May 27th, 2008

Google AdWords: A Brief History Of Online Advertising Innovation


All innovation looks inevitable, except while it’s happening.

Google’s search advertising model didn’t spring forth fully formed. It was iterated, and many of the key concepts were borrowed — something many people don’t realize. But a few key market-defying decisions, and one stunning insight, made it all work. Here is a brief history to inspire, taken from John Battelle’s The Search (required reading for anyone who wants to innovate anything on the web):

In late 1999, Google began testing a program to sell ads on a CMP basis, the dominant ad model of the time.

But instead of using banner ads, the dominant ad format of the time, Google decided to sell only unobtrusive text ads. And they decided to target those ads based on search terms, and to keep the ads separate from the main search results.

Advertising first appeared on in January 2000 — text ads were sold by a sales rep on a CPM basis. (Yes, that’s right, there was no pay-per-click, no self-serve, no bidding.)

“It didn’t generate much money.” – Sergey Brin

In what would turn out to be a massive irony, based on its initial lack of success with advertising, Google had planned to give its inventory over to DoubleClick, the largest banner ad business of the time.

But then the bubble burst in Spring 2000, and the online ad banner market crashed.

In the wake of the bust, Google introduced a self-serve model for buying text ads — they got the idea from, although they did not then adopt GoTo’s pay-per-click model.

In October 2000, Google introduced AdWords, with this announcement on the main page, “Have a credit card and 5 minutes? Get your ad on Google today.”

This first version of AdWords still sold ads on a CPM basis — the program was successful.

In 2001, Google’s ad revenue was on pace to hit $85 million, but was outpaced by Overture (the renamed GoTo), which earned $288 million in ad revenue selling pay-per-click ads on an auction basis.

(Nobody remembers Overture because it was never a destination — it powered PPC advertising for other sites, and was later acquired by Yahoo, to become Yahoo Search Marketing — yes, more irony.)

In February 2002, Google introduced a new version of AdWords, two years after launching its first ad program and nearly a year and a half after first launching AdWords.

The new version of AdWords adopted Overture’s pay-per-click auction model, where advertisers bid on how much they will pay per click.

If Google had copied Overture entirely, the history of the web might be very different… but they didn’t.

Instead, Google introduced a breathtaking innovation.

Overture’s pay-per-click auction model allowed advertisers to buy their way to the top of the listings — highest bid gets the most exposure.

Google realized there was a problem with this approach. If an advertiser bid their way to the top of the ranking with an irrelevant ad, and no one clicked on it, then nobody made any money from the advertising.

Fortunately for Google, they understood relevance.

Google introduced clickthrough rate, as a measure of the ad’s relevance, into the ranking algorithm. So if an ad with a lower bid per click got clicked more often, it would rank higher.

The result — a lower bid ad with more clicks generated more revenue than a higher bid ad with fewer clicks.

As Battelle put it:

Google’s decision to factor clickthrough into an advertiser’s ranking forced an economy of relevance and profit into the pay-per-click model.

That “economy” turned Google into the great money-making machine that it is today.

Google had two moments of pure brilliance. The first was PageRank. The second was introducing relevance into the pay-per-click auction model.

So brilliantly obvious — yet nobody else at the time thought of it.

What’s notable is that Google didn’t invent search or auction-based pay-per-click advertising — their innovation was perfecting it.

The challenge of innovation is that we are all boxed in by what we know, by our assumptions about how things work.

Nobody at the time thought there was anything wrong with Overture’s model — it was making lots of money.

Nobody at the time thought search was a business — it was expensive and resource intensive, so most portals like Netscape, AOL, and Yahoo outsourced it — ultimately, to Google.

The next Google-like innovation is right in front of us — we just need to see past our own assumptions.

Forget what you know.

Comments (17 Responses so far)

  1. Google gave a concept of beta ( iteration) to the world — hats of to them!

  2. Time for GOOGs to reap the benifits of its innovation — well done guys!

  3. The phrase “Google-like” rather then Google per se should be emphasized. One should also keep in mind that innovation need not coincide with profit. Real innovation nearly always comes from the margins: i.e., right out of left field, as was the case (at the time) with Google. Think web logs, wikis and so-called Web 2.0 which is really about USG. The latter, USG, as a model that seeks to profit from extracting surplus value from content created by third parties will likely fail in terms of an money-making models because there is little, if any, profit to be made from USG. If, generally speaking, people will not pay for content, regardless of quality, which is demonstrably the case, where is the contemplated next “Google-like” innovation, if by “Google-like” we mean “tranformative” or, perhaps, “disruptive”?

    I wish to argue here for Interchangeable Master Channels (IMCs). Overview of what they are and how they work can be found at
    The reason I mention IMCs is because, among other things, a “tethered” rather than “untethered” (after Zittrain) Web removes many of the obvious problems with the laissez faire Internet, including spam, scams, phishing and otherwise trashy content. In other words private Internets that coexist within the public Internet in the form of level three “brand channels” where brand name entities in the form of can sponsor in product or service specific master channels content deemed by a brand channel sponsor to be worth sponsoring. The phrase “worth sponsoring” means content gets aggregated thematically, in content relevant master channels which are “neutral” zones, in the sense that an otherwise copyright item posted to a dot-com master channel is deemed by the author to be in the public domain: ie., royalty free. Brand channel sponsors can thereby retrieve and aggregate the related content in their own brand channels. There are presently 750 dot-com master channels in the IMC network and its growing as more people learn about the innovation.

  4. Really great fantastic article!
    What next then?
    Will it be with the Semantic Web?
    Searching for “holiday in Rome” and getting ads packed from an intelligent agent.

    Are you searching from the UK (?), are you logged in google (?), do you have any cookies on your pc (?).
    The more of it, the better the “agent pack” will be!

    The results is that it won’t look like an ad, but a complete full service including: place to stay, place to eat, travel/tour around Rome, museums and monuments to visit, cars to rent, tickets to buy, shopping places, and so on… (including the most relevant events during your staying).

    A perfect agent who provides a list of services relevant to you… and you will only need to check the box: yes or not.

    Personalized/customized relevant useful (and probably pleasant) “hidden” ad service!

    And it can always be improved!!! ;)

  5. [...] Read the full article in – Google AdWords: A Brief History Of Online Advertising Innovation [...]

  6. [...] Batelles Buch “The Search” noch einmal zur eindringlichen Lektüre und arbeitet in diesem Beitrag einen zentralen Meilenstein in der Entwicklung von Googles Werbeprogramm AdWords heraus: 2002 hat [...]

  7. Great post. What it illuminates is that genuine innovation – that aha moment – is really what creates great value (with persistence of course). What is fascinating is that one innovation alone and Google might be nowhere, but it was the combo of the two just at the right time…

  8. [...] this excellent article makes clear, Google did not invent the keyword auction model, but it did perfect it. By perfecting [...]

  9. [...] | Publishing 2.0 trackback ¿Recomendarías este post? Más noticias sobre: Empresas, Sectores, Marketing [...]

  10. Very interesting read, but Google hasn’t done anything new in quite a while.

  11. I was actually working (and selling) pay-per-click search engine results at a meta search engine called years before Google. I even remember them struggling at the beginning and people saying that they could never catch up to

    It was (and still is) an exciting time.

    I find it even more fascinating that in the multimedia world of the Web, the big win in online advertising is text… speaks volumes to the written word and it’s happening even more so now with stuff like twitter.

  12. [...] Google AdWords: A Brief History Of Online Advertising Innovation [...]

  13. [...] muy interesante la nota que escribe Publishing 2.0 acerca de Google AdWords y como se convirtió en el producto que reconvirtió el mercado del [...]

  14. [...] next media business to connect those dots will be the next [...]

  15. [...] Scott Carp explains how Google adwords was adopted from outside then adapted [...]

  16. Thank you for such a great post and for reminding all of us that there was actually a company before Google that offered pay per click advertising. Today, everyone is so caught up with Google that they forget about Overture. The reality is that PPC is nothing new. However, it has spawned a whole entire industry that never existed before.

  17. Online advertising has come so far since I attended my first website promotion workshop back in 2001. And now that video advertising is available more options are at our disposal.

Add Your Comment


Receive new posts by email