What is PPC Cannibalization ?
PPC cannibalization occurs usually when you have strong organic search engine ranking and you are running a PPC campaign on top of that. It’s the % of organic clicks you would have had but that you ended up paying for because the PPC campaign attracted the click.
You may have heard that running both increases the % of searchers going to your website and that it is basically a good thing. It’s sadly not always the case, but you can measure that, at least, and it’s one of the best things you can do to optimize a PPC campaign. The basic problem is, a % of the traffic you would get for free is becoming paid traffic, and that messes up the ROI of both the SEO and the PPC campaign. Knowing exactly where to put your SEO and PPC money for the best ROI is an admirable advantage in the SERPs war.
Avinash Kaushik suggests in his book: “Web Analytics: An Hour a Day” a method to measure cannibalization. You can either pause some Ads Groups of your PPC campaign and then measure Key Performance Indicators of the organic traffic during a week and then put the Ad Groups switch to On and measure the difference in the organic traffic. Or, you can even stop PPC completely for a week and do the same thing. The difference between the 2 weeks of results is what is called cannibalization. Try to choose a week that you do not expect to be affected seasonality (don’t use Christmas week, for example).
Using this method you have a way to calculate the real return on investment for your PPC campaign. Let’s say your PPC and organic campaign convert at 3% on your brand name keyword. After having made the blackout test, you realize than you lose 35% of your organic traffic on this particular keyword when the PPC campaign is on. To make thing simpler, let’s assume you get 1000 visitors a week by organic (when PPC if off) and 500 visitors from PPC in a week on that keyword. The average CPC of this particular keyword is 1$ and the average conversion worth 50$ in revenue. Doing some basic math, here is the real cost of your PPC campaign.
You can see from the table that you are making less profit by doing PPC on this particular keyword, even if you are generating more conversions. You may want to go that way because more sales can equal quantity discounts for you and end up with more profit. But usually, you are wasting money pursuing this kind of strategy, and you should move your PPC budget towards more profitable keywords (like ones you don’t rank on organically).
- SES Toronto : Web Analytics Track
- Live Blogging SES Toronto 2009: Introduction to Paid Search
- 6 Steps to Keyword Success
- Using Google Universal Search to Achieve Top Rankings
- 4 Musts for AdWords Optimization
Posted by Gary said on 21.03.09 @ 4:13 am :
Great post! I agree that you should measuring the synergy between paid & organic. I have to say that to measure true cannibalization you have to test terms that have high organic ranking & more often times than not It’s these terms are going to be branded terms. We just recently did a test at my job to measure cannibalization & we found that for the terms that had the highest cannibalization the CPC’s were so low that there was no real savings by cutting back.
Posted by FUJIO said on 23.03.09 @ 9:34 am :
Great Post too! I found this cannibalization in the last year. PPC is not so effective for B2B site depends on Keyword and Ad messages. We lost too much Ad costs to PPC. I had checked by ROAS against CTR.
Posted by radith said on 23.03.09 @ 11:10 am :
nice article. i’ve RT it on twitter
Posted by Henk Jan Dulon Barre said on 23.03.09 @ 11:46 am :
Greatpost! I will use it in my argumentation to customers
Posted by Chris said on 23.03.09 @ 1:35 pm :
I think if you’re only getting a 15% boost on your overall traffic from your PPC campaign, you’ve either :
1) done a great job with your SEO campaign
2) done a terrible job on your PPC campaign
And, ^^ agree on the comment about the branded terms.
Posted by Francis Vallieres said on 23.03.09 @ 1:44 pm :
The 15% boost is totaly hypothetical. it was just to make the example clear.
Posted by Jeff James said on 24.03.09 @ 9:27 am :
Many people would gladly pay more for additional customers…especially since either the competition will get them, or they will.
Extreme case - I would demote the paid ad to position 5-7 and see what happens.
Jeff
Posted by James Faulkes said on 30.03.09 @ 5:29 am :
In principle, the logic you put forward makes sense, however…
You’ve neglected to mention that in the “real world” PPC and organic traffic often have a much more complex relationship than straight forward traffic and conversion contribution. Tracking conversions on a “last click wins” basis will not give you the true picture as SEO and PPC can effect the conversion rates of each channel. And that’s before you take into account other media which might be running alongside.
Because of this more complex contribution effect, natural fluctuations and many products having a sales cycle which is longer than a week, comparing week on week performance is pretty rough and ready as there can be huge variations in performance week on week for many businesses.
If users directly copy this for their own businesses without truly understanding their user behaviour there is a risk they will be acting under false assumptions which might cause their sales harm.
Just to re-iterate, I’m totally in agreement that a company can save cost by encouraging users to click on organic listings rather than paid ads but I’m sceptical that a limited analysis such as the above gives users enough information on which to base that decision.
Apply the cannibalisation calculations to your marketing but be sure to apply them in a way which makes sense for your business and not based on generic advice.
Posted by Boris Platt said on 30.03.09 @ 7:05 am :
I have read Web Analytics: An Hour a Day. A great insight into how to best harness the power of google analytics. I did feel rather inadequate when comparing myself against the section headed ‘Ten signs of a great web analyst’ (which I am not). I do however concur with Gary and James’ comments above.
Posted by Guillaume said on 02.04.09 @ 9:26 am :
@James 100% with you there. I think Francis wanted to lay out the concept, but it requires much more in-depth analysis to understand the real effects. I am very surprised there is so few studies about this - and I am pretty sure Google would not be really keen on that subject matter.
@Jeff Agreed with you - but considering the CPA is on a normal distribution curve, as soon as you are in the downward spiral, you might get to a point where you don’t want those extra conversions!
Posted by Mark van Loon said on 19.11.09 @ 3:00 pm :
I always disagree with my clients if they want to run a PPC campaign on business or branded keywords. They always show up first on the organic position, and those customers are looking for you (and nobody else). Otherwise they would have typed a generic keyword.