Since @geoffc is satisfied with, and handsomely rewarded for unsourced answers from time to time, I'll finish the job by quoting the sources I provided there in comments 1, 2 since comments are considered temporary.
Spaceflight Now's Falcon 9 rocket to give SES 9 telecom satellite an extra boost:
But Martin Halliwell, chief technology officer of SES, sang SpaceX’s praises in a preflight press briefing at Cape Canaveral on Tuesday.
“I know we’re a little bit late, but we can live with that,” he said. “We’ll move on. What’s most important is that we get the perfect mission.”
SES is the world’s largest operator of geostationary communications satellites, and the Luxembourg-based company approached SpaceX for some relief after the launch delay.
“We sat with SpaceX and said, ‘Guys, how can you improve our mission profile? How can you get us to orbit a little bit quicker?’ We have agreed with SpaceX that we will change the mission from a guidance controlled shutdown of the second stage to what we call a minimum residual shutdown of the upper stage,” Halliwell said.
In plain speak, the adjustment is a change in the Falcon 9 rocket’s control logic.
Instead of programming a target orbit into the rocket’s guidance system, the Falcon 9’s second stage will burn its single Merlin engine until the launcher’s supply of kerosene and liquid oxygen propellants are nearly gone.
“We’re going to burn the motor on the second stage for a few more seconds,” Halliwell said. “That’s all it really means.”
[...] “Initially, it was going to take about 93 days to get to orbit, but those few more seconds of burn will bring us down to about 45 days, so we should be in operational service towards the end of May or right at the very beginning of June,” Halliwell said. “That is a big deal for us.”
This reddit thread contains the following. Unfortunately their links to sources are now broken:
Apparently SES-8 was delivered to a supersynchronous orbit as well and I found a reason for that here
SES-8 is being inserted into a Super-Synchronous orbit to reduce the total velocity change required by the satellite to reach Geostationary Orbit. A conventional insertion into GEO from an elliptical transfer orbit would require a delta-v of 1,800m/s while the SSTO design reduces the total dV requirement to 1,500m/s.
edit: Just wanted to clarify that SSTO stands for Super-Synchronous Transfer Orbit in this case, not Single Stage to Orbit like it often does.
and also
The supersynchronous orbit saved SES-8 300 m/s of ∆v, or about 6 extra years of station-keeping in GEO.