[NelsonMinar]

I was also at Google when the AOL deal was signed and remember the decision and outcome quite differently. It was definitely a high risk deal, deliberately so, but in no way was it only "the best-case scenario had us breaking even". Nor was it the case that "all of our models were wrong". The product manager in charge of ads at the time had a clear understanding of exactly how the deal could be hugely profitable for Google because of the value of extra advertisers attracted to the Google platform thanks to the added AOL inventory. It was by no means a sure thing, but it was a likely outcome. Fortunately the decision makers believed him, they took the risk, and it paid off enormously.

[nostrademons]

That's what I've heard from other early Googlers who were there at the time. It was a calculated risk. If the deal went badly, it would bankrupt the company, but the best estimates from existing revenue numbers and likely ad inventory increases had it being hugely profitable for Google. Their internal numbers supported that but few people outside the company believed it, hence the need for a guaranteed revenue clause. It certainly wasn't a shot-in-the-dark, though.

[mathattack]

Execs rewriting corporate history after the fact. Go figure.

[ido]

Seeing as this happened 20 years ago and she was quoting from memory, I think we can give her the benefit of the doubt. It could have also been that her view at the time was different than OP's/the person who made the deal & she recalls her opinion as "how it was".

Such things in-fact happened to me (even tho I'm a bit younger than Mayer & my career started a few years later than hers), human memory is really unreliable & a lot of times you remember your feelings & translate those to facts upon recollection even when that wasn't quite the case.

[joesb]

As if only evil CEO can remember things differently.

"This CEO rewrites history! My friend's memory is more trustworthy!"

[eledumb]

And unless the people who were at Google at the time were in the meetings, they are merely speculating.

I was at Netscape and was actually in meetings that marca, and others were not in, and it's funny to hear what they say happened, when they weren't in the room, nor were they consulted. Even the people at the top can be guilty of speculation.

[mathattack]

Speculation, and creating a corporate mythology. It’s deliberate.

[shostack]

Can you speak at all to what sort of modeling work is done around this sort of major partnership? I'm really curious how many people are involved in that sort of exercise at that level and how clean the data is, what sorts of data are used, etc.

[exelius]

I do this kind of work! I don’t know at Google, but typically there will be some amount of third-party due diligence for deals over a certain size / risk profile. Some companies have an internal strategic deals group; which I would assume Google likely does (but probably didn’t back then).

If both companies’ data needs to be combined an analyzed, they usually bring in an outside deals consulting firm. Those teams tend to be very small due to the sensitive nature of the discussions involved — usually 2 or 3 people (backed by a large shared support staff and tooling) over the course of a few weeks.

Often the data used is a combination of proprietary data from both companies, commercially sourced data or proprietary data platforms built by the consulting companies. Deals are a big, sensitive, relationship-driven business.

[johnrgrace]

  It's often one guy with a spreadsheet, and 1-2 people to review the spreadsheet and a few business people to validate assumptions.  The modeling is easy, getting the assumptions right is hard.                                
  To be honest, if you're skilled very simple models that you can do in your head or in a few minutes usually give you a perfectly good answer.  The more complex exhaustive models are usually there to make sure you didn't overlook something or 20 small inputs all cross multiplied to throw your answer off.                             
    Getting the answer exactly right also doesn't matter, if you make $90m in profit off a deal or $100m your going to do it.  What you are most concerned about is making sure that you don't lose money and what factors would push you to do that.

[kayhi]

What does that type of deal look like?

AOL had ad inventory and Google had to get enough eye balls?

[toast0]

Inventory is slots to put ads. AOL had the eyeballs, Google had the platform to turn eyeballs into money, better than AOL was doing by itself. Google provided a guaranteed rate (cpm? cpc?) so that AOL would have confidence it wasn't bait and switch.

See also the Yahoo/Bing deal which didn't work out as well. Microsoft didn't end up actually hitting the targets, and convinced Yahoo to take less; and Yahoo also didn't reduce employee count anywhere near plan on searchy/advertising stuff, so they missed targets on revenue and cost and user experience.

[nostrademons]

"guaranteed rate (cpm? cpc?)"

It was guaranteed revenue IIUC, something like "If you do not make at least $150M from this deal, Google will pay you the difference." That makes the deal a no-brainer for AOL, but puts Google on the hook for any shortfall, which could have potentially ended up bankrupting the company.

[toast0]

I couldn't find a great description of the deal terms. I did find mention that at least some of the guarantee needed to be paid up front. I suspect the guaranteed payment was contingent on some level of traffic though -- if AOL had turned off the placement, or lost a huge amount of users, then they probably couldn't keep the money.

[C1sc0cat]

Cost per mille or cost for 1000 clicks -its an old-school advertising term

[reallydude]

CPM = Cost per mill (cost per 1000 views)

CPC = Cost per Click (cost per 1 click)

These terms are widely used today.