Facebook, Ad Servers, and $344B in Media

There is $344B in media* market cap that own and operate ad serving systems now.  

Google acquired DoubleClick ($274B mkt cap), AOL acquired AdTech ($2.86B mkt cap), and Facebook acquired Atlas ($65.4B mkt cap).  ValueClick owns MOJO and retained Mediaplex ad-server ($2.05B mkt cap).

When we think about ad servers for Madison Avenue, our guts tell us DART and Atlas**.   Both of these two ad serving solutions are now owned by larger-than-life media platforms.  MediaMind, the challenger of ad serving solutions is making inroads across Madison Avenue and believe it or not, has surpassed Atlas as the number two platform.***

Having heard the speculation turned news recently about Facebook acquiring Atlas and reading Gokul’s post on AdExchanger, I still do not understand why they did this acquisition unless Facebook thinks they can convince Madison Avenue to use them as their 3rd party ad serving tool of record.

My question to Madison Avenue:  Wouldn’t you want an impartial 3rd party to be your ad serving tool?  Why would you rely on a media property who is going to make more money off media than ad serving to deliver you your attribution models?

And with this, I’m not saying Google is any better.  It’s a big reason why the majority of our clients are not on the DART ad server.

In the finance world, there is significant rules around proprietary trading (prop desk) and analyst/research work.  The two basically do not intermingle and in the recent laws, the two might have to split.   This is FINRA rule 5280.

(a) No member shall establish, increase, decrease or liquidate an inventory position in a security or a derivative of such security based on non-public advance knowledge of the content or timing of a research report in that security.
(b) A member must establish, maintain and enforce policies and procedures reasonably designed to restrict or limit the information flow between research department personnel, or other persons with knowledge of the content or timing of a research report, and trading department personnel, so as to prevent trading department personnel from utilizing non-public advance knowledge of the issuance or content of a research report for the benefit of the member or any other person.

I understand that advertising is not finance, but wouldn’t we take clues from a more robust industry?

If you are a marketer or agency and put all your media plan data in a company who is selling you millions of dollars of advertising media, don’t you think that the data will be used against you?

Here is an example, purely from illustrious purposes:
Property A – $6/cpm $3/cpa
Property B – $8/cpm $3.50/cpa
Property C – $7.75/cpm $3.40/cpa

Imagine the three properties above have their data in an ad server controlled by Google, AOL, ValueClick, and now, Facebook.

When you go to purchase media from any of these four properties, they can see what you are currently paying and what the actual performance is.

This gives these media platforms a significant leg up on pricing & performance as they know where they need maintain or beat.

Is it just me that’s skeptical?

On a completely other note, I do not run M&A for Facebook but I would have suspected they would have built their own Ad Server and maybe acquired an attribution company such as Adometry, C3 or VisualIQ (or the many others in that space).

* Companies who own significant media properties.  Google, AOL, Facebook, ValueClick.
** There used to be a trade magazine that showed ads served each month by ad server, but I haven’t seen it in a while.  Purely based by my conversations with other agency heads, Atlas and DART are the primary ad servers that come up in conversation.  MediaMind is coming up more and more.
*** Updated after an email conversation with MediaMind.

  • http://twitter.com/robleathern Rob Leathern

    Thought-provoking as always, Darren, enjoyed the post. I think a real story might be in order here to help explain why I think it eventually won’t really matter for most advertisers. (BTW heard an ad for Adometry on NPR yesterday which was a trip)

    We had a customer when we were XA.net before our switched focus on social, that had used our platform to purchase Google AdX inventory. Our display platform allowed you to create a single campaign and publish it via API to multiple ad exchanges/locations including Adwords, Google AdX, Right Media, and other RTB exchanges. This was an always-on kind of campaign, serving through their own Doubleclick tags within AdX and the campaign went on for a few months at fairly low spend (about $1,000 total). Anyway, as these things sometimes do, campaign ends, and company eventually responds to inquiry on a couple of unpaid invoices with the “we logged into the system and the numbers didn’t match our adserver numbers” response.

    We dig in further. Turns out that the person who started the campaign no longer worked at the company etc. etc. Long story short, we know adserver discrepancies sometimes happen and when we know about them we work with the advertiser to fix them. This didn’t happen in this case and our terms of use state that the advertiser pays based on the numbers from the Inventory Partner (such as Google or Yahoo! or whomever). Our fee is 100% transparent markup on top of the actual numbers from the partner.

    We furnish Google data for those campaigns, screenshots etc. Company still only wants to pay $600 of the $1000 bill. Now this normally wouldn’t get to my level, but it was an interesting situation. I’d never ever been to small claims court and was curious if the (no capital letter L) legal system could work this stuff out, so we took these guys to small claims court over the now very very late $1,000 bill.

    The contract was clear – payment was to be based on the number coming from Google.

    The judge/magistrate/whatever says “well, how do you know that Google’s numbers aren’t wrong?”. I almost fell over. I come from a long line of lawyers (for better or usually for worse) and my obvious response was twofold – 1) Google has over 1 million advertisers. I said to him “well if they are, I’m sure there’s millions of advertisers spending billions of dollars who’d love to know that”, and it’s extremely unlikely they’re systematically wrong (there was no possible clickfraud issue here since this was CPM-based) but more importantly (2) IT DOESN”T MATTER. The contract said payment was based on this number coming from a third party. We have to pay Google the actual cost to us before markup. If you say you think you got fewer impressions, I can’t pay them less very easily. If you bring this issue to me early on I can contact them (usually not for a $1000 campaign but you get the idea) and our analyst can figure out what’s going on… but that’s why our lawyers write the contract that way.

    There is no objective truth in the media business. It’s a trust-based business, and it’s a market. Technology can help you find bad actors but at the end of the day we need to work with entities we trust, and the amount of trust we’re going to have to put in the likes of Google and Facebook from a media buying perspective, will only increase. We should anticipate this and our contracts will have to reflect this.

    BTW the small claims thing left me a bit cold. There was no dramatic delivery of verdict from the bench like on TV! Two weeks later, in the mail we learned we had “won” (our $70 court costs were reimbursed) but that our “award” was $750 and not the full $1000. This was less than the media cost. But the experience was very valuable for me and also reinforced what I’d learned working with smaller advertisers using credit cards; good luck media-geek getting anyone outside our industry to understand how this nonsense REALLY works and make sure you get what you pay for!!

  • Domenico Tassone

    Great post Darren. No, you are not the only skeptic and I’m glad I’m not the only person that notices. There is definitely potential for conflicted interests in the technology stack in the way you talk about it and also as I’ve posted about on tipofthespearblog.com. In addition to the aggregate performance and spending data there is also the sharing of actual user behavioral data which is of course – extremely valuable; even more so for ecommerce. Not everyone has armies of engineers and dataminers to make use of this data. Still, it just doesn;t make sense to so blithely share this intellectual property in what are often de facto co-ops. Digital marketers should methodically remove vendors from their stack where the risk of co-mingling data and the vendor gaining an unfair advantage are likely. The reality is that too many digital marketers and their agencies just follow the crowd.

    That said, the corollary to this is another insidious kind of conflict but on the actual digital analytics reporting services – usually bundled with multiple agency services. The problem is that agenicies often (but not always) have a direct interest in the results. Facts are routinely spun and manipulated to fit the story of the day. As an independent digital analytics practitioner we see this all the time and as recently as last week, a digital client was left hanging with neither their media agency (and certainly nor the media vendor) speaking up about measurement problem about which they both knew (we did). And that is just the tip of the iceberg. For the same reason why you don’t ask the same doctor for a second opinion, digital advertisers ought look more closely at the objectiveness of their analytics. If they don’t care about biased results then they shouldn’t be surprised when their agencies don’t either.

    As the old saw goes, clients get the advertising they deserve.

  • hubert

    Why don’t you put Appnexus / Microsoft in the list?