Once upon a time we had graphologists. Since people don't know how to hold a pen any more, or even write anything worth a damn (not that it keeps them from blabbering anyway, unfortunately), it follows that we are now in the process of acquiring Graphology 2.0 - to analyse just how badly you text.
I predict - sadly without any merit - that the validity of the process will be even more hotly debated than version 1.0 used to be before falling into obscurity.
On the other hand, this "innovation" almost makes me regret Graphology 1.0, where you could listen to endless debates on the angle of the characters, the thickness of the lines, the depth of the impression left.
Makes it almost sound scientific now.

It's nonsensical crap which lawyers will use when it benefits them and judges can't throw out because there's nothing legally wrong with and which probably leads to a good number of people getting in jail for no reason, and the company becoming rich, and several bystanders getting so depressed about humanity and its idiocy that they jump off bridges.. 
So it's another great addition to society.

"Grant believes this type of analysis will be far more effective in comparison to current methods, as he says he can distinguish the sex of the sender as well as how well the recipients know each other."

What if you're a homosexual, or a very aggressive woman who likes to pretend she has a penis on irc? And I'm certain that there are plenty of people who behave differently, depending on what they think the people they're talking to know, or think, about them.

This is going to be more trouble than it's worth, I think.

Unless they also test it against everyone else in the world, the best they can "prove" is that whoever sent the message wrote in a style more similar to the suspect than to the victim. And it probably won't be hard for an attorney to find another 10 or 20 people who are an even better "match" for the messages.