Partners Share Results as Facebook Exchange Exits Beta

By
  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

fbxFacebook today removes the "beta" label from its fledgling real-time-bidding ad exchange (official blog post), and has given its 15 platform partners permission to crow about their early results. And crow they have…

Here are two big takeaways from the wave of FBX data coming out this morning: (1) Facebook Exchange dramatically increases the amount of "brand safe" inventory available through RTB during 9-5 working hours, amounting to a substantial step forward in reach for programmatic media, and (2) Acquisition costs are substantially lower through FBX than through other programmatic channels, possibly as a result of the still-limited sources of demand being exposed to the inventory Facebook has made available.

For those keeping count, the current tally of 15 RTB partners is one more than we reported in our last count -- with search and social bidding platform Kenshoo recently joining the pack.

Facebook's partners are pushing out data points including campaign cost, user ad response, return on ad spend, and other metrics. Not surprisingly, the results are flattering to FBX, indicating solid ROI and conversion rates for clients sourcing campaigns on the platform during the alpha and beta trials. To get at the marketplace's potential weak points, we may have to read between the lines a bit.

So here goes, partner by partner…

Triggit: For two different clients (Shutterfly and SurveyMonkey) Triggit has accomplished 4x the return-on-ad-spend at one-fifth the cost per acquisition of their traditional (non-Facebook) RTB ad campaigns. Additionally, their conversion rates have been 2.2x higher than traditional RTB, and conversion lift has been 18 to 30% higher than a control group of users who did not see FBX ads.

Triggit also found Facebook Exchange brings huge amounts of retargeting inventory into the middle of the day, in contrast to traffic bursts for traditional RTB that tend to pile up during non-work hours (mostly at lunch and in the evening). Of course, the reason is that large numbers of workers are on Facebook throughout the day.

CEO Zach Coelius tells AdExchanger, "We went into this thinking, 'Hey, if this works as well as normal retargeting we're great.' It's better – amazingly better." The DSP, which has pivoted to focus on Facebook platform, has "a couple dozen" clients now live on FBX.

There's more in Triggit's blog post.

Xaxis: WPP's trading desk is the only holding company entity to be on Facebook's list of RTB partners, and has run campaigns on behalf of a handful of clients. Among them is an unidentified fast-food restaurant, which used FBX to drive new orders while lowering cost-per-order by 22 percent compared with similar campaigns that didn't run on the exchange.

CEO Brian Lesser is over the moon about the possibilities: "With the launch of FBX, Facebook not only instantly multiplies the overall size of the RTB market, it provides high-quality inventory that is brand safe and extremely targeted," he says.

AdRoll: AdRoll has run campaigns for 60 brands, including Room & Board, Hootsuite and GoPro. It says FBX campaigns have generally launched within one hour of receiving approved creative, and advertisers are seeing an average ROI of 16x.

Turn: Demand-side platform Turn has done an integration with FBX that can support billions of daily requests from Facebook. Since it went live with the exchange in June, it has enabled 50 campaigns and identified "a significant number" of unique users there that can't be found on other exchanges.

Turn's agency trading desk partners that have leveraged FBX include Omnicom's Accuen, Aegis's Amnet and IPG's Mediabrands Audience Platform. The latter of these is working with Chrysler Group to leverage the Facebook Exchange.

On results, Turn CEO Bill Demas said, "Advertisers have achieved their overall campaign goals very quickly, which underscores the quality of FBX inventory and its potential as an important part of the media mix."

TellApart: FBX ads for OnlineShoes.com earned clicks from 15 percent of users that were served them, says the retargeter. (That does not mean the click-through rate is 15%. The number of impressions per user was surely much higher than one.)  Across all customers, this "user CTR" metric is 6.6% on both the Facebook Exchange and Doubleclick Ad Exchange.

The Trade Desk: Acquisition costs are sublime, says CEO Jeff Green. One client is achieving conversions at half the CPA of other sources of inventory.

AppNexus: For indie trading desk platform Accordant, FBX delivered increased reach through AppNexus, while reducing advertisers’ CPAs by as much as 25 percent. Meanwhile Bizo delivered a 30 percent lift in reach "while maintaining key metrics such as impression to conversion rates."

Like Xaxis's Lesser, AppNexus SVP Andy Atherton sees revolutionary potential in Facebooks' sheer volume of "quality" ad space. "There is no doubt in my mind that the Facebook Exchange will open new doors for many of our clients," he writes in a blog post.

DataXu: CEO Mike Baker calls initial results "very positive, particularly around customer acquisition efforts."

Nanigans: One of only two FBX partners to already be certified with Facebook Ads API badge, Nanigans sees potential for granular creative. "A retailer could…deliver Facebook ads to people who have added products to their shopping carts with creative specific to those products."

Conclusions

A few observations based on what the partners have shared:

-Campaign costs are reasonable. Much of the new data has to do with very low cost-per-acquisition and impressive ROAS. In addition to the above, Bizo's Russell Glass wrote a piece for Digiday this week disclosing FBX delivered 300% "better" cost-per-leads, at prices 66 to 80% less than "an equivalent high-quality display impression."

-Retargeting is still the rule. Back during the launch, Facebook billed its exchange as a means to capture retargeting dollars. Very little in the early slate of results indicates it has been used otherwise, though there are indications that major brands (Chrysler, a major QSR chain) are sniffing around.

-Wither user response and price data? Notably absent is click-through rate and CPM data. We explicitly asked for this from one partner and were told clients had declined to share. We assume Facebook is not eager for user response rates and pricing information to be out there either.

-Cross-channel analytics. There's no data available as yet on the interrelation of Facebook ads with ads on other sites. Xaxis says, "An advertiser might identify an increase in performance when pairing a Facebook ad with a specific online video ad shown on a different site." Might, but will they?

-Facebook-to-Facebook comparison still missing. Comparing results data to native Facebook advertising is certainly possible, but no one in the early results have shared such results.

-Volume is still an open question. Facebook’s ad inventory comprises more than a quarter of the available US display ad impressions, according to comScore, and many of the beta partners are focused on the reach possible on Facebook. But it's not clear how much of that inventory will be allowed to flow into the exchange.

Footnote: Several FBX partners announced their involvement in the exchange this morning, but did not release specific results. These include Optimal, Criteo, Nanigans, and Rocket Fuel.

  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

Email This Post Email This Post

By on at

12 Responses to “Partners Share Results as Facebook Exchange Exits Beta”


  1. Alejandro Correa says:

    Question:

    Can anyone comment on whether these results de-duped against conversions recorded by the business-as-usual 3rd party adservers that I assume are delivering non-FBX ads? Are these results factoring in view-through conversions or just click-thru?

    thanks!

  2. xuehua says:

    Hope to see more absolute numbers instead of ratio. Also interested in seeing whether more 9~5 inventory on FBX is due to overall traffic paterns of Facebook website or the control mechanism of how Facebook inventory is moved to FBX.

  3. @Alejandro - I can comment for TellApart. Our ROI calcs are based on click-throughs. (We only get paid by our retailers for post-click conversions... check us out.)

    Re: de-duping, FBX uniquely finds 14% of users impressioned across all major exchanges.

  4. > Triggit also found Facebook Exchange brings huge amounts of retargeting inventory into the middle of the day

    This is inaccurate. RTB bid request volume (aka impression inventory) in the US exactly follows working/waking hours for both Facebook and other exchanges.

    Note, if you bid lowly on major exchanges, your ads serve at night because that's when competition is lowest. Because FBX's clearing CPMs are comparatively so cheap, those same low bids receive more impressions throughout the day.

  5. @xueha - The pattern of conversions/impressions during the post-work hours are typical for *traditional* retargeting exchanges. However, the availability of impressions on the Facebook exchange can't be denied - we've been seeing requests and conversions all day long because Facebook is quite simply "always on." @xueha - The pattern of conversions/impressions during the post-work hours are typical for *traditional* retargeting exchanges. However, the availability of impressions on the Facebook exchange can't be denied - we've been seeing requests and conversions all day long because Facebook is quite simply "always on." We'll be doing a deep dive on this very topic in the next few days - watch out for it on our blog.

    • xuehua says:

      Christina, thanks. I am looking forward to it. When you write new blogs about FBX, make sure to write clearly about what is your control group in your A/B test.

  6. Andy Hunn says:

    As a company that focuses on helping online advertising clients focused on BRAND advertising objectives, I hold out some hope for the utility of FBX to help with those goals. Every case discussed here is really focused on retargeting or bottom of funnel conversion. And I'm a bit concerned if the DSPs/Exchanges see this as the unlocking of a huge volume of PREMIUM inventory...

  7. Dylan says:

    Remember when Right Media first burst on the scene and the ROI/eCPA improvement was through the roof compared to ad networks and other high-impression sources? A significant factor in that was low CPMs due to the low initial RTB demand in the market. Over time, as more and more people began buying on an RTB basis, prices rose and return shrank.

    As FBX exits beta and more people bid for a limited amount of impressions, eCPM will rise and returns will decline. Not to say that FBX won't be a valuable tool in any marketer's arsenal (I'm sure it will be quite valuable), but I'm going to temper my expectations on this one. Numbers like "16x ROI" and "25% reduction in eCPA", if accurate, aren't sustainable. The market has ways of correcting these things. Competition will drive prices up and return down--I'm sure Facebook's counting on that.

  8. Joe Doran says:

    I haven't bought any FBX but have bought a ton of RTB and a ton of Facebook Ads through the marketplace. All I can say is that these results bode well for FBX but not for RTB. It really underscores the lack of quality inventory in RTB systems today (and most likely in the future) and not to mention the lack of scaled quality (converting) audiences.

    If the big two take aways are that - 1) FBX is Brand Safe & 2) Cheaper - people are just confirming what they knew going in. FBX is very Brand Safe compared to RTB - that is a known fact before the test. I mean there is stuff on the edges but Facebook is militant about maintaining their standards. Did you see the Nipplegate article this week where the venerable New Yorker was banned - http://www.newyorker.com/online/blogs/cartoonists/2012/09/nipplegate-why-the-new-yorker-cartoon-department-is-about-to-be-banned-from-facebook.html

    And 2) Cheaper - come on, every ad network or marketplace out of the gate claims performance lift. I think Zack sees through this where he notes that these marketplaces have limited demand. In an auction, lack of demand creates inefficiency, inefficiency means that the item sold is well below the market value. And anyone playing in the real Facebook market place knows the real swings in demand and pricing in the auction everyday as well as the constant pressure of rising costs.

    But the real issue is that the keepers of the RTB ecosystem (who ever they are) have created a beast that will be their undoing with FBX. They are going to let FBX see the pink underbelly of RTB and the weaknesses that are inherent with the system - lack of true transparency, low quality publishers, non-brand safe, low quality audiences, lack of scaled quality audiences, ect and Facebook will steal the demand away from them.

    Bringing those advertisers and the RTB buy side vendors into a Facebook controlled ecosystem where margin and control will be wrestled back by Facebook through the strong arm of a high volume, high frequency, high quality/ converting audience.

  9. Wilf says:

    All these claims about FBX providing better ROI and a cheaper source of inventory are completely irrelevant during a beta program. FBX has a huge volume of supply and they have restricted the demand. Essentially FBX has so much supply that they could (if they were so inclined) ramp up the supply to come up with any ROI story that they want during this beta.

    So right now FBX is cheap - oh great, a controlled price is said to be cheap during a beta. The true numbers will come out when they open up the demand (and of course they will to drive up the price to FB).

    Don't get me wrong, for people working in RTB, anyone that has a goal of opening up a large volume of supply is a good thing for the business. More RTB = more efficient buying = good results for advertisers.

    But obviously good ROI drives up the price and the inventory will simply become more expensive until the same CPA exists in FBX as everywhere else.

Leave a Reply