Deciphering the AdWords discount

What’s Google up to with Checkout? And is the new payments system a sign of strength for the company, or of weakness?

Google’s an odd duck, businesswise, but the oddest thing about it is that it sells its product (AdWords ads) through an auction that it controls. Auction pricing is pretty unusual to begin with, but auction pricing through an auction you control is very, very unusual. One consequence of allowing customers to set the price of your product, through competitive bids, is that you can’t do what most companies do routinely: manipulate demand by altering prices. At least, you can’t do that directly. What you can do, though, is provide indirect discounts by reducing the price of other products that you bundle with your core product. That’s what Google’s doing with Checkout. It’s providing certain AdWords customers – those that also use Checkout – with a discount, or subsidy, on their AdWords purchases in the form of a rebate on their Checkout fees. For every dollar a customer spends on AdWords, it gets to sell $10 worth of goods through Checkout without incurring the usual payment-processing fee (20 cents plus 2% of the transaction amount).

So how big of a discount is it? Well, let’s say you sell widgets for $20 a pop, and you already use AdWords to promote your widgets. And let’s say it costs you on average $2 in AdWords fees to sell one widget. And let’s also say that you sell 100 widgets a month, earning revenues of $2,000 and incurring $200 in AdWords fees. So you’ve earned a Checkout credit of $2,000 ($200 x 10), which means that you get to process all $2,000 of your sales without any fee. That translates into a savings of $60 [(100 x .20) + ($2000 x .02)], which means that Google has given you a 30% discount on your AdWords purchases. Any way you look at it, that’s quite a large discount.

Now, this is just one hypothetical example, and I have no idea how it matches up with the typical profile of an AdWords customer. But I think it’s fair to say that the AdWords discount earned by Checkout users will often be substantial.

What does this indirect discount mean in the context of auction pricing, though? If all Google’s customers signed up for Checkout and received the same discount, it would be a wash. The prices of auctioned AdWords keywords would simply go up to offset the general discount. On net, Google would make the same amount of money, and its customers would spend the same amount of money. But not every AdWords customer will use Checkout. A lot of them, in fact, don’t use AdWords to sell merchandise or services over the web, so they have little need for Checkout. That means that keyword prices should go up a bit, but not enough to offset the discount Google’s paying. Nor will the company make up the discount in added sales. The customers receiving the discount will tend to buy more, since the prices won’t go up enough to fully eat up their discount, but those not receiving the discount will tend to buy less as the prices go up. (I encourage readers with deeper knowledge of mathematics and economics to check my reasoning here, as I may well be missing something.)

Now, certainly Google has some ulterior motives. By encouraging companies to use Checkout, it makes Checkout more attractive to consumers – and as consumers sign up, Google will be able to harvest more information about them and their buying habits. That will be worth something, as it will allow Google to target its ads more precisely. And if Checkout takes off, Google could eventually increase its fees or reduce the discount, thereby making money off the service itself. But these are theoretical benefits that may or may not materialize in the future. Right now, it’s hard to see Checkout as anything other than a discount program for AdWords. Google CEO Eric Schmidt admitted as much to the New York Times: “Mr. Schmidt said the company was willing to lose money on [Checkout] transaction fees because it felt the package would increase advertising spending. ‘The math works because we can have lower prices and higher volume,’ he said.” I don’t think “the math” works as he describes it, though. Yes, the net prices of keywords should go down for Checkpoint users, but the actual prices should go up (as those receiving the discount bid more), and I’m not sure that will translate into higher volume overall.

In this context, the program seems to signal weakness in the AdWords market rather than strength. Companies rarely give price discounts when demand is buoyant. Discounts are something you use when demand is softening. One thing Checkout will do is give an artificial boost to keywords prices, which may be helpful to Google when it comes time to report its results and discuss market trends with analysts. Any sign of erosion in keyword prices would make investors nervous.

There’s one last twist here. By effectively boosting keyword prices for AdWords, the launch of Checkout should increase demand for alternative keyword-based ad programs and boost their own prices. The biggest immediate beneficiaries of the Checkout discount may, in other words, be Yahoo and Microsoft. That’s another odd effect of auction pricing: one competitor’s attempt to provide a discount, through an indirect subsidy, may actually raise the prices of other competitors’ products. Weird.

7 thoughts on “Deciphering the AdWords discount

  1. quodlibetor

    Despite the fact that this seemed to me to be yet another step in google’s quest for total information domination of the world, the release of google checkout seems like it might the beginning of a strategy to eliminate click-fraud. Or at least another step towards converting from cost-per-click to cost-per-action ad pricing. Because, as somebody said: how profitable can any advertising system really be for advertisers when it relies on, not only impulse buys, but impulses that are so strong that you make the buy before you even close your browser window. They (and i really wish i could remember who i’m plagiarising here) suggested that you would need to be able to keep track of a customer’s buying habits for quite a while to determine the effectiveness of cpa advertising. And a checkout account, when combined with google’s 32 year cookie, seems like a pretty solid way to check out who buys what.

  2. Christian

    Interesting analysis. Although I see your point, I think the main reason the discount was offered is to allow Checkout to scale up exponentially and not necessary a reflection of weakness in Adsense. The side effects may play out to your scenario which is something perhaps Google did not fully thought out the consequences.

  3. _oh

    I got the very same impression from the eBay interview linked here earlier. That the real competition between the large dot.coms at the moment, is in trying to grow a market. Not necessarily making profits out of those markets, just in commanding the largest share and gaining a stranglehold, a strategic position or something.

    I compare it to the semi conductor industry, and the tactics adopted by Intel down through the years to become the very biggest player. Not many of us here, remember a time in the semi conductor industry, when there was discussion about diversity in the cpu industry. The notion of multiple providers of cpu chip technology. There was a huge consolidation at one point. The same thing is repeating now in web services I think.

    Brian O’ Hanlon.

  4. _oh

    “I think the main reason the discount was offered is to allow Checkout to scale up exponentially”

    Reed’s Law, exponential growth of group forming networks. I would assume, that wikis, are another manifestation of this, like auction formats are. Bringing people together to create connections, that last only for a while, and serve a certain purpose. The growth experienced is often much greater than that of services which approximate to Metcalfe’s Law, that scales to a power, or something like Moore’s Law, which scales only linearly.

    I guess, it is a necessary hazard of getting involved in any of these third generation networking formations – you need to grasp Reed’s Law, to fully grasp the whole concept.

    Brian O’ Hanlon.

  5. Nige

    Auction pricing is not so unusual. Further, it represents markets at their most pure and efficient.

    There are (at least) two, linked, reasons why all products/services are not sold via auction: 1. there are transaction costs involved that exceed the benefits; 2. bidders know that the “correct” market clearing price is a function of how many products the producer releases for sale.

    So, auctions really only work with discrete products, where supply constraints are clearly understood. Which is surely the case with Google: although bidders can seek any number of pages and locations to place an ad, each one is different and discrete. Further, there is no sensible reason for Google not to accept some price, no matter how low, for every spot. A producer would rather store a car or other hard asset than sell for too low a price, but some products, like Google’s ad positions, are not able to be stored for future sale (electricity is another example of this effect). In that sense, Google is being efficient and logical.

    Further, I would say all examples of spot markets (commodities, stocks/bonds) are actually broad examples of auctions at work: the only difference is that producers are also bidding the price(s) at which they will sell, in conjunction with, and influenced by, ongoing real-time bids from buyers.

  6. mohit_mahendra

    You argue that keyword prices will be up for all users, ad volumes up among Checkout users and ad volumes down among Checkout non-users. So if the advertiser/volume mix is more heavily weighted with Checkout users, overall volumes should be up. I think the math Mr. Schmidt talks about is right. I bet he knows his customer mix pretty well.

  7. mohit_mahendra

    You’re way too serious about the dismalness of economics in calling this ‘discounting’! Anyways, its really a discount if you reduce price for a widget without changing any other parameter. If its a better product for the same price, its technically not a discount. More a value driven decision you could say, from a product manager’s perspective ;) But think of Checkout as an enhancement of their advertising platform, rather than another online payment widget.

Comments are closed.