As we now know, it was too much of a good thing.
A last-minute crush of orders KO’d many pay-TV operators’ ability to process them, forcing promoters to delay the start of the fight to give operators time to catch up, but even with that several operators were forced to offer refunds to thousands of subscribers who were never able to access the PPV stream.
HBO and Showtime, which jointly produced the broadcast, actually saw the problem coming in the days leading up to the fight and tried to warn operators. Advanced orders were running well ahead of any previous event, but if MayPac held true to form, the vast majority of orders — as much as 90 percent of the total — would still come in the final hours. Given the baseline, the networks feared — correctly as it turned out — that a huge wave was building.
The previous record for a pay-per-view event had been 2.48 million buys, for Mayweather’s 2007 bout with Oscar de la Hoya, so the Pacquiao fight reached nearly twice the previous high-water mark. And as with any issue involving network capacity, provisioning for peak demand is financially risky for network operators, especially when the potential peak is so grossly out of proportion to the baseline. Some rate of failure was all-but inevitable.
Yet HBO and Showtime had plenty going for them, too, in producing the fight, particularly in the U.S. The pay-per-view was available exclusively to pay-TV operators; there was not online streaming option available. The pay-TV ecosystem in the U.S. is, by now, well understood: There are fewer than 10 major operators nationally, all of which have many years’ experience selling pay-per-view events. Their device portfolios — set-top boxes — are limited and generally similar, and billing is asynchronous: PPV orders are simply added to a subscriber’s regular monthly bill, not invoiced and collected separately. Apart from the volume of orders there was nothing unusual about the Mayweather-Pacquiao PPV.
According to Gilles Domartini, founder and CEO of video e-commerce startup Cleeng, however, that was part of the problem.
“The truth is, nobody knows how to handle a million concurrent transactions, and there were certainly more than a million orders at the last minute in this case,” Domartini told me. “Uber can’t do it, Google can’t do, even Amazon can’t do it. We’re fairly confident up to about 500,000, but beyond that nobody really knows how to do it. Once you get up to that scale it just introduces too many variables.”
Cleeng, which bills itself as an EventBrite for video, handled order processing for the MayPac fight in the Caribbean region for PPV distributor Sportmax TV, where things went smoothly. While a far smaller market than the U.S. — roughly 15 million people vs. 300 million — the Caribbean presents its own challenges, including a fragmented pay-TV universe, multiple languages and currencies, and different payment preferences.
But the real challenge, Domartini said, will come if and when (probably when) big event PPV follows the rest of the pay-TV business off dedicated platforms and onto the internet.
“Once you try to migrate it online, you’ll be facing many more unknowns,” he said. “You won’t know all of the devices, or all of the players they’re using. You have to deal with different operating systems, different browsers. There are thousands of combinations just from a device perspective. But then you have different networks, different payment methods — there are many more ways things can go wrong.”
Netherlands-based Cleeng this week announced it had raised €1.1 million (U.S. $1.25 million) in series A financing that it plans to use in part to fund expansion in the U.S. Up to now, most of its U.S. business has come by referral as a preferred e-commerce partner by leading online video platforms, including Brightcove, Livestream, Ustream, Kaltura, and Ooyala.
“There are a number of industry players working on the problem [on handling video e-commerce at scale],” Domartini said. “The first one who comes up with a simple solution that is easy to implement and doesn’t cost a million dollars per event to set up will have a very interesting future.”