The use of the ONI is restrospective-- it is one simple index to categorize past historical ENSO episodes.  We do not use it for our real-time, month to month updates because it would lag what is happening.  For example, we would not wait until the end of 5 consecutive 3mo. overlapping seasons to say El Nino is here... otherwise we would be hopelessly behind the curve. 

In our monthly updates, we issue an El Nino Advisory when we see *1-month* SST and atmospheric indexes consistent with El Nino.  To issue an advisory, we also have to expect (>50% chance) that El Nino will continue.  Here is an article documenting several of the indexes we consider:

http://www.climate.gov/news-features/blogs/enso/why-are-there-so-many-e…

February 2015 was the first month where we felt El Nino conditions were in place.  The ONI might end up suggesting a start point that is earlier, but it is just one index based on one variable (SST only) and ENSO is a coupled ocean-atmosphere phenomenon.