Is CPM Bidding a Waste of Your Money?

By ,

103 SHARES

Yes.  Full stop.

Answering that question and then calling it a day doesn’t make for a rewarding blog post, though.  I have found reasons that allow you to say CPM bidding wastes your money, and there are also some cool insights I’ve found about two of the biggest categories on the Google Display Network.  So even though my titular question has already been answered, please read on to find out exactly how CPM bidding wastes your money.

Like us, you may have thought it was a waste previously.  Also like us, you may have wanted some data to back that up.  Within the walls of Hanapin I asked around and found that no one had used it before (since we are all so direct response oriented).  It was time to experiment! After concluding a lengthy (albeit small) test with Google’s CPM bidding option, I’ve found that CPM bidding fails the smell test in pretty much every metric I could think of.

First, a quick explanation of CPM bidding: CPM stands for cost-per-thousand impressions (I’m almost positive they’re using mil {which is any number of romance languages for “thousand”} in case you’re wondering where the “M” comes from).  Your CPM bid is how much you’re willing to pay for a set of 1,000 impressions on the Display Network.

Google says that CPM bidding is best “for advertisers who are focused on brand awareness.”  For people who want “sales or website traffic” they helpfully point out that CPC bidding “might be a better option.”  Might it, dear Google?  Might it?

Both personally and professionally, my mind boggles at the thought of these mythical advertisers who are interested merely in brand awareness.  I’m sure there are a couple out there, but all of my clients, all of my hypothetical clients and even all of my hypothetical hypotheticals would follow a model where they want some sort of interaction based on their advertising.  If that wasn’t the case I would just recommend that I buy them an ice cream truck and drive it around shouting their name through the PA system (this has the added bonus of free ice cream for me).  Someone’s bound to hear that, right?  It’s branding! (Further note: even if this stupid ice cream truck were a business of mine, I still would want my advertising to be “sales or website traffic” focused, thereby proving my earlier statement about hypothetical hypotheticals.)

Putting aside all of my internal concerns about the logic of “brand awareness,” CPM bidding has some potential – maybe it’ll give you access to sites you wouldn’t normally get with your traditional CPC bids or maybe your CTR will be so good on CPM bidding that you’ll be able to save some money on clicks.  Based on my (again, small) experiment, this just isn’t the case.  On to the numbers!

Cost Per Thousand Impressions Performance

So over the same amount of time, here’s what gains you can expect CPM to get you:

-More impressions

-Cheaper CPM

Here’s where it will hurt you:

-Click volume

-Click through rate (I used the same ads throughout the whole experiment, so it can’t be blamed on ad testing)

-Cost per click

-View Through Conversions

That last one is probably the most damning one of all.  If the point of CPM bidding is to get your ad out there, then why do all of those increased impressions actually decrease your view through conversions?  I recognize that view through conversions aren’t a perfect metric, but they’re still an indicator as to whether or not your ads are at least in the proper range of users you want to attract.

I’ve heard arguments that CPM bidding allows you to get on better web sites that normally wouldn’t show your ad.  The argument basically boils down to higher site quality or at least higher quality placements in general.  (I wanted to consider CPM bidding from all of its angles before giving it the headline of doom.) **SPOILER ALERT** This argument is also false.

By looking at category performance over the time frame of my test I was able to see how AdPlanner 1000 sites did against non-AdPlanner 1000 sites (just by subtracting out non-AdPlanner category statistics from the overall numbers), since these are theoretically high-traffic sites that can do a lot of branding for you.

AdPlanner 1000 Performance with CPM Bidding

Key takeaways from here:

-AdPlanner 1000 actually has a lower percentage of the impressions on CPM bidding (91.4% AdPlanner sites on CPM bidding vs. 96.7% AdPlanner sites on CPC bidding)

-Non-AdPlanner traffic has a much higher CTR for both campaigns (although the 3.33% CTR during the CPC bidding phase of the experiment is disconcertingly high and warrants further investigation)

-The actual CPM of AdPlanner sites is a lot cheaper with CPC bidding

-View Through Conversions are drastically better with CPC bidding, even more so when you look at that stat per 1000 impressions (labeled in the chart as View Through/Impression)

So, the whole “quality of site” argument surrounding CPM bidding was put to bed, but there was also one more category that I took a look at specifically: whether or not the ad was above or below the fold (basically, the quality of your impression).  Again, you’d expect (or at least I expected) CPM bidding to have a higher percentage of impressions above the fold.  You’re bidding on visibility, after all.

CPM Bidding Above/Below the Fold

First off, as with the AdPlanner breakdown, the impressions that are theoretically more valuable (above the fold) don’t perform as well.  That holds true across both CPC and CPM campaigns – CTR and View Through Conversions are both a lot better below the fold.  (It’s a shame that you can’t track performance by site within each of these categories.  It’d be great to see which sites on your placement reports are above or below the fold.)

In any case, here are the main items that I took away from the data:

-Actual CPM (for either CPC or CPM bidding) is cheaper with impressions above the fold.  It’d be great to see the quality of those sites, but I don’t know of a way to do that.  (You can only exclude categories at the campaign level, and it seems like a bit of overkill to have one campaign for above and below in addition to another campaign with the same targeting that only targets above the fold placements.  That’d be tough to manage.)

-Impressions below the fold do a lot more in terms of view through conversions.  Another mystery that I’d love to solve some day is what percentage of those impressions involved someone scrolling down to your ad.  “Below the fold” can be way, way down the page depending on the site, so I’m very curious as to what percentage of users really scrolled down that far.  The high preponderance of impressions to view throughs could also just mean that the sites that have below the fold placements are more relevant to whatever I was pushing online during this experiment.

-CPM bidding does not buy you a higher percentage of impressions above the fold.  CPC bidding got 97.7% above the fold.  And while the crazy CTR of the below the fold placements may account for so few impressions there, I still think that it shows CPM bidding isn’t going to get you a more visible spot online, even though its purpose is one of branding.

So, the stats are worse, the sites are of lower quality on average and the placement on those sites is also of lower quality.  I think I’ve proved why CPM bidding wastes your money.

Has anyone seen anything different?  It’s always been a scary idea to us here at PPC Hero, and this isn’t doing it any favors.  Who wants to speak for CPM bidding?  It won’t speak for itself (it’s too busy wasting your money {Ba-zing!}).

Get more weekly links with our Fast Five newsletter! Five Fast Links in Your Email Every Friday.

Also send me a daily RSS digest

PPC Paradise Lost

Twitter Facebook LinkedIn Google+ Email Print More
  • Jan B

    Thanks for this topic. It’s an interesting one. But I have a fundamental question:

    Why didn’t you adjust the CPM bid of your test campaign to the CPM that your CPC campaign achieved? So basically bidding 1.99 $ or 2.00 $ as a CPM?

    As the content provider is selling his ad space on CPM basis it is the key criteria to get comparable results. For this test case I would assume that you got completely different placement (and quality of placements) because you only wanted to pay 1 $ for CPM campaign, while you granted 1.99 $ CPM in the CPC campaign.

    For me you priced yourself out of quality ad space with this bidding and basically ensured that your CPM campaign gets lower CTR and therefore a higher average CPC. In the end CPM is the final currency to buy space. If you just invest 50% with your CPM bidding, you are not into the same neighborhood that your CPC campaign is in.

    • http://www.hanapinmarketing.com/ Sean Quadlin

      Hi Jan,

      Thanks for your question! I’m just showing the results here. The bid was actually set just about where you suggested it should have been. It was a small amount lower, so your points could have been at play a little bit here, so thanks for pointing them out.

      There was plenty of headroom between my bid and what we ended up paying, though. The placements should have been a lot more comparable based on our bid vs. the previous results of the CPC bidding period, in my opinion and based on the logic that you point out.

      A fair point, so anyone trying out CPM bidding should absolutely keep that in mind. Thanks for passing it along!

      • Jan B

        Ok thanks for the feedback! I would propose for anyone testing this to even bid higher on Max CPM to achieve a comparable CPM in the end to his CPC campaign. It’s really tricky and we are playing around with it a lot.

        But I agree overall. We would always try CPC bids before CPM bids, especially as clicks are much more valuable since remarketing became such a powerful tool, and normally you know the CPC you want to pay better than what you want to pay for a view :)

        • http://www.hanapinmarketing.com/ Sean Quadlin

          You are absolutely right. We should have put that bid higher to be as fair as possible with the experiment. If you find out anything cool with CPM bidding please pass it along! We’d love to hear any results.

  • AdamBerke

    There’s a bit of a false dichotomy here. Using CPC bidding in Google Adwords actually is CPM bidding. Google is just managing the CPM bids for you. When you use the CPC pricing option, you’re using Google’s (very good) CPC optimization engine to manage your CPM bids. In fact, Google themselves is bidding on a CPM and pricing (via arbitrage) back to you on a CPC. Google AdWords is the largest buyer on the DoubleClick Ad Exchange, which is all CPM bidding.

    When you use Google’s CPM bidding, you’re basically just turning off their CPC optimization. So it’s not surprising that the CPCs would be worse. However, there are ways via RTB enabled DSPs to bid on a CPM, but optimize towards a CPC (or other performance models) very effectively. For example, if you want to bid on specific users based on what you know about them, ie put something in their shopping cart vs bouncing off the homepage, it would be better to do this using a DSP that can take that info into account and adjust your CPM bids accordingly. Only bid a little for the bouncers, but a lot for the cart abandoners. That would be a more effective strategy than just using Google’s out of the box CPC optimization. As @disqus_9mspVOij70:disqus noted, you’re just placing a flat CPM bid without adjusting it any way to optimize towards a goal, so that’s not really a good portrayal of best practices or what can be accomplished today via CPM bidding and RTB.

    Google’s bread and butter is CPC, so they obviously do a great job at optimizing for it. However, simply turning the optimization off and saying CPM is bad, doesn’t quite paint the full picture. I hope that’s helpful. I’d be happy to discuss further.

    • http://www.hanapinmarketing.com/ Sean Quadlin

      Hey Adam,

      This is helpful, so I appreciate you passing it along. I should have mentioned how AdWords handles CPM/CPC stuff behind the scenes, so thanks for doing that for me!

      My experiment was definitely a low level one that was meant to just try out the little-used (within my own company) bidding option within AdWords itself. You mention a lot of high level bidding strategies that are always going to be the better way to go. My focus here was just on whether or not you might want to experiment with AdWords CPM bidding, and based on my admittedly small experiment it was a no.

      • AdamBerke

        Right on Sean. Just wanted to weigh in on the CPM side of things per you’re closing statement. Those CPMs are notoriously tight lipped when speaking for themselves ;)

  • http://twitter.com/GnosisArts Gnosis Media Group

    Thank you for doing a use case and for a really thorough breakdown, Sean.

    It seems that one conclusion that can be drawn from your case study is that CPM is really only preferable when impressions is your primary objective. I can see this in cases where we put together ad creative that we *don’t* want the viewer to click. Rather, we want them to take an action on the ad content itself. e.g., “Text DO IT to 00000 to ..”. It isn’t necessarily just a brand awareness campaign, nor is it just a conversion-oriented campaign. We want them to take an action, but we don’t want that action to be a click. It seems to me that, in this scenario, you would want as many sheer impressions on the ad (eyeballs), from the relevant audience, as possible. So, if CPM gives you a greater number of impressions than CPC on average, then this seems to be the one case where you’d want to choose a CPM bidding approacj over a CPC bidding strategy.

    Eric Bryant, CSMO
    Gnosis Media Group

    • http://www.hanapinmarketing.com/ Sean Quadlin

      That’s a very fair point, Eric. I completely agree with your point. I would caution to keep an eye on site quality, though, as not all impressions are made equal.

      Thanks for pointing it out!

  • David Wolf

    Hey Mark, I agree with everything you checked out in your test. Thanks for sharing all the data. I am going to speak for CPM for one reason and one reason only……. Many national media buyers are used to cpm models and it is an inconvenience for them use cpc for ppc campaign and cpm for everything else. CPM meant access to more buyers of Google inventory. Other than that, from one direct response to another……long live CPC (CPA is a whole different story)

    • http://www.hanapinmarketing.com/ Sean Quadlin

      Thanks for passing along the thought, David! If you’re working with national media buyers it’d probably be a better idea to still bid on CPC and then convert it to CPM for reporting purposes. It doesn’t seem to me like Google’s CPM option is worth fooling with. That might work for any advertisers tasked with buying everything on a CPM basis.

      Long live CPC!

  • Santosh Prasad

    Sometime it really is..but doesn’t it proportionally depends upon the nature of campaign too??

    • http://www.hanapinmarketing.com/ Sean Quadlin

      As is the case with pretty much anything in PPC, it absolutely depends. If you’re looking for direct response on Google’s Display Network, though, I’d recommend using CPC bidding. If your client wants to see CPM numbers you can still provide them, but this option gives better control and works better for the advertiser from my experience.

  • Gnosis Media Group

    I’m not sure I’m in agreement with the premise of this post now. We have a GDN campaign on CPM that is generating a ton of VTCs. Now, leaving aside the question of the validity of VTCs, I think that VTCs might tip the scale a bit towards finding value in CPM bidding.

    • http://www.hanapinmarketing.com/ Sean Quadlin

      My display campaigns run either CPC or CPA bidding and they still get plenty of VTCs. I think it may be possible for a CPM campaign to generate more VTCs, but just because you’re seeing success on a CPM campaign it doesn’t mean that you wouldn’t have even more success with a different bidding option.

      Is the CPM campaign a brand new one? Or did you switch it from CPC bidding and see a nice lift? I could see a CPM campaign type being successful, but I think Google’s systems handle CPC/CPA bidding a bit better (especially for direct response campaigns).

      Whatever the case, if CPM is working for you, that’s great. Maybe I’ll give it another shot sometime and see what happens.

      • Gnosis Media Group

        Oh I see your point. You’re saying that you may get better performance overall by bidding CPC or CPA vs CPM. Ours is a new campaign that we started off using CPM.

        • http://www.hanapinmarketing.com/ Sean Quadlin

          If you find out anything cool let me know! I hope that performance remains strong in your campaign.

  • Gnosis Media Group

    Has anyone tried or considered creating two identical GDN campaigns, and running one on CPC and the other on CPM for same targets (same placements, same remarketing audience, e.g.) and competing them against one another? Just brainstorming out loud.

    • http://www.hanapinmarketing.com/ Sean Quadlin

      That would be a great test to run! If anyone ends up doing it please share the results with us.

  • Gnosis Media Group

    Also an argument could be made that, for a remarketing campaign, the placement is less important than you suggest above.

    • http://www.hanapinmarketing.com/ Sean Quadlin

      Remarketing definitely plays by a different set of rules. Thanks for calling that out!

  • http://www.hanapinmarketing.com/ Sean Quadlin

    As is the case with so much of PPC, your results will definitely vary. From what I’ve seen in Facebook it’s great to use CPM if you can greatly exceed average CTRs. But I’d recommend CPC bidding within AdWords for most advertisers.

  • Sandra Taramtamtam

    CPM marketing is more for brands to be seen, not to sell or generate clicks, thats for sure. Thx for the analyses http://www.hotels-aroundtheglobe.info

  • Jasonmailley

    Strategically, there is no “one-size fits all” bidding model for ads. I will usually set my campaigns on CPC, tweak, optimize and test my conversion. Once I know all my stats, I will migrate it to a test-campaign on CPM, verify if my CTR matches and check if the conversion is similar. If CPM ends up getting me cheaper lead-acquisition cost, the decision is not hard to take. Obviously, on similar performances, I will opt for the less riskier model, CPC.

  • http://nettoyageimperial.com/fr/montreal/nettoyage-de-conduits-dair David lax

    Thanks nice shearing