Google’s eCPC Can Solve All Your Bidding Problems – Or Can It?

Over the past couple of months we’ve been testing Google’s flexible bid strategy ‘eCPC’ –  or ‘Enhanced CPC’ –  with our clients to discover which strategy works best for which kind of account and in which circumstances. This method allows Google to increase your bid by up to 30% or decrease it by up to 100% in each auction depending on the likeliness of a conversion. Factors that are taken into consideration are real-time details such as browser, location, device and time of day. By activating the eCPC Google claims that you can achieve more conversions at a similar or even better CPO.


The test was run across several accounts from different customers and several markets in both Search and Shopping campaigns. Each test campaign was duplicated twice while the original one was paused so that performance wouldn’t be influenced by history. To avoid cannibalisation the campaigns ran on an alternating 2-hour schedule.

As Google needs performance information in order for the eCPC to work, we chose campaigns that had at least 100 conversions per month and let the test run for several months in order to allow for a four week period of data collection.

For the last weeks of the testing period we uploaded some standard RLSA lists but without bid modifiers to some of the campaigns as we wanted to find out if Google bases the anticipation of a conversion partly on information that’s provided by remarketing audiences.


eCPC Shows no Noticeably Positive Effect on Shopping Campaigns

In Shopping the eCPC campaigns did not perform better than the non-eCPC campaigns:


While the overall differences were very small, there is a tendency for CPCs, CR, and number of Conversions to be higher in eCPC campaigns. However, the higher CPCs outweigh the higher CR so that the effect on the CPO is negative.


Search Keeps What Google Promises

In our Search campaigns on the other hand we could see performance develop as promised:


CPCs were only 1% higher in the eCPC campaigns but CR was more than 7% higher, leading to a better CPO.


While our Shopping campaigns showed a similar amount of traffic our Search eCPC campaigns could achieve more clicks even though the number of Impressions was lower. This was due to a noticeably higher CTR.

The combination of more clicks and a better CR indeed lead to 10% more Conversions than in the non-eCPC campaigns as google promises.


Traffic is Shifted Towards RLSA

The RLSA test shows pretty interesting results:

Clicks & CPC

While the amount of traffic coming in via RLSA lists is only 1% higher in our eCPC campaigns than in our non-eCPC campaigns, there is quite a big difference in CPCs.

As mentioned above we did not push our RLSA lists via bid modifiers and while the CPCs in the non-eCPC campaigns are almost identical, the eCPC campaigns show quite a variance. CPCs of clicks generated via RLSA lists are 10% higher than the CPCs of non-audience clicks. This shows that google does indeed shift traffic towards audiences as here the probability of a conversion is higher.

The eCPC is a Great Concept but Google Still Has Some Work to Do

Overall the variance in performance between eCPC and non-eCPC campaigns was not as significant as we had hoped for. However, activating the eCPC in Search can give your campaigns an additional push towards efficiency and additional conversions that could increase the amount of new customers for your shop. In Shopping however, the difference between eCPC and non-eCPC campaigns was not big enough to draw any conclusions. It does not seem to have a big influence on performance neither in a positive nor a negative way.

The RLSA results showed us that Google tries to work with users they already know which is a great concept but still needs some work to be done in order to work significantly better than the standard campaigns, especially in Shopping.



[shareaholic app="share_buttons" id="19406647"]


Madeleine Bitsch

Madeleine Bitsch works as a Business Intelligence Manager for crealytics and has 3+ years experience in PPC and in managing our biggest customers in the area of fashion. She has a proven track record in client communication and strategic testing.

    Find more about me on:
  • linkedin
  • Patric

    Very interesting test.

    Did the test run on enhanced campaigns or did you isolate desktop and mobile traffic?

    • Madeleine

      Hi Patric,

      Good question! The test did run on enhanced campaigns with Mobile Bid Modifiers in place. These however were not adjusted during the duration of the test and the mobile traffic volume was quite low so we did not isolate it for the analysis used in the post above.

  • Sam Knapp

    Very interesting Madeleine! Going a little further into the device segmentation, I have been searching for a way to place modifiers on tablet traffic and eCPC supposedly can do that. Did you notice any big difference between the tablet performance between the control and experiment sides of the test?

    • Madeleine

      Hi Sam,
      performance on tablets actually was a bit better in the campaigns with activated eCPC.
      While traffic was about 10% lower, conversions ‘only’ were 5% lower.
      CPCs were slightly lower, too, while CR was a little higher. The result was CPO being about 10% lower in eCPC campaigns.

      One of the things that Google takes into consideration when adjusting the bids is device, so as tablet usually performs weaker than desktop it makes sense that Google decreases the bid for tablet impressions, thereby leading to lower CPCs and less traffic but better CPOs.

      So it seems if you want to boost your traffic on tablet eCPC might not be the best method but if you aim to improve tablet performance eCPC can be a way to give your campaigns a push.

      • Sam Knapp

        Thanks Madeleine!
        That is exactly what I was looking to find out more about–tablet performance has not been as good as desktop so hopefully eCPC will lower the tablet bids since there is no way to set that manually!

        • Madeleine

          Great that I could be of help! Hopefully activating the eCPC will help you optimize your tablet performance!

  • Lucas Ertola

    Hi Madeleine!
    Great Post! Ive just discovered and Ive read all the post from last year until today in one day 😉
    I have 2 questions
    1-Ive noticed that a common practice you have is to use this 2-hours schedule to split imp during the day between campaigns. Is there a reason why you use 2 hours, and not 1 hour. My first though when I read it was that you are probably not distributing equally daily impressions as if you were using 1-hour. Performance ratios wont change, but volume might be more evenly. But maybe there is reason you decided to do it this way.
    2-One question not related to this topic.
    I would like to know your opinion regarding Google’s Algorithm to pick the winner ad when you create a new campaign or adding a new ad? Do you blindly trust on it, or do you always set rotate evenly to take your on conclusion when testing new ads?

    Thanks and keep writing!

    • Madeleine

      Hi Lucas,
      Sorry for my late reply and thanks for your comment! Happy that you find our posts to be informative.

      Regarding your questions:
      1. Google accepts only 6 schedules per 12 hours, so it’s not possible to run a test on a one hour alternation. We do however make sure that the campaigns also alternate daily: If one campaign is active from 2-4, it won’t be the next day. This minimizes the influence of performance differences according to hour of day.

      2. My colleague Andreas wrote a blog post regarding this topic a while ago. It might help answering your question:


  • sem champ

    White label PPC
    Really it was a great post, got to learn something new.
    Thanks for sharing the post.