FYI: My native language is English, and I have studied Spanish, French, and am currently studying Japanese.
As already mentioned, no irregular verbs
No verb conjugation
Japanese has the concept of particles, which is brilliant, and can solve a great many problems that are present in most natural languages
The problems with Japanese (being a natural language, it has problems like any other) could be solved by more extensive use of particles. Verb conjugation particles could be added; counters could be replaced by a counter particle, etc. Verb particles could also let you put the verb anywhere in the sentence you want, making for a very flexible language. Particles for various levels of politeness could make that very easy for those societies where that is a thing.
I would evaluate the major languages of the world and see if there is a sizable enough set of sounds they have in common that would be sufficient for the new language. One of the problems learners of second languages have is their new language often has sounds that simply do not exist in their native language. If you don't start learning your new language before puberty, the chances of you being able to make native-level sounds in your target language (when those sounds don't exist in your native language) become very unlikely. Some people are able to, but most are not, especially if they don't have the opportunity to immerse themselves in that language every day, which will never happen in a new constructed language.
I'm learning Japanese, and the no spaces between words is VERY difficult to adapt to. I would recommend against no spaces. :)
Logographs like Chinese characters (which are also used in Japanese) are VERY efficient for a native language (and I can read those far faster than the Japanese words which do not use them), but learning them is a total pain in the ass, and takes far longer than I would think a good idea for a secondary, universal language. For this reason, I would use an alphabet system that is already in widespread use, and well-understood by more people than any other - the Latin character set used by English and the Romanesque/Romance languages. I would avoid the use of diacritical marks, if possible, due to being harder to type.