I'd say, "We have a bad connection. We can only hear every other word you say."
In real life it's unlikely that you would really hear the first word, not the second, hear the 3rd, not the 4th, etc, in such a regular pattern. I might still say "we can only hear every other word" as an approximation of the real problem. No one's going to keep track carefully enough to say, "I heard the first, third, seventh, ninth, and twelfth words you said but not the second, fourth, fifth ..." etc. But we're more likely to say, "I'm sorry, this is a bad connection, you keep breaking up" or something more general like that.