Marketing 101 Part 3: Advertising Part I
by Joseph Lieberman


Everything you wanted to know about advertising but were afraid to ask.

Just yesterday I was talking to the owner of an online strategy game and when I mentioned my services him his reply was "I don't need any advertising." For those of you who have read the previous articles you certainly know there is much more to marketing than advertising; however, advertising is still an important part of any marketing plan and tends to be the first thing people think of when they think of marketing.

Many indie developers tell me that advertising never pays for their products. It is true that if your game is converting far below 1% that finding advertising with positive ROI is pretty unlikely. However, as you approach a 1% conversion rate (even if you are somewhat below it), advertising becomes an increasingly important factor.

Part 1: The Math

Before you jump to the conclusion that advertising won't work for your game, you need to do a little math and find out some key figures. These figures are useful for many more things than advertising, but for now we will concentrate on their advertising implication. If you are planning on launching a game soon, make sure that you are collecting all the necessary parts to calculate all of these figures.

First you are going to need to know your conversion rate. Conversion rate is calculated as follows:

Number of sales / Number of downloads

You also need to know the likelihood of a visitor to download your product. This can be found by taking the number of clicks / number of downloads. A good site should have 20% or more. If you are below 20% then you may be having trouble with your site design or your target demographics.

Part 2: The Demographics

Next, hopefully, you can get some demographic information on your customers. Some useful demographic questions are: Age, Occupation, Gender, Income, Hobbies, and statistics on what their PC is.

Using demographics you can get a good grip on who your customer is. You can divide these up into probabilities by taking each sale and examining it's demographic. If you recall way back to middle school algebra, you probably wondered why the average "mode" would ever be used, since everything back then seemed to revolve around "mean" and occasionally "median." This is why! Figuring out who the most likely buyers are is a key to demographic mining. If ages 18-24 comprise 50% of your purchases and ages 50-65 the other 50% you don't want to know the MEDIAN buyer is 35! You want to know that the modes are 18-24 and 50-65.

Mode is defined as the most frequent value of a set of data

Also interesting could be sex: Are those buyers mostly male? Mostly female? Evenly distributed?

Occupation is probably not too important; however, if your game is about plants you may find that a lot of the people who buy your game have some kind of botanical background. This could be useful in determining where to advertise, but odds are most occupational demographics you will be able to make assumptions about or simply discount it as a factor. Same thing goes with hobbies; the odds are you should be able to assume whether your game is geared towards a specific hobby. However, demographic research may yield things that you hadn't thought of before.

Income is always a factor, but rarely in videogames do you see it as a determining factor. It is not often that people with incomes of $60,000 are more prone to buy a game than people with incomes of $30,000. If it IS a factor, you may want to look at income as a direct relation to education (therefore, directly related to occupation) as a clue to why that is. This means income analysis should usually come before occupational or hobby analysis and only if the data is very strange should you begin figuring out why. Advertising should never be done on income data alone as a factor.

Last but not least I mentioned PC configuration. Really, you should know this yourself. If your game requires a 800mhz processor you had better come to the conclusion that your buying demographic is 800mhz or above.

A parting note: Keep in mind that demographics can show the people who are willing buyers as well as the people who are NOT willing buyers. If you find that a group of people are the most likely to download your game it does NOT mean they are the most likely to make a purchase! Keep this in mind when polling your users for demographic information.

Part 3: The Lingo

There are four basic types of ads, and technically they can all be broken down into a single figure (CPM). However, until you know other factors this may not be so easy.

CPM stands per thousand impressions. An impression is each load or view of a website page. CPM is the most commonly used advertising type. If an ad is $1 CPM then it costs you $1 to display the ad 1,000 times. Most sites do this on a NON UNIQUE basis! This means if you buy 100% of the ad inventory for a day, if someone views 10 pages you get charged 10 times (if its $1 CPM, 10 page views is 1 cent). Many websites, however, have advanced options such as limiting the displays per session or displays per day. Be sure to talk to the ad representative about the various options.

There are some more important numbers to get to know and you should ask the person you are thinking of advertising with these things. First, you need to know how many unique visitors they get per day and how many banner impressions they show per day. This will give you the average number of page views per person (Displays / Unique Visitors). In order to have an effective ad, you should make sure the average person sees your advertisement multiple times, however, buying 100% is too much. Each marketing person will vary what they tell you and there is no rule on how much is the perfect amount. I suggest ensuring your ad is displayed 2-3 times per visit as a starting point and probably no more than 50% of the time, each website will vary based on its demographic and really on how well your banner and their site are designed.

CPC ads are uncommon, but you do find them in places. CPC stands for cost per click. CPC ads are nice in that you are guarantied traffic for your money. A lot of inexperienced people believe that CPC is the best way to advertise; however, due to the low risk involved, CPC rates tend to be higher than CPM rates when your banner is well designed and targeting the right group. Advertisers dislike CPC because if you design a crappy banner they end up having to display it that many more times, which costs them money! The ironic reality of CPC ads is that you WANT them to generate only qualified leads. If you can make your CPC ad unappealing to all people except the ones most likely to buy it, all the better for you! CPC is most commonly used in Seach Engine Advertising.

CPA ads are the least common. It stands for Cost per ACTION. Advertisers hate these because not only does someone have to click on the banner but they also have to perform some kind of action (download the product, register, ect). You are welcome to try to negotiate a CPA payment for an ad, but the cards will be stacked against you. However, if you CAN, a CPA action eliminates almost all of the risk on your side.

Finally, a fairly common method is a month (or general time) based cost system. There are several important things to know. First, how many other advertisers are there? Second, how many visitors and how many page views (again, so you can get average number of page views per person)? Make sure that in these scenarios your banner will be displayed more than once on average per visitor or demerit the site accordingly. Let's look at an example:

Site 1 has 100,000 visitors a day and 6 advertising slots that are going to be full (you would be #6). The average page views per person is 3, meaning the site generates 300,000 page views a day. The cost of the ad is the same as site 2.

Site 2 has 50,000 visitors a day, 3 advertising slots, and an average of 5 page views per person. We'll assume both have identical demographics. Site 2, even though it has less visitors, is the better buy because you will reach almost every visitor and reach most more than once, whereas site 1 you would completely miss half the visitors that day.

There are exceptions to this rule based on how frequently visitors return to sites and other factors. For instance, Comic based sites have a lot of people who visit multiple times a week. This means if you missed half the visitors that day it is not so important since they will be back tomorrow and then next day as well. The best thing to do may be ask the ad rep why you should pick them and see how they respond.

So I had mentioned these all go back to CPM models. What this means is CPM can be a common denominator for determining ad cost. Keep in mind, however, that this is a FALSE number. In order to accurately represent each ad you need to adjust for many more factors, such as the type of site, the demographic match, how many other competing ads there are, ect. In order to do this easy math you simply take a CPM ad and determine your average click through rate (How often someone clicks on the ad, also known as CTR). If your ad is getting .5% CTR (Which is somewhat low for a good ad) and costs 50 cents per 1,000 impressions you are paying .10 cents per click (CPC). If, from this site, it takes 3 visitors to download or sign up for a game, you are paying .30 CPA. So, basically, all these terms are interchangeable and each carries a specific risk factor for both parties involved.

Part 4: The Example

Now that we know all the background information we can put it all together and make some real decisions on advertising. The following will demonstrate the process step by step.

First, get your conversion rate and demographic information ready. We will say that your conversion rate in this example is 1%.

Next, find out the price, demographics, and web statistics (visitors, page views, ect.) of the website in question. We will say that this site is a great match on your best demographic and has 100,000 visitors, 300,000 page views, and charges .50 CPM.

To run an effective ad on this site, I would suggest buying about 120,000 page views. That is 60 dollars in ads. If you have designed a good banner (which will be discussed in the next article) and the site has a good ad layout (also discussed in the next article), you can expect a CTR of .5% or more. If your ad is below .5% you should be looking for causes. So, 120,000 page views at .5% is 600 visitors going to your site. If you have a well designed site and great targeting of your ad that may be 300 downloads of your product. Since you were converting at 1% that is three sales that you should get, right? Wrong! This site had a better demographic than your average trend as stated above. It is therefore logical to assume that you will convert at a higher than 1% rate! Lets assume it boosts you to 1.33% from good targeting. That means you got 4 sales from this 60 dollar ad. Lets assume you get $18 per sale net profit (a 20 dollar game with 10% processing/BW cost). You just turned your 60 dollar investment into $72. All of these numbers are fairly average. You can find better deals than these, I promise you. Even using very typical numbers we were able to turn a profit on a game that converts 1%. Plus this does NOT take into account word of mouth generated or additional publicity your ad may generate! If, on an ad, you spend 60 dollars and make 60 dollars in sales you have come out ahead. Why? Because there are now more people who are familiar with your brand, your product, and there are some new customers whom you can sell your NEXT product to.

Tune in next time for Advertising Part 2, where we will delve into the actual pros and cons of each advertising type, the best design methods, and much more.

About the Author

Joseph Lieberman is the founder and owner of, the only company dedicated to independent game marketing. To receive more helpful tips via e-mail, sign up for our newsletter at

Discuss this article in the forums

Date this article was posted to 12/9/2004
(Note that this date does not necessarily correspond to the date the article was written)

See Also:
Featured Articles
Running your own Game Company

© 1999-2011 All rights reserved. Terms of Use Privacy Policy
Comments? Questions? Feedback? Click here!