The single biggest thing to fix is your timing problem. That pulled your score down in both sections—a lot. And you *still* scored V30 and Q48 anyway! If you can fix the timing problems, you'll likely increase your score another 40-ish points—without even learning how to do anything else. (And if you don't fix the timing problem, you won't get to 700+ no matter how much more you learn.)
Look back over those earlier problems in the section. Where did you spend time that was, in hindsight, wasted?
—You spent more than 1.5 minutes when you knew the whole time that you really didn't know what you were doing. Next time, admit this to yourself and cut yourself off much faster.
—You spent >>1 min about the average for that type (regardless of whether you got it right or wrong). For instance, let's say you spent 3.5m on one quant question. Even if you got it right, you spent nearly as much time as you should have spent on 2 questions, not just 1. So that 1 question wasn't worth it.
—And don't forget to include problems you got right but you realize you just got lucky.
For each one of those, start to learn how to make decisions factoring in your timing. When should you bail immediately? When should you cut yourself off after a while—how will you know that the odds are not good / your ROI is bad and you should let go?
Have you read / seen this yet? It will help you to answer those questions.
https://www.manhattanprep.com/gmat/blog ... -the-gmat/(Try the webinar that's linked at the beginning of the article.)
Before you get to the real test, you *must* identify some small number of questions on which you know you will guess immediately when you see them. You won't even try. Do this even if you see one of these questions early in the section (and you're not behind on time yet). Do not wait until you are already behind to try to fix the problem.
The last time I took the real test (two months ago), I did this on 8 questions in the quant section (I literally did not even try to make an educated guess—I wanted to see how high of a score I could get if I didn't even try on these questions). I still scored a 48. So if you do this on, say, 4 questions, you can still get a 50 / 51. (I have had students do this on 4 questions and still score 51 on quant.) And 4 questions = 8 minutes, a BIG difference at the end of the section.
Since verbal is your lower section, I would do this on maybe 5-6 questions there. What do you just hate on verbal? Don't even try—guess immediately and move on.
Also, don't do sets of 37 and 41 questions out of the GMATPrep question pack. The questions are not adaptive, so that is not pushing you in the same way the test is—you won't learn now to make timing decisions in the right way. If you need more adaptive practice on quant, you can try GMAT Focus (another product from mba.com). This a 24-question adaptive set (they don't have the same product for verbal, unfortunately).
If you feel comfortable with your overall benchmark tracking, you can keep that method. If you think that that method isn't working for you, here's another method to try:
blog/2014/09/23/set-gmat-scratch-paper/
You might also look at the Question Pack add-on for GMATPrep (about 200 quant and 200 verbal questions). There's also GMAT Focus for more specific quant practice (no verbal). You can look these up in the mba.com store to read more details about what they are / how they work.