Brad Geddes / PPC Geek
Official Google Ads Seminar Leader.
Author of Advanced Google AdWords.
Co-Founder, Adalysis.
(703) 828-5811‬
Brad Geddes's Theories on Marketing Why traditional campaign structures belong in the past

Why traditional campaign structures belong in the past

This is a guest post by Bjorn Espenes, founder of several Internet businesses including one that processed over $1 billion in online sales for clients. Most recently he co-founded Finch with Eric Maas bringing 10 years of building optimized eCommerce software experience to the PPC industry. You can read more about how to optimize your Pay Per Click on the Finch Blog.

Most companies with an eCommerce site structure and organize their Google AdWords campaigns like this:

  • 10+ related keywords in an AdGroup (usually a product or product line)
  • 3-5 ads that are relevant to the related keywords
  • Set the CPC for the AdGroup based on CPA target for the products

Sounds familiar? This structure is very effective for someone managing the campaigns manually or by a “power tool” with a CPA performance target to hit. For the real performance of the campaign (and for the business value) it is worthless. Why?

Related keywords in an AdGroup

You may think that by putting product related keywords into an AdGroup you can now control the ad spend on those products and as such manage the cost of selling those products. Well, guess what? You have no control. Those keywords may drive the sales of other products on the site, and other keywords may drive the sales of the products you are trying to control. You simply cannot control what people are searching for and what they are buying; especially not by grouping keywords in AdGroups!

3-5 ads to a group of keywords

The objective is to create relevant ad copy to related keywords and make the ad copy management more scalable (and convenient for the marketer). Let’s say that one of the ads is performing poorly and you make changes to try to drive the CTR (click through rate) up, and two days later you check in to see how the changes performed. You see that the CTR is the same and draw the conclusion that it made no difference. The reality may be that the ad copy change doubled the CTR for one of the keywords and cut it in half for another keyword; you would never know because you are looking at averages and totals. AdWords success is a result of the keyword (triggers the ad), the Ad (generates the click), and the landing page (captures the conversion). If you could isolate each ad to the keyword that triggers its display, would you like that? Would that enable you to optimize your ad copy for each keyword? See the benefit in doing this when competing for the 1 click there is to be had per Google search?

Rather than using one Ad Group for many keywords, try one Ad Group for each keyword. This allows you to look at and optimize the performance of each keyword separately. Now it’s easy to sort by the keywords that are getting the most impressions, look for the ones with below-average click through rates, and then get to work on optimizing those keywords with ads with laser-pointed accuracy.

Set the CPC for the Ad Group for related keywords

By now you are getting the drift I assume. Each keyword earns its own right to spend your money. If all the keywords have the same conversion rate (or pull in the same amount of revenues relative to ad spend), then the traditional strategy works. But it is never like that; each keyword is unique and needs to be treated accordingly (this gets even more complicated with match types, geo targets, day of the week, etc.). If you spend the same across the board, you live a life of averages with AdWords, and that is no way to compete if you want to WIN! To capture the clicks that drive the most sales/revenues for you, you need to bid accordingly and based on the history of each keyword.

In the new campaign structure you will have the following:

  1. For each keyword, there will be a broad, phrase, and exact match equivalent
  2. Each keyword will be bid based on its own conversion rate (or preferably revenue/profit returned). Exact and phrase matches will have higher rates, and you’re willing to pay higher CPCs for these.
  3. Each Ad Group now has exactly 1 keyword. You may start with template based ads, but over time you’ll tune these to be very specific to the keyword – focusing on the ads with the most impressions.

Setting the same CPC for all keywords in an Ad Group is just silly and will waste your money while you lose the valuable clicks to your competitors.

Campaigns need to be structured with the purpose of isolating the variables that matter and impact conversion rates (or revenues/profits) inside AdWords, not for campaign manager convenience. If you want to see an example of what Finch can do for your account, why we would do it, and what you can expect — visit us at

This is a guest post by Bjorn Espenes, founder of several Internet businesses including one that processed over $1 billion in online sales for clients. Most recently he co-founded Finch with Eric Maas bringing 10 years of building optimized eCommerce software experience to the PPC industry. You can read more about how to optimize your Pay Per Click on the Finch Blog.

Opinions expressed in the article are those of the guest author and not necessarily bgTheory. If you would like to write for Certified Knowledge, please let us know.


  1. tdwhalen
    May 24, 2012 at 10:38 am · Reply

    Hi Bjorn,

    I agree that a lot of folks structure their PPC accounts similarly to your first 2 bullet points:

    — 10+ related keywords in an AdGroup (usually a product or product line)
    — 3-5 ads that are relevant to the related keywords

    But I don’t think many people bid only at the ad group level (bullet point #3). Certainly not folks who read this blog, eh? Bidding at the keyword level has really nothing to do with how many keywords you have in each ad group. Plenty of good PPC folks have more than one keyword in some or all ad groups, but they would never bid only at the ad group level.

    Also, I am a fan of 1 keyword to 1 ad group in many different scenarios, but I also think there are many scenarios – including PPC programs that have recently been started from scratch – where this is not the best solution. Although tying 1 keyword to 1 ad group almost always sounds good in a blog post or in conversation, there are challenges with this approach. One challenge is that if you have 3 ads in each 1-keyword ad group as you suggest, you will have hundreds of ad groups where you have to wait months to get any meaningful ad test data. Better in some cases to (thoughtfully) roll these up and do your ad testing that way; then, once you have some good ad data, go ahead and do your re-structure.

    Again, I am a fan of your structure in many scenarios and with many advertisers – I just don’t agree that this is the only way to skin all cats!

    • bjorn
      May 25, 2012 at 4:25 am · Reply

      Thanks for the comments!

      I can see your point for new campaigns as it would help a new campaign get off the ground faster, but on the flip side why not collect granular data if you can?

      The biggest argument for making the ads this granular is to optimize more effectively, and while it may take an eternity to optimize the low traffic ads…the high traffic ads can be optimized very quickly this way at a keyword level.

      There are definitely more ways to structure campaigns, I just happen to be very passionate about this one 🙂

  2. Adrian Bold
    May 25, 2012 at 2:41 am · Reply

    Excellent write-up; thanks.

    I’ve long said that, in an ideal world, each keyword would have it’s own ad group. The issue, as always, is one of time to implement this practically, particularly when managing multiple smaller accounts.

    • bjorn
      May 25, 2012 at 4:28 am · Reply

      Hi Adrian,
      Thanks for the comments! You are right on the time constraint on doing this, both the setup and the daily management. It is only feasible in my opinion to do this through software, both the campaign build and the daily management.

      • bezz80
        July 15, 2012 at 5:01 pm · Reply

        advice which software you use to build and manage campaigns optimization account?

        • bjorn
          July 17, 2012 at 4:36 am · Reply

 (I am a founder for disclosure) has developed the software that builds campaigns with this structure and manages the ongoing optimization. If you go to our site you can register for a free audit that will give you a detailed analysis of your current account setup/configuration with specific actions for how we would optimize your account.

  3. olafpijl
    May 25, 2012 at 9:15 am · Reply

    Granular campaign setup is definately the most important way to optimize PPC campaigns. And averages mean nothing. Good points!

    Just one thing: why would you already look for results after two days? In such a small period of time, differences in CTR could occur because of many other reasons (good weather, no weekend, etc.).

    This is one of the bottlenecks I often see in PPC campaigns: bid management decisions are taken on insignificant (read: not enough) information. That’s why I usually focus on impressions when deciding which keywords “earn” their own ad group.

    • bjorn
      May 29, 2012 at 7:04 am · Reply

      More data is always better, agreed! One of the benefits for structuring campaigns like this is that it takes very little time to get sufficient data around your high traffic ads, and a few days for your top ads should tell if the change was effective.

      If you look at your top 25 ads by impressions and pick the 3-5 with the lowest CTR (there are always a few that really stands out) and make adjustments to those where you can control both they keyword (match type) and landing page, chances are that you will gain major insights just after a short period of time.

  4. adhummer
    May 29, 2012 at 5:16 am · Reply

    While I agree your approach as an ideal scene, it isn’t very practical for smaller advertisers but it could be done.

    The campaigns can be done in Excel, which isn’t too difficult, using ad templates. Data could be later aggregated in Excel using a pivotal table (Google doesn’t do anything different here, aggregating the data within ad groups). Managing the campaigns would need more time but it should yield more results.

    For small advertisers, like myself, it makes sense to do this only for their top KWs.

    But, how do I get KWs into new ad groups without the QS dropping to 3? I have ad groups with 10 KWs, each with a fairly decent QS. Moving them doesn’t seem right.

    I noticed the CTR varies drastically on a per KWs basis, so it makes sense to separate them.

    • emma
      March 31, 2013 at 12:44 pm · Reply

      If you have a nice structure with display-URLs and relevant ads to your adgroups the QS should not decrease but the oposite. The historical keyword data does not disappear inside an account but stays with your keywords when you move them around. Would could be the 3/10 issue is if it’s low traffic keywords, when you break them out some match types might not have had proper traffic before either, so the 3/10 is an estimate google set before having real data. In many cases I just click my own ad once for that specific keyword and can see how the QS jumps up to 7-10/10.

  5. bjorn
    May 29, 2012 at 7:15 am · Reply

    We do both the campaign build and management with software so it makes it a lot easier. The easy part is building the campaigns, the trickier part is how to pull in data to keep the campaigns optimized based on their ability to deliver revenues/conversions in a run-time auction environment for clicks. The results from doing it this way are often game-changing, simply because you can put so much more leverage behind specific keywords based on more/better information than the other competing for the same click.

    I am not sure what is going on with your QS. This structure has in my experience improved the QS because you can tie the ad to the individual keyword/landing page which makes it much easier to optimize the CTR.

  6. idolw
    May 30, 2012 at 2:52 am · Reply

    And Google could fix all this in 1 day – simply show stats for each keyword-ad combination…

  7. davisj80
    May 30, 2012 at 1:20 pm · Reply

    Bjorn, Nice article. I assume that the next part of this optimized strategy would be to negative match the exact keywords in the broad adgroups to direct traffic to the appropriate adgroups. I’d like to add an experience that I have had that people should watch out for. In your post above you separate the match type at the adgroup level. I have implemented this strategy but at a higher level, the account level so that I could add 100’s of 1000’s of products at the adgroup level. Then I negative matched all exact terms in the broad accounts hoping to direct all of the exact traffic in the exact accounts. The problem is that eventhough the accounts were all in the same MCC, Google does not link the accounts match types to each other. So when developing the account structure keep this in mind.

    • bjorn
      June 1, 2012 at 5:07 am · Reply

      Thanks! Use of negatives is key to this strategy, and it has to be done in the same account.

  8. srulez
    October 25, 2012 at 8:35 pm · Reply

    I can see the advantages to this, but wont the account grow to extremely large very quickly? What If you mix the traditional method with this one. Once you find a high traffic converting keyword add it as a negative and then create a new campaign where it as the only keyword.

    • bjorn
      October 30, 2012 at 12:57 pm · Reply

      Hi! There is no penalty for growing the size of the account like this. The reason why we use this Ad Group structure is to handle both match type separation and ad isolation by keyword/matchtype.

  9. JasonT
    October 26, 2012 at 5:20 am · Reply

    srulez Personally, I like the alpha/beta structure more – it’s very concrete:

    • bjorn
      October 30, 2012 at 1:06 pm · Reply

      Hi Jason! The campaign just lets you group the keywords/ads/ad groups so that it is easier to segment. How you manage your work should direct the campaign structure and for manually managing a campaign this looks very logic.

      In our case we use software to stay on top of the keyword/ad Max CPC so it makes the campaign structure insignificant, even when we add new search queries as new keywords.

Leave a Reply