I ran an ad on reddit recently on the /games subreddit. it’s the place for fairly serious gaming debate, and I thought it would thus be a good match for my very serious political strategy game Democracy 3. It was a ‘sponsored link’ right at the top of the page. First things first, don’t let all that information on reddits ad page about ‘how ads work’ fool you. That’s an old comic they drew years ago and is completely wrong. they just sell impressions at $0.75 CPM. they really should update that…(I think this is a recent change).

Anyway….normally $0.75 CPM is a bit high for indie game advertising, but it was targeted, and a good experiment in a potentially rich new source of players. I put $250 into it, which lasted about 3 days.

According to the traffic stats provided by reddit I got:

stats1Which says 1,140 clicks. On the face of it, this isn’t too bad at all, 22 cents for a quality-traffic click. However, if I check google analytics for that page and its source I get this

stats2Only 199 views? that’s insane where are all the other views? Let me check my webalizer server logs instead… I see about 2,200 hits from reddit/r/Games for the month (which is almost certainly all this ad, as I have been listed only in /indiegaming until this. If I look at different logs from awstats I get 1,000 visits for the month from /r/Games.

So what value do I believe?

Well in a sense, who cares? All I care about is how many of those people bought the game, but so far I am suspicious about the tracking I have in place there, although it seems to work very well for other games so…arggghhhh!. Can I conclusively say the ad was worth it? I suspect not. Something I can definitely say from all this is that google analytics is CRAP at actually tracking the number of hits you get from a site. It could be php acting up, maybe people have javascript turned off or ad-block software turned on that blocks google analytics? All of my online research has given me long lists of reasons a few clicks might be missed, but not 90% of them. If we believe awstats & reddit 9and I’d be surprised if any site selling ads had the cheek to lie about them, they WOULD get caught by someone, and then the site is toast overnight), then it means that all of my google analytcis data has to be taken with a huge helping of salt.

Would I advertise on reddit again? Maybe, but I have other methods I’m experimenting with for now. (BTW don’t bother with google video ads, total waste of time and money).

10 Responses to “Analyzing advert effectiveness. (Where did all my clicks go?)”

  1. Joel Lamotte says:

    I might be wrong, but doesn’t Google Analytics ignore very low time spent on a page? I mean when you open a page, quickly realize it’s not what you want and close it, only spent less than 5 seconds on it?
    I don’t know the details of Google Analytics but I always assumed they were sorting a lot based on what they think is meaningful. If I’m correct, it might suggests that maybe there is a way to obtain “raw” data.

  2. Martin says:

    You appended a custom source parameter for that particular campaign to the URL and I assume you track that. Wouldn’t that be the most dependably measure?

  3. LK says:

    I bought the game after being directed here from Reddit, Cliffski. :)

  4. cliffski says:

    Cool :D

    Awstats and webalizer both strip the src parameter, making exact tracking by them…*challenging*

  5. Aleksander L. says:

    Hello!
    Because I do not know your e-mail address, decided to ask here. A few years ago I bought GSB DLCs through the website (Swarm, Tribe, Order, Nomads): is it possible to get the Steam keys for them? Thanks.

  6. Sam says:

    I would guess Reddit has a higher than average proportion of users running some sort of tracker blocking tool, but the discrepancy in Google Analytics does seem too high.

    Can you do a simple grep of your web logs for that src string in the URL (you’ll have to filter out lines where it appears in the referer)? That should at least eliminate any magic going on in the stats packages.

    The other thing that occurs is bots – it’s possible there’s something sufficiently clever to pull up Reddit ads and click through, though one would hope not. Are there any unusual patterns in the user-agents for incoming hits with that src string?

  7. Kashaar says:

    This may be a different discussion altogether, but I personally can’t stand source parameters as parts of URLs. If I want to share a page I find interesting with someone who I think might enjoy it, I have to strip those manually. If I give them the unchanged URL, it probably distorts statistics all over. And most people share e.g. youtube links by just clicking the browser’s address bar and copying what’s there.

    It also feeds into that whole discourse of browser tracking that just comes with all sorts of negative feelings. If I see that the link I’ve just clicked on has source parameters in the URL, I immediately get annoyed at the page.

    Of course, I may not be anywhere near the norm in that regard…

  8. bills6693 says:

    I bought the game following an ad on Rock Paper Shotgun, so thats one sale from me. I also bought Redshirt while I was at it, which I would not have pre-ordered if I hadn’t already been here preordering Dem3

    I have to say, the only reason I even wanted to preorder them was also because of RPS’s coverage. Plus the good will you earned from GSB with me.

    So really, RPS was responsible for both purchases :)

    Also one other thing, you are UK based, but is the company UK based or overseas? Because it seemed like it was saying it was overseas which was odd.

  9. cliffski says:

    I’m definitely UK based. I’m as British as Chelsea Buns.

  10. bills6693 says:

    Glad to hear!

    Anyway still. Yes, the RPS ad made money since otherwise I can almost guarantee I would have waited for 75% off sales on steam for those titles! Instead I got full price preorders :)