Scoop: We found the deck DoubleClick used to sell itself to Google

This article is being published in cooperation with Insider’s Lara O’Reilly.

In our first newsletter we published our history of Google’s DoubleClick acquisition and made the case that within that deal were the seeds of the company’s eventual antitrust struggle. By piping the enormous demand from Google Ads into the near-monopoly publisher software platform DFP (now GAM), Google was able to build a reinforcing loop of advantages that helped turn the fledgling DoubleClick Ad Exchange into the majority marketshare behemoth informally known as AdX.

As part of that story we mentioned that the economic model used for selling DoubleClick, known as YMAG, was lost to history. This seemed important to me at the time since it would help align the story of what happened at Google back to the intentions it had when making the acquisition.

After I published the original article, a reader who will remain confidential sent me a copy of the entire YMAG presentation used to sell the company. I can attest that this person is someone who would reasonably have had access to the document, and I showed it to another former executive of the company who confirmed it appears authentic. Also, the first five or so slides of the deck are in the DOJ’s case as plaintiff’s exhibit 104, but for some reason they do not include any of the actually interesting details.

So, for the first time, here’s the YMAG deck.

Let’s just say this right away — this deck isn’t going to win any design awards. We’re in DoubleClick’s early aughts design era. Lots of burgundy and black and the old fashioned Click-Click logo. It wouldn’t be until 2008-9 that we would see the nice yellow and green color scheme emerge.

The complete deck is 49 slides. I’m only publishing the ones that are of interest to the antitrust issues at hand.

What is DoubleClick and what has changed?

The deck includes quite a bit of exposition about the company and its products, which is omitted here.

Starting with the company overview. Tough to argue you’re not a monopoly with “9 out of 10 agencies” and “9 out of 10 sites” and “35 out of Top 50 web publishers”.

Below, the “marketecture”, if you’ll allow me, shows how the company used to position its products on the buy- and sell-sides, with each burgundy square representing a distinct offering.

Editor’s note: At the time this slide was created, I was personally in charge of DART Motif and Video. At a later point I took over MediaVisor and DFA.

In the previous newsletter I made the assertion that the post-private management team had done a bang-up job on cleaning up the company and positioning it for growth. The deck agrees.

Let’s talk DoubleClick Marketplace (Ad Exchange)

At the time of this deck’s creation the ad exchange was nascent, and still called “DoubleClick Marketplace”. Elsewhere in the deck the projected 2007 revenue for the exchange is listed as just $800,000.

Below is the not-at-all-confusing diagram of what it is (or was at the time). Keep in mind, real-time bidding (RTB) did not yet exist, so the assumption was that buyer demand was in the same system as supply via an ad tag.

Even in this early stage, DoubleClick touted “Dynamic Allocation” as the killer feature of their exchange. In the slide below you see this feature highlighted as a unique value proposition in the upper-right quadrant.

Interestingly, they also list “boomerang” as a buy-side advantage. Boomerang was DoubleClick’s retargeting solution, which was a largely moribund business. The ability for buyers to use their first-party data turns out to be the basis of much of the multi-billion-dollar programmatic ecosystem and the key innovation of multi-billion dollar companies like Criteo. I am pretty confident that DoubleClick didn’t really know what they had cooking here.

The previous newsletter included a kind of crappy diagram I whipped up showing how Dynamic Allocation benefits publishers. Behold! We now have the official diagram explaining this! The diagram shows that if you overlay marketplace demand (upper-right) on top of direct demand (upper left), with Dynamic Allocation you end up always choosing the best yielding demand (bottom).

Editor’s Note: I am 99% sure this beautiful diagram is the work of advertising savant Scott Spencer, so lets give credit where credit is due.

Synergies are where its at

In the previous newsletter I made the point that the sell-side synergies were what sold the company. The deck agrees. See the bullet under sell-side reading “Most strategic and time-sensitive opportunity.”

Let’s get into the sell-side opportunity. First the slide, then we’ll translate the text:

• “the Big Four” : Yahoo, Microsoft, AOL, Google. It’s a little hard to wrap your head around the fact that Google was in tough competitive battles against these companies given their growth and dominance over the subsequent decade.

• “off network inventory”: Ads on webpages not owned by the companies, i.e. "everything is an ad network”

• “largest pool of liquidity”: When running a marketplace, having more demand and supply is the winning formula. I don’t really know why “liquidity” as a term has stuck to this market, I find it confusing.

•  “Network effects…ebay”: We are the biggest source of supply.

• “Become the primary ad server to get a ‘first look’”: Yeah, that’s the good stuff. This is before anyone ever dreamed of header bidding, so being able to snipe inventory that no one else has access to is a huge advantage.

• “The race to the finish line…”: Should we interpret this as a realization that there would be a winner-take-all dynamic and that the buyer of DoubleClick would win? In retrospect it kind of looks like that. Could AOL, MSFT, or Yahoo have done anything to win after Google bought DoubleClick?

Let’s model this out

After explaining the strategic rationale, the deck then goes into a model to explain the value of the deal to a buyer. It appears that the version of the deck that I have obtained was meant for Yahoo, and presumably very similar decks were created for the other companies.

First, the deck defines terms. The only interesting thing here is the definitions of “Available Inventory” which is clearly stating that the buyer (Yahoo) would be able to grab inventory from the DFP client base.

The next slide shows how the calculation works.

Here’s the main course. The YMAG synergy model. Let’s walk through it:

• The columns may be confusing.

  • “DFP” is obvious.

  • “DE” is DART Enterprise, which was a customer-hosted ad server used by many large publishers that has since been retired.

  • “Affiliate” is through a division of the company called Performics which, as the smallest input, was ironically the only part of the company that got spun out as part of the deal terms with the DOJ.

• Impressions:

  • 19 billion impressions per month. That’s a lot.

  • Take out 9 billion as direct sold, low quality, or part of MySpace. I think MySpace is called out here separately because they are so big and have very low CPM ads.

  • Net-net, you are acquiring the ability to monetize about 10 billion impressions per month.

• Monetization

  • There are estimates for “Publisher Network Lift”. This is the amount they are saying the buyer could sell the inventory for, above the baseline CPMs the publishers are already getting.

The calculations are a little hard to understand so I’ve spelled them out:

DFP Tier 1: 1.8 billion imps x $6 CPM x 150% lift = $15 million/day

DFP Tier 2: 3.3 billion imps x $0.50 CPM x 300% lift = $4.8 million/day

The end result is $2.6 billion per year in incremental revenue! Not surprisingly, the company sold for $3.1 billion. A bargain in retrospect.

The buy-side is an afterthought

The last two slides of the deck made me chuckle a bit. At the time of the acquisition I was running product management for the buy-side suite (DFA, etc), and it was palpably clear that Google didn’t really know what to do with them or why they even owned them in the first place.

The deck reflects this reality. The buy-side slides talk about using “Boomerang” to essentially retarget users on the exchange. This is presented as a pretty speculative opportunity as retargeting wasn’t really possible before programmatic.

The punch line here is that DV360 is now a $10+ billion programmatic powerhouse, arguable more strategic that AdX, and this was barely contemplated at the time.

Predicting the future

The YMAG document is of historical interest and lends context to the current antitrust case. It largely predicts the effect of bringing Google Ads demand into the exchange, along with the power of the ad serving stickiness to build a large business. In the end, the amounts earned by Google dwarfed the eye-watering $2.6 billion prediction in the deck, as programmatic exploded, DFP share continued to grow, and monetization opportunities increased.