2.2K members online now
2.2K members online now
Improve your AdWords performance and boost your ROI, CTR, and Quality Score
Guide Me

Generating traffic with keyword "Discovery" campaign

Follower ✭ ☆ ☆
# 1
Follower ✭ ☆ ☆

This may be a question with many answers (and a lengthy explanation), but if anyone is willing to stick with me here, I'd really appreciate it.


I have 2 long term goals for this account:

--- Maintain a positive ROI at the account level; 110-115% is ideal. (When I started with the account they were over-spending by quite a bit)

--- Increase conversions (Thank-you page after trial download form)


For each product (there are only a few) I have the following campaigns:

Product #1 - Branded - Exact & Phrase match - CPC (Low Volume/Impressions, High CTR, High Conv. Rate)

Product #1 - High Conv. Rate - Exact & Phrase match - CPC (Low Volume/Impressions, High CTR, High Conv. Rate)

Product #1 - "Discovery" - Phrase & mBroad match" - Budget Optimizer (High volume/impressions, Low CTR, New keywords)


Within these campaigns are ad groups based on kw theme and match type.


My highest converting campaigns are the "Branded" campaign and the "High Conv. Rate" campaign --- which makes sense, but these campaigns don't get many impressions. Not enough to generate the kind of download leads the folks upstairs are expecting...


So, I've been trying to find new high performing keywords with a Budget Optimizer "Discovery" campaign that I've been dumping new keywords and ad groups in to. As keywords hit 100 impressions I look at CTR, then again at 200 impressions. If the keyword is below 1-2% CTR I pause it and tag it with a "Failed CTR Test" label, then add some new keywords to replace the paused ones.


I've been at it a couple weeks now with this structure and have yet to find anything worthy of moving into a high conversion campaign. I'm just not getting enough clicks on any keyword to evaluate conversion rate with any significance.


Generating clicks with broader keywords seems like it may be out of the question due to the high CPC --- with a download value of $41, even if a kw converts @ 10% my max CPC would be $3.74. This puts many keywords that might be more relevant out of my budget...

Does it make sense to over-pay for keywords that have high conversion (even if RPC < CPC) then try to make up the difference with low CPC keywords?


Any thoughts on how I could generate more traffic? I'm quickly running out of time to improve the account Smiley Sad

2 Expert replyverified_user

Re: Generating traffic with keyword "Discovery" campaign

[ Edited ]
Follower ✭ ✭ ✭
# 2
Follower ✭ ✭ ✭

take all the relevant keywords and put them in the high conversion rate campaign. make them exact match, and/or modified broad match.


don't bother so much about the conversion rate, just take care of the relevance- the conversion rates will sort themselves out.


you can use cross products to generate new, relevant KWs



column 1                          column 2              column 3:


blue                                  widget                 online

red                                   widgets                website



a cross multiplication of the three columns will yield 12 relevant keywords.


blue widget online

blue widget website 




to summarize:


a) use modified broad match instead of phrase match

b) add as many and all the relevant keywords(in exact match and modified broad match form) to the second campaign.

c) use columns and cross multiplication to generate new keywords.

Re: Generating traffic with keyword "Discovery" campaign

Top Contributor
# 3
Top Contributor

Hello, S-Klep.


In my view you're very close to the right path, but looking at the wrong sign Smiley Happy. It's not keywords that you should be looking at in the discovery campaign, but rather search terms. One broad match keyword can bring you thousands of different search terms, and the fact that you look at that keyword's performance and decide what to do with it does not help.


The discovery campaign's job is to discover search terms. If you download search terms reports often enough, you may identify searches that occur often enough (you'll have to decide a threshold), you may bring them to your high-converting campaigns and try them out. Also, if you find searches which are completely irrelevant, you should use them as negatives in the discovery campaign and prevent them from re-occuring, and thus reducing your future costs for that keyword.


Also, if you see that new search terms are somewhat different in wording than the new ones, you may create separate ad groups in the high converting campaigns, and make sure the ads are as close to them as possible in terms of wording, keyword usage in the ad, etc.

Calin Sandici, AdWords Top Contributor | Find me on: Google+ | Twitter | LinkedIn | myBlog
Was my response helpful? If yes, please mark it as the ‘Best Answer.’ Learn how here.
Marked as Best Answer.
Accepted by topic author S-Klep
September 2015

Re: Generating traffic with keyword "Discovery" campaign

Top Contributor
# 4
Top Contributor

Hi S-Klep, although you've got a good strategy, like Calin I'm also a little concerned with the degree of your focus and where it's looking.  Poor CTR for example can often indicate an issue with the Ads tied to those Keywords and if I have low CTR Keywords that I believe should be relevant - and which appear to be getting relevant search term triggers - I'd be editing the Ads to improve that CTR, not just dumping the Keywords.


My "discovery" group would probably be a Campaign on its own with a low daily budget and very broad terms.  As Calin has said, I'd use this group to "trawl" the search terms it picks up looking for frequent occurences that I would then place into another, higher budget supply, Group elsewhere.  You'd need to think about how this "broad" group would interact with your exsiting Keywords and you wouldn't want it to run forever (unless it's possible your Keywords will change over time).



AdWords Top Contributor Google+ Profile | Partner Profile | AdWords Audits

Re: Generating traffic with keyword "Discovery" campaign

Follower ✭ ☆ ☆
# 5
Follower ✭ ☆ ☆

Thanks for the help! I appreciate it!


I do also pay close attention to the search query report and I add new keywords to my discovery campaign as I find relevant terms. But often there isn't much variation from the keyword in my campaign --- because I'm using nearly all modBroad and phrase match.

The more I describe it, my discovery campaign sounds less and less like a real "discovery" campaign and more like a campaign full of marginally relevant keywords without much purpose.


Typically the broad match keywords that are relevant to my products have a First Page bid that is way out of my price range. Should I go for them anyway in their own campaign with a very limited budget? Seems like it would be a big waste of money to pay $5.00+ for a click that I have no idea how well it will convert... Maybe I'm missing the point?


I feel like I'm so close to having this figured out, but I haven't had that lightbulb moment where it all clicks yet Smiley Frustrated

Re: Generating traffic with keyword "Discovery" campaign

Top Contributor
# 6
Top Contributor

"The more I describe it, my discovery campaign sounds less and less like a real "discovery" campaign and more like a campaign full of marginally relevant keywords without much purpose."

Well, you said it, not me Smiley Happy  Seriously, there is always a benefit in keeping an eye out for better/more Keywords or different match types so don't be disheartened.  However, like many other things with AdWords you also have to consider that there may be no more useful Keywords for you at this stage.  Just keep aware of the possibility and do keep "playing".


As for the high CPC Keywords... that's a tough one.  It really depends upon the value of your Conversions.  I don't like to talk in terms of "cost" because it moves the focus to the wrong way of thinking.  It's better to think in terms of (return on) investment.  $5 a click is a lot if your product is only retailing at $80, but it's practically a pittance if you're selling new BMWs.


The good news is that in theory you can experiment with Ads and landing pages on the lower CPC Keywords and most of that learning should translate to the higher CPC ones.  In other words, what attracts people to click the Ads and complete the sale on the low CPC Keywords is probably going to be very close to what will attract and sell on the higher ones.  It's not a given, but it's something to think about.



AdWords Top Contributor Google+ Profile | Partner Profile | AdWords Audits

Re: Generating traffic with keyword "Discovery" campaign

Follower ✭ ☆ ☆
# 7
Follower ✭ ☆ ☆

Good tip. Your perspective is really helping me re-evaluate my focus here.


I get what you're saying about the high CPC keywords, and I completely. I should have been a little more clear with the issue/mental-roadblock I'm stuck with.


I've been setting bids based on this:

(Max CPC for target ROAS) = ((Goal Value)/(Target ROAS)) x (Conv. Rate)


So I created a table for reference where I input a target ROAS and a conversion rate based on past performance and I get a Max CPC.


This leaves me with an issue though, even at a high conversion rate, to get a positive return my max CPC is often much lower than the estimated first page bid for keywords. If I bid higher, history shows that I get more conversions, but I also end up paying way more than my target CPA. I lowered bids to return the account to a postive ROI, but I've seen the total number of conversions take a dive as well...


Re: Generating traffic with keyword "Discovery" campaign

Top Contributor
# 8
Top Contributor

Hi S-Klep, I have to say I'm struggling with your equations here - probably because I don't know how you're plugging the numbers in (i.e. as percentages, decimals, dollar values, etc.) but I've played around a bit and I can't make the math work.  My equivalent would be something like:


Max. CPC = CPA/CR where CPA is the desired Cost-Per-Acquisition/Action and CR is the Conversion Rate expressed as a percentage.  So, if we plug some random numbers in here, if I want my CPA to be no more than $80 and my typical conversion rate is 10% (nice easy figures), my Max. CPC is $8 (80/10).


I don't think you're doing anything wrong with your approach - it's a typical problem with AdWords that many advertisers face.  As you increase the spend the conversions go up but so do the costs so it's all about finding a balance.  Have you tried plotting the data as a graph?  This should make it pretty easy to see what the "best" CPC would be in terms of ROI/ROAS (we're just full of abbreviations today!).


I might have a go at building a plotter just for fun, when I've got some time...


The only other thing I'd say is that it's easy to forget the "Max" in Max. CPC.  Although the equation above may say I can't pay more than $8 per click to achieve my desired CPA, it doesn't mean I am going to pay $8 per click.  Given the nature of the auction process and variations over time I may well be paying only $5 or $7.50 there etc.  If this were a real example I might well set my Max. CPC at $8 then come and examine the data in three weeks time and found that on average I only paid $6 with my Max CPC at that setting.  I might then risk putting the Max. CPC up to $10.


In short I wouldn't treat your output CPC figures as too literal.  You need to test and evaluate - carefully - but you may find you can actually afford a higher CPC than raw math predicts because you're not always paying this Max value.



AdWords Top Contributor Google+ Profile | Partner Profile | AdWords Audits

Re: Generating traffic with keyword "Discovery" campaign

Follower ✭ ☆ ☆
# 9
Follower ✭ ☆ ☆

"I might have a go at building a plotter just for fun, when I've got some time..."

You sir are a gentleman and a scholar.


Our forumlas are the same. Minor differences, but they produce the same results in the end.


I believe the plotter you're describing is similar to something I took a stab at when I came up with my reference sheet a few weeks ago. Kind of like the Traffic Estimator tool, but for conversions. But I got stuck at the time. I was attempting to answer the question "How many goal completions/conversions can I expect at a max CPC of $X and a target ROAS of X%?" What I ended up with is a formula that requires me to input the number of clicks I would receive. Everything else seemed to work fine, but I was left with trying to guess how many clicks I would get from a particular keyword.

As things are playing out, my ROI is right on, but I need to scale up clicks.


Just starting out with this AdWords stuff, it can be an overwhelming amount of data to sort through and organize Smiley Frustrated



Re: Generating traffic with keyword "Discovery" campaign

Top Contributor
# 10
Top Contributor

Indeed, you're spot on, clicks would have to be a part of that forumula.  It'd really only work once you had some fairly stable data and you might have to "prepare" for it by recording somehow typical CTR figures for typical positions.  It'd be a challenge.



AdWords Top Contributor Google+ Profile | Partner Profile | AdWords Audits