How does a website make money? The basics of online advertising and the business models that use it

One of the most important things an analyst can do to hone their valuation skills is to understand the business models used by the companies they’re evaluating.  Is it a business that depends on high turnover and low margins, like Wal*Mart or Target?  Is it a business that depends on high margins and lower turnover, like some of the luxury retailers or homebuilders?  Or, is it a spread business, where leverage, the pricing of risk, and cost controls are some of the most important drivers of profits (banks and insurers fall into this category)?

Careful examination of these factors, as well as what drives them, is a tremendously important part of company evaluation.  A solid understanding of the business model – before the question of valuation is even considered – is an absolute necessity for intelligent research.

With that in mind, it’s worth taking a look at a simple, fictional dot-com company to try and better understand how it is they make money.  I wrote a post about this previously as well, so if you read that, you’ll recognize some material.

The basics of online advertising (ad-based sites)

Much like television, many websites make their money from advertising.  With television, the companies spend their money to produce ‘content’ (shows, news) that are broadcast out to a certain area.  This content is interspliced with advertisements, and the advertisers will pay money to the TV station (or network, if it’s part of a network) for playing their ads.  So we have a few driving factors at play (approximately in order from most to least important):

Revenues:

• How many people in our broadcasting area actually watch the network during the time slots the ad is aired?
• How many of these people are in the target demographic (perhaps homeowners who are between the ages of 35-50?)
• How many other TV stations/networks in the area compete with us for that target demographic during that time slot
• How effective is other media (newspapers, radio, billboards, the internet) in reaching the demographic, and are they willing/able to do it for cheaper?

Costs:

• How high are our fixed costs for broadcasting (power, licensing fees, salaries of staff)?
• How much did our content cost us to produce (did we make it in-house, buy it elsewhere, etc)?

There are undoubtedly other factors, but these are probably the big ones.  With ad-based websites, there are many parallels here.  The business model is that the site will produce content (usually articles, but there are other examples – see ‘other business models’, below), and advertisers will pay the site to display ads alongside that content.  For simplicity sake, let’s say I run a blog with a small writing staff.  Here are the driving factors, roughly from greatest to least importance.  I’ve put some industry terms in bold:

Revenues:

• How many unique visitors (different people per day or month) does my website get?  Can also be referred to as ‘eyeballs’ if you want to sound cool.
• What’s the growth in my userbase (growth rate of unique users)?
• What’s their click-through rate (how often are they clicking on ads on my page)?
• What are the payout rates and formats on the ads I show (explained below)?
• What’s their average visit time (length of time spent) on my site?
• Any alternate delivery modes for ads (do I have a mailing list, for example)?

Costs:

• How high are my fixed costs (salaries of programmers, writers, utilities, and of course – for any good dot-com - maintenance on the office halfpipe)?
• How high are my variable costs (mainly server/hosting fees to keep my website online)?

Demographic Targeting and tracking ad viewership

With TV stations, it is very difficult to try and calculate the exact number and type of people who actually see a given ad.  The best that most stations can do is figure out their geographic reach, and then use Nielsen data to estimate the popularity of a given show during a given timeslot and region to figure out the number of households watching.  We can overlay this number of watchers with viewership demographic information, and figure out, for example, approximately how many 18-35 year old males are probably going to see a given ad.  This will largely determine the rate that we charge our advertisers, who will be interested in maximizing their ad dollars by targeting only likely buyers of their product.

The ability to target specific demographics – as well as the ability to track how many viewers have been reached - is where the internet starts to pull away from TV advertising.  By using tracking information for my website, I can tell advertisers exactly how many people read my blog on a given day, nearly exactly where they’re accessing the website from, and exactly how many of them click on the ad or follow up the ad with a purchase.  Furthermore, unlike TV, I don’t have to show all viewers the same ad; I can serve up different ads for people accessing the ad from different countries, states, or even zip codes (singles from available now!).  Depending on the tracking cookies I’ve got floating around and my ability to get user info, I may even be able to show different ads for users that I suspect are male or female, or in certain age brackets.

These factors mean that instead of guesstimating ad viewership and demographic info, I deliver it to my advertisers on a silver platter.  The internet is the first form of media where this is possible, and it is a huge leap forward for advertisers.  It means that instead of basing my ad rates on readership (newspapers), listener base (radio), or viewership (TV), my advertisers can pay me based on how effective my advertisements are in reaching the target demographic.  This has led to the development of new payment agreements for many dot-coms.

Payout rates and formats

Ads on my blog are displayed in a number of places.  I might have them embedded in text – for example, if I review a product, a good idea might be to link the name of the product to a purchasing page on Amazon.  However, more ads these days are image-based ads, referred to as display or banner ads.  Sites generally have banner ads near the top, bottom, or sides of a page.  These text and banner ads generate revenue in one of three ways.  From lowest-payout to highest:

Cost-per-mille, or CPM: The advertiser pays a set rate for every 1,000 individual page loads the webpage gets.  This is the closest to old-media – we’re being paid for simply running the ad.
Cost-per-click, or CPC: The advertiser pays the website whenever a user clicks on the ad.
Cost-per-action, or CPA: The advertiser pays when a user takes a given action (submits a valid e-mail, say, or makes a purchase).

As a rule, the more stringent your criteria for payment, the higher your payment will be. Hence, my generic CPM ad might only pay me $2 per 1,000 views, whereas my CPC ad pays me $2.50 per click, and my CPA ad might run as high as $15 or more, depending on the action.

These payout methods can also be combined, based on an advertisers goals.  For example, AdCo might pay me $2 for every 1,000 people who view an ad, $0.50 for every person who clicks the ad, and $4 for every person who signs up for their mailing list. 

At the end of the day, my websites revenue is determined by:

User base x click-thru rate (what %age of people click ads) x average payout rate per ad.

Factors that will increase these variables will increase my revenue, and thus, what my website is worth.

Sounds good, but how the heck do you pick the ads?

Content, content everywhere, but not an ad in site (pun intended).  So what’s a webmaster to do?  Most websites will quickly realize that they can’t regularly keep up with selecting, maintaining, and rotating a million ads on their pages – nor are they pros at selecting the best ones - so they will farm out ad selection to another company.  The companies that provide this business are said to be in the business of Ad Serving.  Simply put, ad serving means that another company takes a look at your content and then takes care of the task of placing ads on your page for you.  Google’s AdSense is a dominant player here, but there are some others as well.  These Ad Servers typically take a cut of your advertising profits as payment for selecting the ads you run.  Read my original post to see how Google got so good at this, as well as a mini-rant on why I think Facebook just isn’t impressive enough to warrant a $56B (now $75B or so?) valuation.

Variations on a theme:  Other Online business models

So hopefully you now understand the basics of text and display ads, as well as some of the factors that drive a website’s profitability.  Below are a few interesting businesses that use online ads a little differently than my simplified model above. 

Note that a common theme with making more money in online businesses is the idea of providing ‘infrastructure’ for a website, but finding a way to push the task of content generation out to other people, or, in some cases, the userbase.  This means that a big, big part of the websites costs (paying writers, researchers or journalists for content) is no longer a factor, and as a result, our website is almost infinitely more scalable. 

Youtube.com, eHow.com – Online video site and how-to site (respectively) where all content is user-generated.   A portion of ad revenue is returned to content-submitting users as payment for generating content.

Buy.com, Overstock.com – Websites who had the idea of using online ads to supplement loss-making (or at least, deeply discounted) e-tailing businesses.  Since inception, results appear to be pretty weak and both of these businesses are starting to wise up to the fact that the e-tailing business will probably have to stand on their own two feet.  (I encourage you to take a look at the last ten years of OSTK’s financials before you jump into this business.  It doesn’t help that their CEO is kind of insane and easily distracted from his core duties).

Google.com, Yahoo.com, Bing.com – Websites that generate zero content, but instead help their userbase better sort through the web by providing search engine services for free.  Revenue is generated by serving ads up on the pages of results.  Note how well this business model dovetails with the business of serving ads up for 3rdparty websites; this is why you’ll almost always find these companies involved in both search engine marketing as well as ad serving for 3rdparty websites.  Many of the core competencies and software developed in one business can easily be applied to the other.

Facebook.com, Myspace.com, LinkedIn.com – Websites that generate no content but offer online networking capabilities.  (LinkedIn is a little bit of a special case; revenue is generated both from the member base and from advertisers.  The “advertisers” in this case are companies looking for qualified employees, so payout formats also differ a bit).  For networking sites, ads served are sometimes done by an in-house department – many of these companies are in pursuit of the holy grail of social networking, which is the ability to serve up better, more targeted ads based on all the data they have about their userbase. 

A huge amount of the hefty price tags affixed to companies like LinkedIn.com and Facebook.com is the fact that they have all this user-specific, user-provided data and thus should be able to serve up even more targeted ads – i.e. not only slicing an ad demographic down to the zip code, but also down to people who read certain books, watch certain TV shows, have blonde hair or enjoy mountain biking, etc.  While this sounds great in theory, if you read my first post you’ll know that I remain unconvinced.  Some of these sites may in fact someday generate a lot of money based on this concept, but the valuations applied to these sites at present virtually guarantee that most investors at present would be better off putting their money in an index fund.

There are plenty of other business models out there, and new ones spring up all the time.  One of the beautiful things about the internet is that the costs of setting up a site are so much lower than establishing a brick-and-mortar space to try out a business idea.  I’ve also deliberately excluded companies like Groupon or the now-defunct Pets.com from my “other business models” discussion, because companies that hemorrhage money at record-setting rates are not ‘businesses’; they’re incinerators.  While there are indeed loss-making businesses that can turn out to be stellar companies, these extremely early stage ventures are usually best left to venture capitalists or angel investors. If someone explain to you that their business will absolutely be profitable once they get over their never-ending ‘customer acquisition’ costs, or once their userbase reaches a “critical mass”, you are strongly advised to keep both hands on your wallet. Remember that a tremendous amount of money was lost in the dot-com bubble by people who mistakenly thought that growth in a userbase forgave a multitude of sins.  While growth is an important part of valuation, it will not fix an inherently bad business model or a business that is losing hundreds of millions of dollars per year.