Google is not a search engine. Actually, that's a lie. The core of Google's success is a search engine. But that's not the business it's in, and neither are the myriad of other free products and services it offers.
Google's real business tends to get lost in the noise of a hyperbolic press, who sell papers by covering all of the company's latest, greatest and still-in-beta innovations. “Google launches email.” “Google photographs the world.” And, most popular: “Google takes on Microsoft!”
Despite all of these wonderful ideas, when it comes to making money, Google remains a one trick pony. The vast majority of its current income comes from a single idea: advertising.
Let's step back in time, and go on a history trip. The reason Google became such a good search engine is also the reason it's such a good place to put your advertising.
Most of the histories of Google, including the officially sanctioned version, present Larry Page and Sergey Brin - its thirty-something, multi-billionaire founders - as two geeky grad students who got lucky. Make no mistake: luck played a part, but so did outstanding engineering and the right connections.
Brin and Page met on the Computer Science PhD course at Stanford University, where they shared an interest in extracting useful information from large piles of jumbled up data. In a world when commercial search engines were finding web pages by the number of times searched-for keywords appeared, the future co-founders hit on an alternative approach: using links.
Brin and Page's initial search algorithm, ‘PageRank', worked by treating each link to a site as a ‘vote' for that site. Put simply: if someone had linked from their website to another's website, they must see some value in the other site's content.
PageRank counted the number of links to a site to work out its popularity. Some ‘votes' were worth more than others, however: a link from www.bbc.co.uk was (and is) given more significance than a link from the blog ‘musings of a stay-at-home Mom'.
The precursor to Google was a prototype that Brin and Page named ‘Back Rub' for its connotations of reciprocal ‘back links', where one website linked to another and vice-versa. It combined the popularity rankings that came from the PageRank system with more conventional methods of determining the content on a page.
The results were good and got better. Until Google went live (in Beta, natch), its rival search engines were very good at pulling up web pages related to a user's search, but pretty poor at presenting them in a useful order. This is important: finding the information you need online is not about having 300,000 options; rather it's about knowing which one to click on first.
Google did a better job than anyone else of ranking their results. The consequence? Pretty much no web searches these days require a visit to the second page of results; the answer is almost always on the first.
So, the big secret of Google's success was (and, again, is) top results that better matched its users' searches. Brin and Page focused relentlessly on improving, improving, improving this aspect of their machine.
The need for cash
Google did have another, less-obvious, source of success, though. As befits a project that was started under someone's desk, the search engine ran on a series of cheap, mis-matched computers strung together. Brin, Page and their steadily growing team developed technology that could use all of these machines as if they were one. It's the end result of this that is Google's other leg and the source of much of its competitive advantage. But that's another story.
Its relevance to this story is that the young search engine needed more computing horsepower to carry out an ever-increasing number of searches. Cheap computers gave Google good bang for its buck, but more bucks were needed.
As it happened, Page's brother, Carl, had himself started a company and sold it to Yahoo! for a few hundred million dollars. This doesn't get much reported, so perhaps it had little impact on the fledgling firm. Google was able to raise substantial sums for their own efforts. First $100,000, then $1,000,000, then $25,000,000.
With this money came responsibility, however; although Brin and Page retained control of their company and continued to focus on and build ever-better search technology, they needed to pay it back.
Both founders were fundamentally opposed to advertising, not for ideological reasons, but because it had a tendency to pollute the results. Other search engines, now Google's competitors, were selling banners and result rankings that distracted from and compromised their core purpose; helping people find the information they were looking for. Looked at in this way, these businesses were slowly strangling the golden goose.
A combination of factors overcame Brin and Page's resistance: in particular, the realisation that they could separate the screen and show paid results separately from pure, unsullied, ‘organic' unpaid results, and another company called Overture that was having success with simple, text-based adverts.
Overture operated an auction. Advertisers would bid for the right to show their adverts in search results for certain keywords. For example, BMW might bid to show an advert against the keywords ‘luxury car'.
The highest bid led to the most prominent advertising position, the second-highest to the second-most prominent, and so on. The more prominent the advert, the more people clicked on it and the more successful it was. Want more success? Pay more.
The format was staggeringly successful, for one simple reason: when people searched online, they were looking for something. Quite often they were looking for something to buy. An advert that popped up saying ‘Gillette' the moment a user typed in ‘razor' was much more effective than a random roadside billboard. Not only that, its effectiveness could be quantified.
After a slow start and a legal challenge by Overture, Google got in the game. And how.
How Google advertising works
An auction. But not really.
Google's pay per click program for advertisers is called AdWords (they have another pay per click program, AdSense, for publishers, but we're not going to talk about that here). AdWords is often referred to as an ‘auction', where the amount bid by advertisers determines where their ad shows up next to the search results. That's not quite true. But for now we'll pretend it is; it's a good place to start.
First things first: Google's advertising system is separate from its search engine. In the early days of search-based advertising, companies including Yahoo! used to manipulate their search results in favour of big advertisers. If you spent significant money you could expect your site to perform well in what's now called the ‘organic' or non-paid results.
Google thought differently, perhaps as a result of its founders being initially more obsessed with the perfect search results than income. Their view was simple: manipulating the system in favour of those with deep pockets would reduce the quality of search results. Reduced search results would lead to fewer users. And without users there would be no advertising.
It all makes sense. Now, when you get your Google results, the adverts are distinguished by a teeny label that says, euphemistically, ‘sponsored links'. If adverts appear at the top of screen, they also have a yellowish background. The rest of the links on a page are real search results, unsullied by filthy lucre.
Although some queries will return no results, most Google searches return several. These are found on the right-hand side of the screen and, often, across the top. Higher placements tend to attract more clicks and, as such, are more valuable. Here's where the ‘auction' starts.
Pay more, list higher. Or not…
In the beginning, Google knew that advertisers would pay more to be featured in the higher, more prominent positions. Running an auction was the most efficient way of finding out exactly how much more. Advertisers would bid against each keyword, specifying the maximum they'd be prepared to pay if a Google user clicked on their ad, thus visiting their website. The more money they bid, the higher they'd appear.
The trouble with this approach was that advertisers with deep pockets could appear at the top, even if no one was interested in their ad. This is bad for users – it's annoying – and bad for Google – if no-one clicks on the ad, they don't make any money.
The best ads are those that closely relate to the keywords they appear against. Remember from earlier that Google's success is built on providing the best results for people's search. The same applies for ads: the closer they match users' searches, the more likely those people are to click.
Google's solution to this was to record the click-through rate, or CTR. This is simply the percentage of people who click on an ad; or, to put it another way, the number of clicks on an ad for every hundred times it is displayed.
This changed the way they operated the auction system. An ad with a bid of 50p and a click-through rate of 10% brings in £5 for every hundred times it is displayed (100 x 10% x 50p). However, an ad with a lower bid of 40p but a higher click-through rate of 20% brings in £8 for every hundred times it is displayed (100 x 20% x 40p). Google made the click-through rate a key function in the bid process, essentially multiplying the maximum bid by it. This ensured that the highly placed ads would be the most profitable.
But, lest you forget, or in case you didn't know, Google is the land of metrics. The place where no decision is made without a pilot project, millions of unsuspecting participants and reams of statistics. The story where they tried to use user statistics to decide between several shades of blue is not apocryphal, it's true.
A company tends to get like this for a reason. In Google's case, it's a good one: such statistics have made them billions from AdWords. It turns out that putting popular ads in prominent positions is good for both profitability and people who are searching.
This isn't like any auction I've ever been to…
Google attempts to measure all of the factors affecting an ad's success. It measures the text, the click through rate, the keywords it's shown against, and even the landing page on your site.
It does this because it's bringing the same approach to advertising as it did to search: show the user what they want to see. Google is involved in a desperate, single-minded quest to match your search with the most appropriate ad. If it succeeds, you'll click and make them some money.
So, even relatively low bids can make it to the top, if the ad is of sufficient quality. Google, you see, has introduced an ad ‘quality score' that, in conjunction with your bid, determines the placement of your ad. Make it good enough, and you'll get great value for money. Stick anything up, and you'll be harshly penalised.
So, we've got an auction where you don't know how much other people are bidding, and where the auctioneer can adjust your bid however he pleases (and for unknown, unexplained reasons). But, strangely, it works.
In typical Google style, the quality score is a black box. No-one knows exactly what measures are used, what matters most or least and what counts as good. But we can have a good guess. A very good guess.
But there are many things, and this isn't a book, so we need a shortcut. That shortcut is relevance.
You need to make your ad and the page it takes you to (the ‘landing page') as relevant to Google users' searches as possible. While there are countless factors at play, this one simple approach will set you in the right direction.
Put yourself in a web searcher's shoes. With a little effort, this can be done; we're all web searchers. If you're targeting searches for ‘oranges', don't show them an advert for ‘fruit'; show them one for – you guessed it – ‘oranges'. And make sure your landing page talks about, yes, oranges. And, of course, selling oranges helps. Focus, focus, focus.
Set up the campaign with a modest budget and segment it as much as possible to target very small groups of potential buyers. Then test-improve-test-improve-test-improve. Google gives you all the tools you need to succeed and you'll soon be an AdWords winner. Good luck!