I think the stats on divorce rates are actually kind of misleading. The stats on divorce look at for example, 2013 and say, there were X marriages, and Y divorces. Then they figure the divorce rate. However, the marriages that ended in 2013 started in lots of different prior years. The marriages which started in 2013 haven't ended yet, so they shouldn't be part of the statistic yet.
To understand a true divorce rate, one would need to track all the marriages from a particular year, sat every 5 years for 3 decades, to see how many are divorced at each interval. An average of this number from many "starting" years would yield an actual rate of divorce, taking into account the reliability decreases with each increasing year of date of marriage.
The marriages and divorces in one year have nothing to do with each other, so this number has no bearing on the real situation.
Further, marriages can end after 5, 10, 15, 20, or even 30 or more years. So the indicator of a successful marriage is a moving target.
It's also a known fact (sorry don't have a citation but the image in the post above states the same thing) that 2nd and 3rd marriages have even higher failure rates.
I only know one person in recent memory who had a marriage that was "successful". He died of cancer while still married. Therefore, according to the definition of marriage (till death do us part) his marriage was properly successful. Ironically, it was his third marriage.
Everyone who is still married and alive has a possibility to get divorced, and being in it even for 20 years doesn't get them out of the equation.
Therefore, it's not inconceivable that the real rate of divorce for first marriages is actually much higher than 50%.