would bring the total official CATs to six, which will be a good number of practice tests.
4) Redesign GMATFocus so that they are full length quant diagnostics consisting of 37 questions and not the current 24 questions. Also, give a single scaled score for the results and not the range that it currently displays.
5) Update the Official Guides with official GMAT questions that have been tested since 2006 when Pearson Vue took over the writing of GMAT. Retire all of the old ETS/paper test GMAT questions.
6) Allow students to see the results of their official GMAT test. By this I mean they should show the sequence of questions and which ones they missed and how much time they spent on those questions. ETS already does this with the GRE exam, they have been doing it for a long time. There is no reason why GMAC cannot disclose this data.
That is all that comes to my mind for now.
Cheers,
Dabral