Since it seems almost trite to say this, I'm not certain why it doesn't appear to have been said before (judging from a recent, fruitless Google query for this specific phrase): "Keywords are not stocks." But listening to many firms in the SEM industry, one would think that bidding on keywords on Google is governed by the same rules as buying shares of Google, Inc. itself. One glaring problem with this, of course, is that search engine marketers don't "buy and sell" keywords. If you bid on the keyword 'running shoes' on Bing, even if your bid is sufficient to get first position, you haven't "bought" the keyword: that is, you don't own 'running shoes', Bing does. Unlike a security, if the price-per-click goes up, you can't "sell" that keyword on the market for a profit. And if it goes down, you can't deduct that loss from your taxable income either. That's no big deal, just put a minus sign in the portfolio theory equation where there used to be a positive sign and move on, right? Well, it's not as simple as that. For stocks, each company typically issues millions of shares, each of which has exactly the same price and pays exactly the same dividend, if any. Your only decision is how many you want to (and can) purchase. With keywords however, there are only 10 ad slots on the first page, at most, each one of which has a different price, gives a different click-through rate (and therefore, ROI or CPA) and, if you're lucky, you perhaps can get one of those slots. Imagine if every share of Microsoft had a different price and paid a different dividend! But the purpose of portfolio theory is not to help you determine which price you should be willing to pay for any stock. Its purpose is to determine which percentage of your assets should be in stocks and which percentage in bonds. With keywords, however, there is no equivalent of bonds. There is no keyword equivalent of a "risk-free asset", no word that is guaranteed by the U.S. federal government to yield a return on your money. There are subtler differences between stocks and keywords as well. For example, portfolio managers tend to invest in a couple of hundred stocks, at most. The best invest in only a handful. Ken Fisher, the multi-billionaire investor and author of the long-running 'Portfolio Strategy' column in Forbes magazine, recommends for the typical individual to choose 20 to 30 stocks at most. But the average search engine marketer deals with orders of magnitude more keywords - thousands, if not tens or hundreds of thousands, and most by necessity, not choice. That is, unlike search engine marketers, stock portfolio managers don't have to worry much about the 'long tail'. In fact, they intentionally purchase stocks so that each one individually accounts for a non-negligible (but not excessive) portion of their total holdings. They diversify like this for the simple reason that no individual investor (or firm) has any significant impact on the performance of any individual stock or of the stock market as a whole. They are at the mercy of the market and therefore try to do everything they can to insulate themselves from unnecessary risk. But in search engine marketing, the single biggest factor that impacts the performance of any account is the decisions made by that account's manager. The only real similarity between a portfolio of stocks and a portfolio of keywords is the use of the word 'portfolio'. The point is simple -- keywords are not stocks, and we shouldn't talk about them like they are.