meaning

Image credit: Fatheed.

Your CPM rates mean virtually nothing. If you ever negotiate media buys with a website or an advertising agency, you’ll most likely get quoted in CPM rates. For example, you may have a max of $2 CPM (Cost-Per-Milli) set. That means you’ll pay $2 for every 1,000 impressions of your advertisement.

For those of you new to media buying, there are a few things you’ll want to watch out for.

  1. Your ad could be shown to the same person multiple times. Each time the ad is shown to the same person, it counts as an impression. You may have agreed to pay up to a $2 CPM, but you may really be paying $2 for 500 people to see your banner twice, or for a bot to see your ads 1,000 times. To avoid this, you’ll want to tell the network to show your ad to unique IPs only or limit display frequency to one impression per visitor per 24 hour period. Also known as frequency capping.
  2. In most cases, avoid sites that cycle advertisements. A lot of sites will show an ad for XX seconds and then display another in it’s place and constantly cycle in new ads. This way websites are able to display a lot more impressions of many advertisements and as a result earn more per user for themselves. The bad news for you is that a potential buyer could be reading your display advertisement and then a new advertisement takes its place before your potential buyer is able to click the ad. And in this case, the more competition is bad too.

These are just some of the overlooked mistakes in online CPM advertising. Also, remember to check that your ads will be in the visitors’ eye path with the proper targeting (demographics, interest-based, geographic, etc.).

Make sure you get my free newsletter because best practices change as we find new learnings and discoveries. Keep up to date:

Was this free article helpful? Share it!

16 thoughts on “Media Buying: The 2 Mistakes I’ve Made in CPM Buys

  1. Great tip. I do this with CPV and in fact have found playing with freq cap can turn a loser into a winner sometimes. Thanks for reminding us about this with media buying. Do you have a checklist you go off of?

  2. @AffArmy: All networks cycle banners on page load. I was recommending the avoidance of sites that cycle multiple banners in real-time (i.e., after initial page load).

    @Phoenix: Thanks for the comment. I’ve heard a marketer’s expression before, “It takes 7 impressions to get a sale.” That’s one of the reasons retargeting works so well. But, there are always some people that will never buy so I limit frequency.

    I have “mental” checklists on different aspects of media buying, but I’m not sure what type you’re asking about. My obsessions in media buying are copywriting and eye path.

  3. @Jordan: Yeah, I know; that’s what I wrote under the first point. The second point is something entirely different. It appears that you didn’t read the entire post.

  4. My mistake buddy! I’m guilty of skimming I guess; I even checked googles cache to see if you changed it up on me.

    Apologies!

  5. @Roderick: Although I’ve only ran offline advertisements a couple of times, you are correct, but there are differences. For a billboard, you are not paying CPM – you are paying a flat rate for a certain period of time. I’m not certain on how TV commercial rates are setup though.

  6. TV ads are bought on something like CPM, but they use ratings instead, think it was called Cost per Rating Point.

    Aren’t billboards based on the area traffic still? I looked them up recently and I remember something called ‘eyes on traffic’ or something like that.

    Curious, what about amount of creatives for media buys? How much is too much and how much is too little? I can’t decide if you should throw all 80 banners up or do 2 of each size first

    Great site btw, been reading it for the past hour 🙂

  7. @Miro: Great comment!

    For which stage you are in testing, it varies. If you haven’t already, read this post: http://shockmarketer.com/calculate-statistical-significance/

    I’d start with 2-4 of each size to gauge the traffic source, test the offer, and test various targeting channels. Banners are only part of the testing process. And, having too many banners will never give you statistical significance.

  8. Thanks
    Yeah I read it, it’s how I arrived on here actually :).
    Found it on Smaxor’s twitter stream

    @Roderick yes advertisers do pay to have their ad shown to the same people over and over. I believe TV media buyers/sellers use CPG metric that covers repeat viewers (Gross Rating Point).
    It’s complicated though. From the way TV ratings are measured using 1% of the tv market’s audience to stations selling air and online bundled in a package that gets discounted the more you buy. And then you have markets that are still on diaries where everyone has to fill out what they watched, when, how long several days prior. Or the electronic meters where people must push a button every time they are near the TV or leave the room, even their guests. And many forget to do it. Last year Nielsen said average of 8% under reporting or over reporting due to households not using the electronic meters right. Our stations here had almost 50% drop in audience overnight when the meters went live, same has happened at every other market meters went live in.

    Cable and satellite companies though are pushing for everyone to have a box so they can either offer ratings (TV stations can pay between $30-$50k monthly for the privilege to see how they’re doing) or incorporate them into ad sales with neighborhood targeting

    Billboards are measured with Eye On Impressions, evidently they somehow know or factor in how easily seen the board is and on average how many people check it out. Traffic in the area also affects pricing. I found some bilboards in the miami area that were going for $720 to $1000/weekly. Digital bilboards. The ones where you glue your ad cost more as you have to pay to print the whole thing and then for someone to climb up and stick it

Leave a Reply

Your email address will not be published. Required fields are marked *