I know that half of my advertising dollars are wasted... I just don't know which half. -- John Wanamaker

Wow! I'm way ahead of one of the most famous names from modern department store management and modern advertising! I know that ALL of my advertising dollars are wasted!

Well, that may be a bit harsh, but (again) let's do the numbers.


Three little ad campaigns and how they grew

I started the Catena Manor web-comic ad campaign on January 24th, and it ran until part-way through the 26th.

Next, I ran a campaign on Housepets web-comic from January 30th through February 3rd.

At the same time, I ran one on two gaming support web sites. These both ran from January 30th through January 31st... you'll see why I stopped them early in a minute.

The raw, results-oriented numbers

Here are the numbers:

Date Lite downloads Paid sales Web site visits Video plays
01/24 18 2 197 702
01/25 11 4 57 255
01/26 6 0 24 100
01/27 5 2 13 55
01/28 7 0 9 58
01/29 5 0 12 0
01/30 10 2 13 54
01/31 4 2 25 101
02/01 14 0 50 163
02/02 8 1 45 167
02/03 5 1 10 34
02/04 2 0 4 22

Look at that middle column for a while... the one that (theoretically) makes me income... ouch! There's just no correlation there to any of the others.

Well, what lessons can I learn from this? Let's look at the individual campaigns and see what sticks out.

Catena

I mentioned in my previous post on this subject that I thought Catena would be a good place to advertise due to web-comic/gaming crossover there.

The ad numbers:

Date The Day's Expense Displays Unique Users Clicks Unique Clicks Cost Per 1000 views Cost Per Click Percent who Click
01/24 $8.21 90,900 85,400 403 388 0.09 0.02 0.44
01/25 $9.53 27,400 26,350 152 149 0.35 0.06 0.55
01/26 $2.26 5,450 4,900 24 24 0.41 0.09 0.44

Catena updated with a new comic on Monday the 24th. Not surprisingly, that was when the most number of people viewed the site and saw the ad. As I had suspected, there was a drop-off in the number of people coming on the following days. The surprise to me was that the cost per display and cost per click went way up.

Housepets

Chosen simply because it was (advertising-wise) inexpensive and seemed to be a largely inoffensive comic.

Date The Day's Expense Displays Unique Users Clicks Unique Clicks Cost Per 1000 views Cost Per Click Percent who Click
01/30 $0.03 2,000 1,850 3 3 0.01 0.01 0.15
01/31 $0.52 14,850 13,900 64 60 0.03 0.01 0.43
02/01 $3.62 35,700 33,450 146 136 0.10 0.02 0.41
02/02 $3.43 37,250 35,600 166 158 0.09 0.02 0.45
02/03 $0.40 2,800 2,650 13 13 0.14 0.03 0.46

A similar story in many ways. A note on Project Wonderful advertisement bidding: your ad only goes up when you're the high bidder at that moment; with Catena, I paid enough that I (believe) I had the site to myself for the entire day Monday and Tuesday. With Housepets, I often didn't have the whole site.

In fact, I wound up competing with Catena Manor for ad space on Housepets!

The gaming sites

In examining the site numbers and prices for various sites, I discovered that two gaming support sites looked really cheap, from an advertising perspective. The idea of these sites is for discussion (at least; there may be trade, etc. going on there, but I don't play the games so I don't know) of a specific video game.

My thinking was, "Gee, gamer sites, where people go who are people who play games... surely they'd be interested in a game!" Let's see how that thinking pans out.

First, nufonline:

Date The Day's Expense Displays Unique Users Clicks Unique Clicks Cost Per 1000 views Cost Per Click Percent who Click
01/30 $0.05 7,200 3,400 0 0 0.01 0 0
01/31 $0.47 74,000 33,800 0 0 0.01 0 0

Second, the unfortunately-named utopia-pimp ("pimp" as in decorate lavishly, not that other usage):

Date The Day's Expense Displays Unique Users Clicks Unique Clicks Cost Per 1000 views Cost Per Click Percent who Click
01/30 $0.04 58,150 49,750 0 0 0 0 0
01/31 $0.63 1,024,600 843,400 4 4 0 0.16 0

So: over a million page displays to over eight hundred thousand unique users! And... drumroll, please...

Four clicks. Not four percent, or even 0.04 percent, four total. As in, effectively, zero percent! nufonline nailed zero percent by the clever technique of providing exactly zero clicks.

And, to add insult to injury, while those were the least expensive ad campaigns, those were (by far) the most expensive clicks.

I killed those ad placements early.

Some kind of summary

So, it seems like these kinds of ads are not inappropriate for driving views of a web site. I can see why web-comics use these ads a lot -- their business model is getting people to read the comic, buy the merchandise, view the ads there, etc.

But for selling a video game, this doesn't seem to be a fantastic strategy.

It's also entirely possible that my video isn't fantastic at getting people to try the game, so while they came, and saw, it didn't conquer them.

The ad itself may not be fantastic... if I was going further with this, I'd design a few different advertisements and run A/B testing with them.