5 {{ upvoteCount | shortNum }}
5 {{ upvoteCount | shortNum }}

Questions For dburk And Other AdWord Experts

by Trankuility
5 replies
  • PPC/SEM
    • |
I've started an Adwords campaign after reading Perry Marshall's Adwords book. I know much more than before reading the book but I'm also confused as hell.

My setup:
  • I have one campaign.
  • Within than campaign I have 8 ad groups.
  • Each ad group has between 1 - 4 tightly focused keywords.
  • Also, each ad group has 2 ads which I'm split testing against each other.
  • Each ad leads to its own unique landing page.
  • I'm bidding $1 per keyword

Current results:
  • Keyword Quality Scores range between 4 - 6.
  • Most ads don't show on the first page (I expected this)

My questions:

1) I'm testing the bids at $1 as Perry suggested. At what point should I increase the bids?

2) Should I immediately tweak my ads/landing pages to increase QS, or should I let them run to gather more data before tweaking (it's only been 1 day so far)?

3) How long should I split test ads before declaring a winner?

Before testing Adwords I was clear on how to target a niche, but now I'm all mixed up. Perry advises against using long tail keywords. Instead he recommends going for the larger market, and letting your split testing lead you to what people want.

I take that to mean that instead of using the keyword "weight loss for teenage boys", it means to use just "weight loss".

4) But then, how do you actually target your market that way?

If your product is about weight loss for teenage boys, your landing page will be about that. Which means you ad should be about that. Which means your keywords should be about that. So if you use keywords about just "weight loss" you'll get a low quality score when the ad and landing page are more geared for "weight loss for teenage boys".

5) How do you target your market within adwords then?

Last thing, Perry talked about how many advertisers are in a healthy market, but he didn't say much more. A lot of keyword variations I search on have few, or no advertisers. So I assume there's no money to be made there, thus I skip those keywords.

6) How do you know when a market is healthy enough for you to get in and compete?

Whew. Thanks for taking the time to read and reply!
#adword #dburk #experts #questions
  • Profile picture of the author LucidWebMarketing
    As been said again and again, QS is a comparative score of your ads to your competitors. Most of this is click rate, the only thing that can be quantitatively measured. So if your QS is low, it simply means that searchers did not find your ad as appealing as the other guys. A poor choice of keywords, usually being too generic, can also lead to lower QS.

    It is rare for someone to start out and have good QS right off the bat so don't feel bad (it would be worse if they were 3-5).

    >> 1) I’m testing the bids at $1 as Perry suggested. At what point should I increase the bids?

    Why $1?

    What is a typical rate for your keywords? If this was insurance types of keywords, it would be much higher. Maybe that's why you said "Most ads don’t show on the first page (I expected this)".

    Don't increase the bids until your QS is better, 7 or more is what I suggest. First thing to do before increasing bids is always to improve your ads and hence your QS.

    You could bite the bullet if you wanted to and increase bids. I assume you want a higher position. But, with lower QS, you'll be paying more than you should and certainly more than competitors.

    >> gather more data before tweaking?

    You definitely need enough data to make sound decisions. At least 20 clicks for each ad before you can say with confidence that one is better than another and more clicks is better. After one day, I doubt you have that much. But sometimes, you throw those out the window, especially if you see an ad doing so poorly against others there's no point waiting or you are trying to get your QS up.

    >> 3) How long should I split test ads before declaring a winner?

    See comment just above.

    >> I take that to mean that instead of using the keyword “weight loss for teenage boys”, it means to use just “weight loss”.

    Well, at least I agree with Perry there. For one thing, there's diminishing returns. Most people probably don't search much on your first keyword. So is it worth having that keyword in the weight loss niche? It would be if that's your target market but even then I'd use the BMM +weight +loss +teenage +boy first to explore but certainly would not use the broader "weight loss" if teenage boys were my target. I would use the phrase and exact "for" keyword if there's a large enough volume percentage-wise. It's not going to help much if there's only one percent of the overall volume on it so I tend to leave those longer tails out because that's just not how the vast majority of people search for most things.

    >> So if you use keywords about just “weight loss” you’ll get a low quality score when the ad and landing page are more geared for “weight loss for teenage boys”.

    Correct. So don't use "weight loss".

    >> 5) How do you target your market within adwords then?

    Starting with the most correct keywords as you said yourself in 4)

    >> A lot of keyword variations I search on have few, or no advertisers. So I assume there’s no money to be made there

    I've never understood this rationale.

    You use the keywords that pertain and describe your product. So what if there's no other advertisers? Doesn't mean there's no money there to be made. I just means there are no or few advertisers for that product or service.

    Take a client of mine. He sells aquatic weed control services, to clean lakes and ponds. How many other such businesses do you think exist in his area for that service? Just one or two others. There's few that specialize in this in the whole country. How many are using PPC? Even less. Does this mean he should not be using Adwords or forget about using keywords that would be relevant? I don't think so.

    Hope that helps and clears up a few things.
    Signature
    Profitable PPC Campaigns Don't Happen By Accident
    {{ DiscussionBoard.errors[9935580].message }}
    • Profile picture of the author dburk
      Hi Trankuility,

      It really doesn't matter that much where you start your bids, as long as it gets your ads to show on the first page, it is how quickly you adjust them that matters the most. I usually go with the estimated bids provided by the AdWords Keyword Planner tool and then adjust from there depending on what ad positions I want to test first. I usually set exact match bids at about 100% of the estimated bid, the phrase match keywords at 75% and broad match keywords at 50%.

      If you set your bids so low that your ads don't appear on the first page for search results then you are not going to get much data, and you need data to make informed decisions. If you have a very small budget then reduce the scope of your campaigns, focus on a handful of keywords at first and optimize that small set of keywords before moving on.

      For new campaigns I generally make adjustments to keep the bids somewhere around position 5-7 during the first 1000 impressions to give the Quality Scores a chance to update. Then I proceed with testing higher ad positions in stages.

      Budget permitting, you should try to test your all your ads at various ad positions during the first month of your campaign. This will give you data to see how much volume can be achieved at each ad position and what the relative cost, CTR, and conversions will be for those ad positions. This is the data you need to optimize your bids, so don't be squeamish, you have to get the data to make data informed decisions.

      You can begin adjusting your bids after the first day, but it is definitely too soon to start tweaking ads unless you are fixing something that is obviously broken.

      You test your ads for as long as it takes to see actionable data, or until you see that there is no statistically significant difference. Do not react prematurely, that is a common newbie mistake. If you have an ad with only a handful of clicks the results are totally random at that point and cannot be relied upon for valid decision making.

      Search online for a statistical significance calculator, learn how to use it until you have the experience to recognize statistical significance. A rule-of-thumb is that the bigger the difference the smaller the sample size needed to reach statistical significance.

      For example, if one ad has 5 clicks and the other only 1 click, that is far too few clicks to draw a valid conclusion from, however, if both ads have received about the same number of impressions, and one has 50 clicks and the other has only 1 or 2, then you don't need to wait to take action, that is a huge difference even though the sample size is still small and you should act on it.

      I am not sure I agree with your understanding of Perry's advice regarding long tail keywords. Do not get too general in your keyword selection, that is usually a mistake. Perhaps what Perry was suggesting was not to waste time on keywords that get almost no impressions at all. You should set a minimum number of impressions to make it worth your while and include all relevant keywords that meet that minimum, but chose highly specific keywords that have high commercial intent. I usually set a threshold of at least 10 searches a month.

      Don't try to gauge the number of competitors by what you see when you search, unless you plan to target just your neighborhood. Competition level will vary from town to town, and even neighborhood to neighborhood. Use the Keyword Planner to see the true level of competition. I have found that most of the time low competition keywords are the worst performers and high competition keywords are the best performers, especially those with the highest competing bids.

      HTH,
      Signature

      * Get Results - Outsource Your PPC Management |
      *       Don Burk Advertising & Marketing - www.donburk.com

      {{ DiscussionBoard.errors[9935719].message }}
  • Profile picture of the author Trankuility
    Wonderful. Thank you Lucid and dburk for those detailed and insightful posts.

    After reading your posts I'm a lot more clear now.

    Lucid, you said:

    Why $1?

    What is a typical rate for your keywords? If this was insurance types of keywords, it would be much higher. Maybe that's why you said "Most ads don't show on the first page (I expected this)".
    Typical rate is $2 - $5 for first page.

    $1 was for testing purposes so newbies (such as myself) don't blow money when starting out. I don't remember him giving additional reasons or saying when to increase the bid.

    Don't increase the bids until your QS is better, 7 or more is what I suggest. First thing to do before increasing bids is always to improve your ads and hence your QS.

    You could bite the bullet if you wanted to and increase bids. I assume you want a higher position. But, with lower QS, you'll be paying more than you should and certainly more than competitors.
    I created additional ads today. The QS was 6/10 for all the new keywords. About 30 mins later some decreased to 4/10. The diagnosis shows Landing Page and Relevance as Average and Expected CTR as below average. So I assume that's the reason for the sudden decrease. Since landing page and relevance are ok, does that mean it's best to change the ad copy to increase CTR which will then increase QS? If so, should I do that immediately or still wait until I get 20+ clicks?


    dburk, you said:

    ...choose highly specific keywords that have high commercial intent. I usually set a threshold of at least 10 searches a month.

    Don't try to gauge the number of competitors by what you see when you search, unless you plan to target just your neighborhood. Competition level will vary from town to town, and even neighborhood to neighborhood. Use the Keyword Planner to see the true level of competition. I have found that most of the time low competition keywords are the worst performers and high competition keywords are the best performers, especially those with the highest competing bids.
    When I think of "high commercial intent" I think of keywords that include words like
    • Coupon
    • Discount
    • Deal
    • Shipping
    or keywords that include brand names, etc.

    But how viable is a niche that doesn't include those types of phrases? For instance, if you're selling a course on meditation, no one is gonna search on "Meditation Coupon", haha. And if you're not a known meditation guru they won't be searching for your brand.

    There are tons of searches on things like How To Meditate, Meditation Techniques, etc. Those sound more informational intent to me, but are they worthwhile choices?
    {{ DiscussionBoard.errors[9938010].message }}
    • Profile picture of the author dburk
      Hi Trankuility

      Using your example of a "course on meditation"

      "how to meditate" and "meditation techniques" would be examples of low to medium commercial intent keywords, just information seeking searches.

      High commercial intent keywords would look like:
      meditation courses
      meditation dvds
      online meditation courses
      meditation classes
      meditation instructions under $100
      where to buy meditation courses

      It is easy to spot high commercial intent keywords using the AdWords Keyword Planner. Just look for keywords with high competition and higher than average CPC.
      Signature

      * Get Results - Outsource Your PPC Management |
      *       Don Burk Advertising & Marketing - www.donburk.com

      {{ DiscussionBoard.errors[9938068].message }}
  • Profile picture of the author LucidWebMarketing
    When you add new keywords and create ads, the system does not yet have all the information it needs yet. It makes assumptions about your QS based on current data. It then goes out and does a more in-depth analysis. That's why you may have a certain QS the second you make changes and then a short time later it changes dramatically. Don't pay any attention to QS at the moment you make changes.

    Instead of trying new ads when you see your QS being low, run them for some time. Get some data and information about your ads so you can learn from them. For this, you need to run them which you really are not doing right now. So yes, wait until you get some clicks. Otherwise, you are just creating ads with nothing to back up your decision to do so except the low QS. You need to gather data. By the way, the system needs it too for QS calculations.
    Signature
    Profitable PPC Campaigns Don't Happen By Accident
    {{ DiscussionBoard.errors[9938927].message }}

Trending Topics