The big debates over real-time bidding tend to revolve around whether it is good or bad for publishers – but what if there are inbuilt flaws in the technology that could end up breaking it?
The growth and decline of high-frequency trading (HFT) – the name given to the computerised split-second trading that has had a huge impact on the stock markets – suggests RTB’s success could contribute to its own downfall.
So far, for advertisers at least, automated ad buying has produced positive results. It is claimed to have increased targeting accuracy, conversions, ROI and pretty much every metric an online ad buyer spends his days examining.
Yet in the early days after the introduction of HFT – the results were also hugely positive for those using it – but it didn’t last.
HFT peaked in 2009, when it accounted for 61 percent of all US stock trades. It has since fallen back to 51 percent – and profits from HFT have fallen to a quarter of their 2009 levels (see these two graphs from the New York Times).
HFT’s decline only began after it had become a huge part of the market. RTB emerged in 2009, the same year HFT peaked, but its growth curve isn’t too dissimilar.
According to IDC, RTB currently accounts for 16.5 percent of US display ad spending, but that will rise to 26.5 percent by 2015, when it will still be growing at around 3 percent a year. Growth is slowing, but that forecast suggests RTB could hit a similar proportion to HFT’s peak levels within 15 years.
What’s clear is RTB will make up a larger proportion of total spending over the next decade, but as it grows so does the possibility it will run into some of the problems that have afflicted HFT.
Many of the issues affecting HFT won’t translate into the ad buying market. There aren’t really any equivalents to the buy low, sell high mantra in trading, nor is there yet the same level of complexity produced by derivatives.
But there are two trends contributing to the decline of HFT that do have relevance for the ad market:
—Flash crashes: The interconnectedness of financial markets was rammed home by the 2008 financial crash, but HFT makes it even more likely small movements will lead to big earthquakes. With automated systems buying ads in the space of seconds in huge interconnected ad networks, there is a massive increase in the likelihood that a small variation in one part of the market – perhaps a major web property opening up a huge amount of inventory – can cause a ripple with far-reaching consequences.
—Red Queen Syndrome: One of the reasons cited for the decline in HFT is so-called Red Queen Syndrome, where each player is spending ever-increasing sums on incremental technology improvements in a bid to get an edge over the competition. High speed connections and the cost of data are two of the main outlays, and it is easy to see the cost of acquiring both becoming an increasing burden on firms specialising in RTB.
Ad buying and stock trading are very different. On a day-to-day basis many publishers engage with RTB through private exchanges – which are to a large degree sheltered from ripples spreading through the market.
Yet the technologies that changed the stock market and are changing the display ad market still share many basic characteristics.
For HFT – the combination of scale, interconnected markets and more players adopting the technology has begun to wipe out many of its advantages for the companies involved. All these are also issues for RTB. Learning the lessons from the decline of one could help ensure the future stability of the other.