In part 1 I gave an overview of the experiment and what methods were used. Today I'll be explaining what happened. Just as refresher, here's the idea behind this challenge:
"If I had $19.95 to spend on advertising and I had to do everything myself, what would give me the greatest return on investment?"
Results are split into two categories, "paid" and "free". Where applicable the number of ad impressions and clicks are shown. There are also some exciting charts and analysis at the end too.
First up we have the paid advertising, the main focus of this experiment. The total number of hits generated was quite impressive for the amount spent. Sales and newsletter sign-ups were the best I've seen whilst the adverts were running, so something must have worked.
|Total Impressions / Clicks:||4823 / 66|
|Cost Per Click||$0.18|
I wasn't really sure what to expect with AdWords, but I was pleasantly be surprised by the results. The cost per click was a little more than I would have liked, but the clickthrough rate was quite high for a first time ad on some pretty competitive keywords.
First time advertisers must pay a fee to set-up their account, so in effect only $9.95 of the advertising budget actually went on displaying ads.
|Total Impressions / Clicks:||1037451 / 650|
|Cost Per Click||$0.03|
Project Wonderful use an auction based advertisement model, where the highest bidder has their ads displayed on their chosen slot.
To start with I placed a few manual bids to get a feel for how the site worked. Once I was more comfortable with it, I placed a campaign to last for the rest of the challenge. The campaign feature in Project Wonderful is a massive time saver and is quite comprehensive. Simply enter the criteria for sites you'd like to advertise on, and the system will go out and bid on appropriate slots.
One thing you'll notice is that you often go under your bid limit. The total amount spent on Project Wonderful was $18.59, so there was a little change left to run a few more slots. Another important thing to note is that the most expensive ads are not always the best. A $3 bid on a rather busy site (icanhascheezburger.com/) only brought in around 30 visits, whereas $0.57 on rpghost.com brought in over 100.
Text Link Ads
|Cost Per Click||$0.85|
I use Text Link Ads (affil) to generate revenue, so it seemed only right to try them out from an advertisers point of view. The ad cost $17, and with only 20 visits the cost per click wasn't very good. To make matters worse, 80% of visitors left without clicking anything else on the site. Oh dear.
Of course, Text Link Ads is (or rather was) more about boosting search engine rankings by getting links for your keywords. Google has recently punished people who place text link adverts on their sites (including myself), so I'm not sure how this model of advertising will continue. In this case it wasn't really worth the money, as any improvements in the site's ranking have been wiped out by the latest Page Rank update. Oh well.
|Cost Per Click||$0.07|
Everybody knows the WWW is a huge time waster. StumbleUpon is time wasting on steroids. By entering your interests, StumbleUpon will let you "stumble" onto sites that fit into your likes and dislikes. I use it myself and have found hundreds of new sites I never would have seen otherwise. It's really worth installing, especially if you have Firefox.
The idea behind StumbleAds is that you pay for "stumbles" from people interested in your site's theme. Each stumble costs $0.05, and you can set how many will be used each day, which is useful for spreading the server load.
The free advertising was much harder to measure, as it came from a lot of smaller sources which don't usually have their own reporting features. The range of traffic was quite varied, with a few surprises as well as a few disappointments.
LinkReferral.com is quite a simple idea. Your site is placed in a directory, and its position within this directory is dependant on how active within the community you are. By visiting other member sites you gain places, which improves your exposure. Writing (good) reviews and helping in the forums also improves your rank. Other members can also review your site so you get feedback.
Naturally it doesn't work quite like that. Plenty of reviews consist of "Great site link me kthxbye", and there are a fair few sites which are…below average. Having said that, for the amount of work it's a good way of getting feedback and a few hits. Some of the reviews are quite useful, and I've had a few website errors pointed out by members.
Traffic wise, the bounce rate (number of people leaving without interacting) is about the same as StumbleUpon's, which isn't as bad as I was expecting.
Link baiting is the process of crafting articles and content with the deliberate intention to attract a big response. For an idea of what these articles are like, check out the front pages of digg or reddit.
The article that worked for me was "Game Design Lessons: Halo", and a quick email to the editors at the excellent halo.bungie.org quickly got me over 800 visits. There were also a few referrers from other Halo fan sites as the link spread. Sadly Bungie didn't visit my house with cake and an Xbox 360, or even and Xbox 360 cake. Maybe next time.
The only other attempt at link bait was the Pirate Week series of posts, which didn't exactly set the Internet on fire. The Carnival of Video Game Bloggers also brought in a few visitors, but nothing compared to the Halo post.
You might have seen the square adverts if you've visited Flash arcade websites. Most of them animate far too quickly and include gratuitous flashes of cleavage to grab the viewer's attention.
I'll be honest. I ran the ads on this blog for a week and then had to remove them because I felt they were just too horrible. They were generating no traffic and made the place look untidy. I didn't actually realise that one of the services (moregamers.com) gave out 50,000 free credits after a week of posting, which is where all 190 visits came from. Never again.
I didn't exactly go overboard with this technique, but I was still surprised at how poorly it did. I went for a more indirect approach with links in signatures and profiles instead of in actual posts, so perhaps that explains the poor response rate. It's worth a try, but links need to be promoted a little more.
The Big Conclusion
Out of all the advertisements, creating link bait brought in the most visitors. However, most of them only stayed for the article and then left so sales weren't really affected. Project Wonderful brought in lots of traffic from a wide variety of sites, and AdWords had a decent clickthrough rate and brought good visitors.
I promised charts, didn't I?
And finally, two heat maps from before and after the ad campaign (click for bigger images).
The links to Cute Knight and Aveyond attracted a lot of clicks, especially from visitors referred by More Gamers. It must be the pretty pictures (and the fact that they rock).
So where does this leave things? As someone who'd never advertised online, it was easy to get started and quite nice to watch traffic appear with no more effort once the payment had been made.
One thing that's worth pointing out is that it's quite easy to blow a lot of money on advertising and see virtually no return. Different advertisers bring in different visitors, and spending money (or even time) on the wrong audience will quickly drain your resources.
From that perspective, Google AdWords brings in some of the most targeted visitors as they're already looking for what you're selling. Project Wonderful brings in a lot of traffic at a very low price, so it's a great place to get started.
StumbleUpon didn't perform nearly as well as I would have liked, but the content really wasn't targeted well for StumbleUpon audience. Individual, focused articles tend to do much better than site front pages. In the past I've seen articles bring in thousands of visitors within a few hours, so it's worth using to get some good exposure.
Text Link Ads were the biggest disappointment for me. The recent Google update hasn't helped matters, but I did see a slight improvement in my search engine rankings shortly after the ad went live. A few well placed ads might still work for other search engines, but it takes a while for the visitors to show up.
The free advertisers weren't as bad as expected. LinkReferral.com brings in a trickle of traffic, although the site users are looking for free advertising so may not be the biggest spenders. Some reviews are helpful though, and you may be able to build contacts if offering services and products to webmasters.
I still hate those flashing banners, even though they brought in nearly 200 hits. I really don't see how they're worth it for most sites as you have to generate around 100 impressions to get a single click. If you're getting enough impressions to get a worthwhile amount of visitors from them, you really don't need the help ;)
Forum posts and blog comments didn't really do much for the traffic, but they're easy to make and they may just catch someone's eye. A few recommendations between friends can do more for your business than any amount of paid advertising.
More business articles: Sodaware.net - Article index
Very interesting experiment - thanks for taking the time to explain your findings.
I should have stuck with AdWords a bit longer I feel… for some odd reason I decided to try Yahoo. So far not much is going on but it's only been a couple of weeks.
Interesting article, I have actually been curious about the thought of spending money advertising devbump. Up until this point I have only been advertising using the free methods you describe. My problem when it comes to getting visitors is that I need game developers specifically which rules out most mass advertisement methods. I imagine google adwords or being able to bid on a slot at a game development centric website would be the best solution for me.
@michael: Cheers, I've not used AdWords since this experiment but I'm tempted to go back. It was a little steep, but I think some tweaking here and there could cut costs and generate a better ROI.
@David: Have you tried advertising on GameDev.net? Not sure how much an ad on that large block would be, but it might be worth looking into. Personally I'd look for smaller game development communities, like fan game developers and blogs of people just starting out. Maybe even game based webcomics would attract a few developers and drum up some interest.
There's so many options out there that it boggles the mind, so it's worth trying a few small ads and then measuring the results. That way you don't waste too much money before finding out what doesn't work.
One way of gaining a sale's boost from link baiting might be to add a featured game or a few screen shots to the side bar of your blog. If you just read the Halo article you wouldn't know what the rest of the site was about.
Thanks for the suggestion Nate. I've added a little section in the sidebar to see if that encourages readers to visit the main site. We'll see how it goes :)
WebComic Planet - Project Wonderful for Beginners Part One
How much traffic does $19.95 get you? – Part 1 — Sodaware::Blog