With OpenRTB 1.2, for the first time ever, DSPs can build support to objectively understand what they’re buying based on signals passed from the exchange, thus providing better targeting controls to advertisers. We discuss why this is such a huge deal.Read More
To make truly informed decisions, buyers need to know if the impressions they’re bidding on are header bidding impressions, as well as whether the impression is first-, second- or some hybrid auction dynamic. If the industry is going to move entirely towards a first price world, a few things need to start happening.Read More
The RTB auction did a good, but not perfect job of maximizing yield and was quickly embraced as the next generation. We discuss certain drawbacks of RTB - or potentially different ways of thinking about monetization.Read More
You may have seen the graphs showing the increasing percentage of media sold via programmatic channels. If you haven't, here:
TripleLift operates in the red region. That seems like the right region to operate, given a cursory review of that graph. The market has a good sense of what's going on in the grey area (fax machines), but it's the black area in the middle that merits additional discussion. By 2016, "Non-RTB programmatic," according to this somewhat-outdated and rather-optimistic prediction, will be nearly 1/3 of the entire market. What is Non-RTB programmatic and why does it matter? (Note: Automated Guaranteed, Non-RTB Programmatic, Programmatic Direct are all industry synonyms-ish and will be used interchangeably here)
RTB means that the impression is bought and sold in a real-time bidding auction. Programmatic means that the campaign is effected through some automated means. One could conclude, then, that Non-RTB Programmatic means a campaign effected through an automated means where the impressions are not conducted in an RTB auction. Accurate, but not insightful.
Through the lens of a standard publisher, there is the publisher adserver - DFP, OpenX, OAS, Adzerk, etc - which is used to traffic directly sourced campaigns (non-programmatic). These campaigns are generally sold on a guaranteed basis, meaning the publisher does its best job to actually deliver what it sells, and will often incur some sort of penalty for failing to deliver the right number of impressions in the right time period. The more a publisher slices and dices its inventory, the harder it is to guarantee any particular impression will happen in advance, at least within a particular time period - so forecasting is an important driver to maximize the number and price of directly sourced campaigns. The better a publisher can forecast, the more, and more targeted, directly-sourced campaigns it can sell - and the less it has to rely on remnant monetization. As a result, the top publisher adservers generally include some sort of forecasting tool as part of the product.
From the advertiser's perspective, deals and PMPs are good - they can buy impressions they want when they want, but they are not guaranteed. However, it is understood that a particular PMP may deliver no impressions at all. A publisher could sell out its directly sold impressions and never deliver a single PMP-eligible impression. For advertisers that prioritize actual delivery, this can be decidedly suboptimal.
If an advertiser wants to pre-empt the PMP marketplace, but not necessarily enter into the fax-machine world of non-programmatic campaigns, the various solutions that comprise non-RTB programmatic seek to deliver an automated solution. This space includes companies such as AdSlot, Rubicon (through its ShinyAds and iSocket acquisitions), Google (Programmatic Direct in DBM / DFP), AppNexus (through its now-shuttered Twixt project) and others.
The general interface presents a menu wherein publishers display their offerings - various targeting capabilities, media (e.g. pre-roll, interstitial, various size banners), and pricing for each. It's possible for the targeting to overlap (e.g. "first impression per user per day" and "all impressions for all males"). Advertisers select from the various publishers and offerings, negotiate on pricing through some interface, agree on the campaign, upload the assets through some tool, and then the software would automatically sync the assets to the publisher adserver with the correct settings in place. Naturally, this needs to be done in conjunction with the publisher's forecasting and knowledge of the unique specs of each publisher to ensure that guaranteed impressions may actually deliver (e.g. completely overlapping targeting is not allowed, and specs are enforced). It may then be possible for an advertiser to deliver a campaign on hundreds of publishers in only a few hours, without ever lifting a phone or putting pen to paper. Sounds great, but the market doesn't reflect that.
Rubicon Project was able to buy both ShinyAds and iSocket for less than $30M. Adslot has a market cap of $100M (http://www.google.com/finance?q=ASX:ADJ) and was founded a decade ago (most of its revenue comes from Symphony, a Prisma competitor), and there just isn't really much discussion of automated guaranteed in the market relative to other developments. While it may be great conceptually, to be effective this requires simultaneous buy-in from both advertisers and publishers - and automated guaranteed, at some level undermines the RFP process that agencies know and love. So it has not taken off the way many thought it would. That said, the benefits are relatively apparent - a better, easier workflow for those trying to buy guaranteed media. So it's very plausible that one day the market will grow.
At TripleLift, we don't have a strong position. If our buyers wanted to use automated guaranteed, we'd probably support it. This would mean building to the specs of the various solutions in market today. However, buyers don't currently request this with any great frequency, so we're investing in projects with a more direct impact on our business.